You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by tq...@apache.org on 2022/09/12 21:39:09 UTC
[tvm-site] branch asf-site updated: deploying docs (apache/tvm@b22b872da800b0b44feeca67e808319e21b840a2)
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/tvm-site.git
The following commit(s) were added to refs/heads/asf-site by this push:
new f6fbe9315 deploying docs (apache/tvm@b22b872da800b0b44feeca67e808319e21b840a2)
f6fbe9315 is described below
commit f6fbe93150c653faca7ad8beb42157a65a797858
Author: tvm-bot <95...@users.noreply.github.com>
AuthorDate: Mon Sep 12 21:39:01 2022 +0000
deploying docs (apache/tvm@b22b872da800b0b44feeca67e808319e21b840a2)
---
.../how_to/compile_models/from_darknet.rst.txt | 2 +-
.../how_to/compile_models/from_keras.rst.txt | 2 +-
.../how_to/compile_models/from_mxnet.rst.txt | 2 +-
.../how_to/compile_models/from_oneflow.rst.txt | 2 +-
.../how_to/compile_models/from_pytorch.rst.txt | 2 +-
.../how_to/compile_models/from_tensorflow.rst.txt | 2 +-
.../compile_models/sg_execution_times.rst.txt | 22 +-
.../deploy_models/deploy_model_on_android.rst.txt | 2 +-
.../deploy_object_detection_pytorch.rst.txt | 4 +-
.../deploy_models/deploy_prequantized.rst.txt | 6 +-
.../deploy_prequantized_tflite.rst.txt | 4 +-
.../how_to/deploy_models/deploy_quantized.rst.txt | 2 +-
.../deploy_models/deploy_ssd_gluoncv.rst.txt | 4 +-
.../deploy_models/sg_execution_times.rst.txt | 18 +-
.../extend_tvm/bring_your_own_datatypes.rst.txt | 2 +-
.../how_to/extend_tvm/sg_execution_times.rst.txt | 10 +-
.../how_to/extend_tvm/use_pass_instrument.rst.txt | 16 +-
.../optimize_operators/opt_conv_cuda.rst.txt | 2 +-
.../optimize_operators/opt_conv_tensorcore.rst.txt | 2 +-
.../how_to/optimize_operators/opt_gemm.rst.txt | 16 +-
.../optimize_operators/sg_execution_times.rst.txt | 8 +-
.../sg_execution_times.rst.txt | 14 +-
.../tune_conv2d_layer_cuda.rst.txt | 1795 +++-
.../tune_network_cuda.rst.txt | 2 +-
.../tune_network_x86.rst.txt | 4 +-
.../tune_sparse_x86.rst.txt | 162 +-
.../tune_with_autotvm/sg_execution_times.rst.txt | 10 +-
.../tune_with_autotvm/tune_conv2d_cuda.rst.txt | 26 +-
.../work_with_microtvm/micro_autotune.rst.txt | 16 +-
.../how_to/work_with_microtvm/micro_train.rst.txt | 16 +-
.../work_with_microtvm/sg_execution_times.rst.txt | 10 +-
.../work_with_relay/sg_execution_times.rst.txt | 8 +-
.../how_to/work_with_schedules/intrin_math.rst.txt | 2 +-
.../work_with_schedules/sg_execution_times.rst.txt | 14 +-
.../how_to/work_with_schedules/tensorize.rst.txt | 2 +-
.../tutorials/autotvm/sg_execution_times.rst.txt | 6 +-
.../frontend/deploy_classification.rst.txt | 2 +-
.../tutorials/frontend/deploy_detection.rst.txt | 2 +-
.../tutorials/frontend/sg_execution_times.rst.txt | 6 +-
.../tutorials/optimize/sg_execution_times.rst.txt | 6 +-
.../topic/vta/tutorials/sg_execution_times.rst.txt | 6 +-
.../tutorial/auto_scheduler_matmul_x86.rst.txt | 11 +-
docs/_sources/tutorial/autotvm_matmul_x86.rst.txt | 20 +-
docs/_sources/tutorial/autotvm_relay_x86.rst.txt | 58 +-
.../tutorial/cross_compilation_and_rpc.rst.txt | 2 +-
docs/_sources/tutorial/intro_topi.rst.txt | 2 +-
docs/_sources/tutorial/sg_execution_times.rst.txt | 24 +-
.../tutorial/tensor_expr_get_started.rst.txt | 44 +-
docs/commit_hash | 2 +-
docs/how_to/compile_models/from_darknet.html | 2 +-
docs/how_to/compile_models/from_keras.html | 2 +-
docs/how_to/compile_models/from_mxnet.html | 2 +-
docs/how_to/compile_models/from_oneflow.html | 14 +-
docs/how_to/compile_models/from_pytorch.html | 5 +-
docs/how_to/compile_models/from_tensorflow.html | 2 +-
docs/how_to/compile_models/sg_execution_times.html | 26 +-
.../deploy_models/deploy_model_on_android.html | 2 +-
.../deploy_object_detection_pytorch.html | 85 +-
docs/how_to/deploy_models/deploy_prequantized.html | 10 +-
.../deploy_models/deploy_prequantized_tflite.html | 4 +-
docs/how_to/deploy_models/deploy_quantized.html | 2 +-
docs/how_to/deploy_models/deploy_ssd_gluoncv.html | 37 +-
docs/how_to/deploy_models/sg_execution_times.html | 18 +-
.../extend_tvm/bring_your_own_datatypes.html | 2 +-
docs/how_to/extend_tvm/sg_execution_times.html | 10 +-
docs/how_to/extend_tvm/use_pass_instrument.html | 16 +-
docs/how_to/optimize_operators/opt_conv_cuda.html | 2 +-
.../optimize_operators/opt_conv_tensorcore.html | 2 +-
docs/how_to/optimize_operators/opt_gemm.html | 16 +-
.../optimize_operators/sg_execution_times.html | 8 +-
.../sg_execution_times.html | 14 +-
.../tune_conv2d_layer_cuda.html | 1795 +++-
.../tune_with_autoscheduler/tune_network_cuda.html | 2 +-
.../tune_with_autoscheduler/tune_network_x86.html | 4 +-
.../tune_with_autoscheduler/tune_sparse_x86.html | 162 +-
.../tune_with_autotvm/sg_execution_times.html | 10 +-
.../how_to/tune_with_autotvm/tune_conv2d_cuda.html | 26 +-
docs/how_to/work_with_microtvm/micro_autotune.html | 16 +-
docs/how_to/work_with_microtvm/micro_train.html | 16 +-
.../work_with_microtvm/sg_execution_times.html | 10 +-
.../how_to/work_with_relay/sg_execution_times.html | 8 +-
docs/how_to/work_with_schedules/intrin_math.html | 2 +-
.../work_with_schedules/sg_execution_times.html | 14 +-
docs/how_to/work_with_schedules/tensorize.html | 2 +-
docs/install/nnpack.html | 12 +-
docs/reference/api/doxygen/annotated.html | 240 +-
docs/reference/api/doxygen/c__runtime__api_8h.html | 2 +-
.../api/doxygen/c__runtime__api_8h__dep__incl.svg | 1096 +-
docs/reference/api/doxygen/classes.html | 456 +-
..._1runtime_1_1hexagon_1_1SDLTensor-members.html} | 16 +-
...lasstvm_1_1runtime_1_1hexagon_1_1SDLTensor.html | 236 +
...untime_1_1hexagon_1_1SDLTensor__coll__graph.svg | 42 +
...ime_1_1hexagon_1_1SDLTensor__inherit__graph.svg | 42 +
docs/reference/api/doxygen/conv2d_8h.html | 143 +
docs/reference/api/doxygen/conv2d_8h__incl.svg | 937 ++
docs/reference/api/doxygen/conv2d_8h_source.html | 102 +
.../api/doxygen/data__type_8h__dep__incl.svg | 108 +-
docs/reference/api/doxygen/device__api_8h.html | 2 +-
.../api/doxygen/device__api_8h__dep__incl.svg | 164 +-
...r_000003_000029.html => dir_000003_000031.html} | 0
...r_000003_000030.html => dir_000003_000032.html} | 0
...r_000004_000029.html => dir_000004_000031.html} | 0
...r_000005_000029.html => dir_000005_000031.html} | 0
...r_000005_000030.html => dir_000005_000032.html} | 0
...r_000006_000029.html => dir_000006_000031.html} | 0
...r_000006_000030.html => dir_000006_000032.html} | 0
...r_000007_000029.html => dir_000007_000031.html} | 0
...r_000011_000029.html => dir_000011_000031.html} | 0
...r_000014_000029.html => dir_000014_000031.html} | 0
...r_000015_000029.html => dir_000015_000031.html} | 0
...r_000016_000029.html => dir_000016_000031.html} | 0
...r_000016_000030.html => dir_000016_000032.html} | 0
docs/reference/api/doxygen/dir_000026_000007.html | 73 -
docs/reference/api/doxygen/dir_000026_000017.html | 73 -
docs/reference/api/doxygen/dir_000027_000011.html | 73 -
docs/reference/api/doxygen/dir_000028_000007.html | 6 +-
docs/reference/api/doxygen/dir_000028_000008.html | 73 -
docs/reference/api/doxygen/dir_000028_000013.html | 73 -
docs/reference/api/doxygen/dir_000028_000017.html | 6 +-
docs/reference/api/doxygen/dir_000029_000007.html | 73 -
docs/reference/api/doxygen/dir_000029_000011.html | 6 +-
docs/reference/api/doxygen/dir_000029_000013.html | 73 -
docs/reference/api/doxygen/dir_000029_000017.html | 73 -
...r_000027_000026.html => dir_000029_000028.html} | 0
docs/reference/api/doxygen/dir_000030_000002.html | 73 -
...r_000028_000007.html => dir_000030_000007.html} | 0
docs/reference/api/doxygen/dir_000030_000008.html | 6 +-
docs/reference/api/doxygen/dir_000030_000013.html | 6 +-
...r_000028_000017.html => dir_000030_000017.html} | 0
docs/reference/api/doxygen/dir_000031_000007.html | 6 +-
...r_000029_000008.html => dir_000031_000008.html} | 0
...r_000029_000011.html => dir_000031_000011.html} | 0
docs/reference/api/doxygen/dir_000031_000013.html | 6 +-
docs/reference/api/doxygen/dir_000031_000017.html | 6 +-
docs/reference/api/doxygen/dir_000032_000002.html | 6 +-
docs/reference/api/doxygen/dir_000032_000008.html | 6 +-
...r_000030_000011.html => dir_000032_000011.html} | 0
...r_000030_000013.html => dir_000032_000013.html} | 0
...r_000031_000007.html => dir_000033_000007.html} | 0
...r_000031_000013.html => dir_000033_000013.html} | 0
...r_000031_000017.html => dir_000033_000017.html} | 0
...r_000032_000002.html => dir_000034_000002.html} | 0
...r_000032_000008.html => dir_000034_000008.html} | 0
...r_000033_000002.html => dir_000035_000002.html} | 0
...r_000033_000011.html => dir_000035_000011.html} | 0
...r_000033_000030.html => dir_000035_000032.html} | 0
docs/reference/api/doxygen/dir_000035_000033.html | 73 -
docs/reference/api/doxygen/dir_000035_000034.html | 73 -
docs/reference/api/doxygen/dir_000035_000036.html | 73 -
...r_000034_000030.html => dir_000036_000032.html} | 0
...r_000034_000033.html => dir_000036_000035.html} | 0
...r_000035_000029.html => dir_000037_000031.html} | 0
...r_000035_000030.html => dir_000037_000032.html} | 0
docs/reference/api/doxygen/dir_000037_000035.html | 6 +-
docs/reference/api/doxygen/dir_000037_000036.html | 6 +-
docs/reference/api/doxygen/dir_000037_000038.html | 6 +-
...r_000035_000038.html => dir_000037_000040.html} | 0
...r_000036_000002.html => dir_000038_000002.html} | 0
...r_000036_000030.html => dir_000038_000032.html} | 0
...r_000036_000033.html => dir_000038_000035.html} | 0
...r_000037_000029.html => dir_000039_000031.html} | 0
...r_000037_000030.html => dir_000039_000032.html} | 0
...r_000037_000033.html => dir_000039_000035.html} | 0
...r_000037_000034.html => dir_000039_000036.html} | 0
...r_000037_000035.html => dir_000039_000037.html} | 0
...r_000037_000036.html => dir_000039_000038.html} | 0
...r_000037_000038.html => dir_000039_000040.html} | 0
...r_000038_000029.html => dir_000040_000031.html} | 0
...r_000038_000030.html => dir_000040_000032.html} | 0
...r_000038_000033.html => dir_000040_000035.html} | 0
...r_000039_000029.html => dir_000041_000031.html} | 0
...r_000039_000030.html => dir_000041_000032.html} | 0
...r_000039_000033.html => dir_000041_000035.html} | 0
...r_000040_000030.html => dir_000042_000032.html} | 0
...r_000040_000033.html => dir_000042_000035.html} | 0
.../dir_006b1f4ac353a18abb55f74cc4796db6_dep.svg | 6 +-
.../dir_02be2c9d68e402f80df60bd528724ee5_dep.svg | 22 +-
.../dir_05ffda4d144d7985f926507abde48dbb_dep.svg | 12 +-
.../dir_1f1b12d204a071c9e67e47fcbb552b86_dep.svg | 10 +-
.../dir_2b0ef9f1c86b565a92e96353e1195b2c_dep.svg | 8 +-
.../dir_3a038e7bfa2370c6aee2a5aecd5d3ef1_dep.svg | 12 +-
.../dir_404558507ed35459f0d7a6d81d8c508d.html | 4 +-
.../dir_404558507ed35459f0d7a6d81d8c508d_dep.svg | 107 +-
.../dir_4378f18824ae7d4ad48f8d7785cd7ac8_dep.svg | 16 +-
.../dir_519be2d4a83a987dbf989f1de527b870_dep.svg | 10 +-
.../dir_54983dd6d74c59f67ee9e8e5a50aafc4_dep.svg | 42 +-
.../dir_5da96592f3a7c442b838b075c58254c2_dep.svg | 14 +-
.../dir_63946bee875c6d52bce55e72a67a86ad_dep.svg | 20 +-
.../dir_67fdee7a5e0396034822418fa5baa4b4_dep.svg | 4 +-
.../dir_72c2f11201cd7636dc7624de0754daa5_dep.svg | 22 +-
.../dir_8395ded0a3205c0748976a0d4487d38d_dep.svg | 8 +-
.../dir_84875704194fd544d29fe0c7fedd8939_dep.svg | 8 +-
.../dir_8e4e25e66b8623d88c5b5dd2040bca97_dep.svg | 74 +-
.../dir_9e615ec4a59e46584bcc4e2226e148a2_dep.svg | 8 +-
.../dir_a59a89c7dd2e4e6561fe59bf359ce2f3_dep.svg | 8 +-
.../dir_a98464176f1216e334ac3bbacd433085_dep.svg | 16 +-
.../dir_ac57496531ccbad72f774fa62e6de987_dep.svg | 28 +-
...l => dir_ad74dcb015f5755a482b0a16aa23ec44.html} | 22 +-
.../dir_ad74dcb015f5755a482b0a16aa23ec44_dep.svg | 30 +
.../dir_b4c7d8e826c599ba55146c099a14beb5_dep.svg | 40 +-
.../dir_d331277d4303e21ded95616eb56c1a9e_dep.svg | 6 +-
.../dir_d3953cf7eb33eca56fc6850c0e98447d_dep.svg | 6 +-
.../dir_d4a54fa981698f72ef4cd62f8b9e1a8f_dep.svg | 4 +-
...l => dir_da796f6f071225ab18a1002406aa03ea.html} | 22 +-
.../dir_da796f6f071225ab18a1002406aa03ea_dep.svg | 42 +
.../dir_dc867ff9a37cad1764f1670dc7eba6c1_dep.svg | 12 +-
.../dir_e4a1a856a30057b9b1543256279fc7a1_dep.svg | 4 +-
.../dir_f97d855a3173728370e632aa77170e34_dep.svg | 8 +-
docs/reference/api/doxygen/files.html | 57 +-
docs/reference/api/doxygen/functions_func_g.html | 9 +-
docs/reference/api/doxygen/functions_func_s.html | 15 +-
docs/reference/api/doxygen/functions_g.html | 9 +-
docs/reference/api/doxygen/functions_l.html | 6 +-
docs/reference/api/doxygen/functions_s.html | 15 +-
docs/reference/api/doxygen/functions_t.html | 6 +-
docs/reference/api/doxygen/hierarchy.html | 2770 ++---
docs/reference/api/doxygen/inherit_graph_100.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_101.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_102.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_103.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_104.svg | 18 +-
docs/reference/api/doxygen/inherit_graph_105.svg | 18 +-
docs/reference/api/doxygen/inherit_graph_106.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_107.svg | 4 +-
docs/reference/api/doxygen/inherit_graph_108.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_109.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_110.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_111.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_112.svg | 4 +-
docs/reference/api/doxygen/inherit_graph_113.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_114.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_115.svg | 27 +-
docs/reference/api/doxygen/inherit_graph_116.svg | 10434 +------------------
docs/reference/api/doxygen/inherit_graph_117.svg | 10434 ++++++++++++++++++-
docs/reference/api/doxygen/inherit_graph_118.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_119.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_120.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_121.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_122.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_123.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_124.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_125.svg | 7354 +------------
docs/reference/api/doxygen/inherit_graph_126.svg | 7355 ++++++++++++-
docs/reference/api/doxygen/inherit_graph_127.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_128.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_129.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_130.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_131.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_132.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_133.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_134.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_135.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_136.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_137.svg | 4 +-
docs/reference/api/doxygen/inherit_graph_138.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_139.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_140.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_141.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_142.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_143.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_144.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_145.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_146.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_147.svg | 61 +-
docs/reference/api/doxygen/inherit_graph_148.svg | 62 +-
docs/reference/api/doxygen/inherit_graph_149.svg | 21 +-
docs/reference/api/doxygen/inherit_graph_150.svg | 21 +-
docs/reference/api/doxygen/inherit_graph_151.svg | 20 +-
docs/reference/api/doxygen/inherit_graph_152.svg | 20 +-
docs/reference/api/doxygen/inherit_graph_153.svg | 19 +-
docs/reference/api/doxygen/inherit_graph_154.svg | 18 +-
docs/reference/api/doxygen/inherit_graph_155.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_156.svg | 20 +-
docs/reference/api/doxygen/inherit_graph_157.svg | 19 +-
docs/reference/api/doxygen/inherit_graph_158.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_159.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_160.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_161.svg | 4 +-
docs/reference/api/doxygen/inherit_graph_162.svg | 4 +-
docs/reference/api/doxygen/inherit_graph_163.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_164.svg | 6 +-
docs/reference/api/doxygen/inherit_graph_165.svg | 20 +-
docs/reference/api/doxygen/inherit_graph_166.svg | 19 +-
docs/reference/api/doxygen/inherit_graph_167.svg | 4 +-
docs/reference/api/doxygen/inherit_graph_168.svg | 4 +-
docs/reference/api/doxygen/inherit_graph_169.svg | 21 +-
docs/reference/api/doxygen/inherit_graph_170.svg | 21 +-
docs/reference/api/doxygen/inherit_graph_171.svg | 19 +-
docs/reference/api/doxygen/inherit_graph_172.svg | 19 +-
docs/reference/api/doxygen/inherit_graph_173.svg | 18 +-
docs/reference/api/doxygen/inherit_graph_174.svg | 21 +-
docs/reference/api/doxygen/inherit_graph_175.svg | 24 +-
docs/reference/api/doxygen/inherit_graph_176.svg | 21 +-
docs/reference/api/doxygen/inherit_graph_177.svg | 18 +-
docs/reference/api/doxygen/inherit_graph_178.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_179.svg | 18 +-
docs/reference/api/doxygen/inherit_graph_180.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_181.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_182.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_183.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_184.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_185.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_186.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_187.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_188.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_189.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_190.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_191.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_192.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_193.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_194.svg | 16 +-
docs/reference/api/doxygen/inherit_graph_195.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_196.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_197.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_198.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_199.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_200.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_201.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_202.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_203.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_204.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_205.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_206.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_207.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_208.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_209.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_210.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_211.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_212.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_213.svg | 17 +-
docs/reference/api/doxygen/inherit_graph_214.svg | 80 +-
docs/reference/api/doxygen/inherit_graph_215.svg | 70 +-
docs/reference/api/doxygen/inherit_graph_216.svg | 79 +-
docs/reference/api/doxygen/inherit_graph_217.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_218.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_219.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_220.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_221.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_222.svg | 14 +-
docs/reference/api/doxygen/inherit_graph_223.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_224.svg | 18 +-
docs/reference/api/doxygen/inherit_graph_225.svg | 19 +-
docs/reference/api/doxygen/inherit_graph_226.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_227.svg | 15 +-
docs/reference/api/doxygen/inherit_graph_228.svg | 29 +-
docs/reference/api/doxygen/inherit_graph_229.svg | 24 +-
docs/reference/api/doxygen/inherit_graph_230.svg | 30 +-
docs/reference/api/doxygen/inherit_graph_231.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_232.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_233.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_234.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_235.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_236.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_237.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_238.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_239.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_240.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_241.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_242.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_243.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_244.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_245.svg | 12 +-
docs/reference/api/doxygen/inherit_graph_246.svg | 12 +-
...inherit_graph_246.svg => inherit_graph_247.svg} | 0
docs/reference/api/doxygen/inherit_graph_99.svg | 31 +-
docs/reference/api/doxygen/inherits.html | 282 +-
docs/reference/api/doxygen/map_8h__dep__incl.svg | 48 +-
docs/reference/api/doxygen/menudata.js | 2 +
docs/reference/api/doxygen/namespacemembers_b.html | 7 +-
docs/reference/api/doxygen/namespacemembers_c.html | 10 +-
docs/reference/api/doxygen/namespacemembers_d.html | 7 +-
.../api/doxygen/namespacemembers_func_b.html | 7 +-
.../api/doxygen/namespacemembers_func_c.html | 10 +-
.../api/doxygen/namespacemembers_func_d.html | 7 +-
.../api/doxygen/namespacemembers_func_h.html | 6 +
.../api/doxygen/namespacemembers_func_n.html | 5 +-
.../api/doxygen/namespacemembers_func_p.html | 8 +-
.../api/doxygen/namespacemembers_func_r.html | 10 +-
.../api/doxygen/namespacemembers_func_t.html | 10 +-
...rs_func_h.html => namespacemembers_func_x.html} | 18 +-
docs/reference/api/doxygen/namespacemembers_h.html | 6 +
docs/reference/api/doxygen/namespacemembers_n.html | 5 +-
docs/reference/api/doxygen/namespacemembers_p.html | 6 +
docs/reference/api/doxygen/namespacemembers_r.html | 10 +-
docs/reference/api/doxygen/namespacemembers_t.html | 6 +
...members_func_h.html => namespacemembers_x.html} | 20 +-
docs/reference/api/doxygen/namespaces.html | 63 +-
.../api/doxygen/namespacetvm_1_1runtime.html | 2 +
.../namespacetvm_1_1runtime_1_1hexagon.html | 725 ++
docs/reference/api/doxygen/ndarray_8h.html | 2 +-
.../api/doxygen/ndarray_8h__dep__incl.svg | 1070 +-
.../reference/api/doxygen/object_8h__dep__incl.svg | 212 +-
.../api/doxygen/optional_8h__dep__incl.svg | 116 +-
.../api/doxygen/packed__func_8h__dep__incl.svg | 80 +-
.../runtime_2container_2base_8h__dep__incl.svg | 320 +-
.../api/doxygen/runtime_2memory_8h__dep__incl.svg | 80 +-
.../api/doxygen/runtime_2module_8h__dep__incl.svg | 68 +-
docs/reference/api/doxygen/search/all_11.js | 2 +
docs/reference/api/doxygen/search/all_13.js | 8 +-
docs/reference/api/doxygen/search/all_14.js | 17 +-
docs/reference/api/doxygen/search/all_15.js | 9 +-
docs/reference/api/doxygen/search/all_19.js | 3 +-
docs/reference/api/doxygen/search/all_1a.js | 3 +-
docs/reference/api/doxygen/search/all_1b.js | 57 +-
docs/reference/api/doxygen/search/all_1c.html | 26 +
.../api/doxygen/search/{all_1b.js => all_1c.js} | 0
docs/reference/api/doxygen/search/all_3.js | 1 +
docs/reference/api/doxygen/search/all_4.js | 3 +
docs/reference/api/doxygen/search/all_5.js | 1 +
docs/reference/api/doxygen/search/all_7.js | 2 +-
docs/reference/api/doxygen/search/all_8.js | 1 +
docs/reference/api/doxygen/search/all_9.js | 4 +-
docs/reference/api/doxygen/search/all_d.js | 4 +-
docs/reference/api/doxygen/search/all_f.js | 1 +
docs/reference/api/doxygen/search/classes_10.js | 3 +-
docs/reference/api/doxygen/search/classes_11.js | 2 +-
docs/reference/api/doxygen/search/classes_5.js | 2 +-
docs/reference/api/doxygen/search/classes_7.js | 2 +-
docs/reference/api/doxygen/search/classes_9.js | 2 +-
docs/reference/api/doxygen/search/files_2.js | 1 +
docs/reference/api/doxygen/search/functions_10.js | 2 +
docs/reference/api/doxygen/search/functions_12.js | 6 +-
docs/reference/api/doxygen/search/functions_13.js | 9 +-
docs/reference/api/doxygen/search/functions_14.js | 2 +
docs/reference/api/doxygen/search/functions_18.js | 2 +-
docs/reference/api/doxygen/search/functions_19.js | 57 +-
.../reference/api/doxygen/search/functions_1a.html | 26 +
.../doxygen/search/{all_1b.js => functions_1a.js} | 0
docs/reference/api/doxygen/search/functions_2.js | 1 +
docs/reference/api/doxygen/search/functions_3.js | 2 +
docs/reference/api/doxygen/search/functions_4.js | 1 +
docs/reference/api/doxygen/search/functions_7.js | 1 +
docs/reference/api/doxygen/search/functions_8.js | 2 +
docs/reference/api/doxygen/search/functions_e.js | 1 +
docs/reference/api/doxygen/search/namespaces_1.js | 1 +
docs/reference/api/doxygen/search/searchdata.js | 4 +-
docs/reference/api/doxygen/serializer_8h.html | 2 +-
.../api/doxygen/serializer_8h__dep__incl.svg | 1052 +-
docs/reference/api/doxygen/shape__tuple_8h.html | 2 +-
.../api/doxygen/shape__tuple_8h__dep__incl.svg | 1104 +-
.../reference/api/doxygen/string_8h__dep__incl.svg | 28 +-
docs/reference/api/python/auto_scheduler.html | 4 +-
.../api/typedoc/classes/bytestreamreader.html | 12 +-
.../api/typedoc/classes/cachedcallstack.html | 34 +-
docs/reference/api/typedoc/classes/dldatatype.html | 12 +-
docs/reference/api/typedoc/classes/dldevice.html | 10 +-
.../reference/api/typedoc/classes/environment.html | 12 +-
docs/reference/api/typedoc/classes/ffilibrary.html | 20 +-
.../api/typedoc/classes/graphexecutor.html | 16 +-
docs/reference/api/typedoc/classes/instance.html | 40 +-
docs/reference/api/typedoc/classes/memory.html | 34 +-
docs/reference/api/typedoc/classes/module.html | 10 +-
docs/reference/api/typedoc/classes/ndarray.html | 22 +-
.../api/typedoc/classes/packedfunccell.html | 6 +-
docs/reference/api/typedoc/classes/rpcserver.html | 14 +-
docs/reference/api/typedoc/classes/scalar.html | 6 +-
.../api/typedoc/classes/webgpucontext.html | 12 +-
docs/reference/api/typedoc/enums/argtypecode.html | 30 +-
.../api/typedoc/enums/aynccallbackcode.html | 4 +-
.../api/typedoc/enums/dldatatypecode.html | 8 +-
.../api/typedoc/enums/rpcserverstate.html | 12 +-
docs/reference/api/typedoc/enums/sizeof.html | 18 +-
docs/reference/api/typedoc/index.html | 112 +-
.../api/typedoc/interfaces/disposable.html | 2 +-
.../api/typedoc/interfaces/functioninfo.html | 6 +-
.../api/typedoc/interfaces/libraryprovider.html | 4 +-
docs/searchindex.js | 2 +-
.../vta/tutorials/autotvm/sg_execution_times.html | 6 +-
.../tutorials/frontend/deploy_classification.html | 2 +-
.../vta/tutorials/frontend/deploy_detection.html | 2 +-
.../vta/tutorials/frontend/sg_execution_times.html | 6 +-
.../vta/tutorials/optimize/sg_execution_times.html | 6 +-
docs/topic/vta/tutorials/sg_execution_times.html | 6 +-
docs/tutorial/auto_scheduler_matmul_x86.html | 7 +-
docs/tutorial/autotvm_matmul_x86.html | 20 +-
docs/tutorial/autotvm_relay_x86.html | 262 +-
docs/tutorial/cross_compilation_and_rpc.html | 2 +-
docs/tutorial/intro_topi.html | 2 +-
docs/tutorial/sg_execution_times.html | 30 +-
docs/tutorial/tensor_expr_get_started.html | 44 +-
480 files changed, 31144 insertions(+), 26211 deletions(-)
diff --git a/docs/_sources/how_to/compile_models/from_darknet.rst.txt b/docs/_sources/how_to/compile_models/from_darknet.rst.txt
index 432b99691..17684c91f 100644
--- a/docs/_sources/how_to/compile_models/from_darknet.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_darknet.rst.txt
@@ -317,7 +317,7 @@ The process is no different from other examples.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 1.980 seconds)
+ **Total running time of the script:** ( 1 minutes 3.157 seconds)
.. _sphx_glr_download_how_to_compile_models_from_darknet.py:
diff --git a/docs/_sources/how_to/compile_models/from_keras.rst.txt b/docs/_sources/how_to/compile_models/from_keras.rst.txt
index 067953a63..717a534f2 100644
--- a/docs/_sources/how_to/compile_models/from_keras.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_keras.rst.txt
@@ -228,7 +228,7 @@ Look up prediction top 1 index in 1000 class synset.
.. code-block:: none
Relay top-1 id: 285, class name: Egyptian cat
-
1/1 [==============================] - ETA: 0s
1/1 [==============================] - 1s 897ms/step
+
1/1 [==============================] - ETA: 0s
1/1 [==============================] - 1s 916ms/step
Keras top-1 id: 285, class name: Egyptian cat
diff --git a/docs/_sources/how_to/compile_models/from_mxnet.rst.txt b/docs/_sources/how_to/compile_models/from_mxnet.rst.txt
index def9e7965..8ca0ac448 100644
--- a/docs/_sources/how_to/compile_models/from_mxnet.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_mxnet.rst.txt
@@ -115,7 +115,7 @@ In this section, we download a pretrained imagenet model and classify an image.
.. code-block:: none
- Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zip621d8ff1-749e-4f6f-ba07-09560314ab54 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip...
+ Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zip19c5496d-5bd0-456d-86d0-04121012ec1c from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip...
x (1, 3, 224, 224)
diff --git a/docs/_sources/how_to/compile_models/from_oneflow.rst.txt b/docs/_sources/how_to/compile_models/from_oneflow.rst.txt
index a905a5afc..b071137e3 100644
--- a/docs/_sources/how_to/compile_models/from_oneflow.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_oneflow.rst.txt
@@ -116,7 +116,7 @@ Load a pretrained OneFlow model and save model
.. code-block:: none
Downloading: "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/ResNet/resnet18.zip" to /workspace/.oneflow/flowvision_cache/resnet18.zip
-
0%| | 0.00/41.5M [00:00<?, ?B/s]
20%|#9 | 8.12M/41.5M [00:00<00:00, 85.1MB/s]
39%|###9 | 16.2M/41.5M [00:00<00:00, 71.5MB/s]
56%|#####5 | 23.2M/41.5M [00:00<00:00, 67.7MB/s]
72%|#######1 | 29.7M/41.5M [00:00<00:00, 62.8MB/s]
86%|########6 | 35.8M/41.5M [00:00<00:00, 55.7MB/s]
99%|#########9| 41.2M/41.5M [00:00<00:00, 53.2MB/s]
100%|##########| 41.5M/41.5M [00:00<00:00, 59.5MB/s]
+
0%| | 0.00/41.5M [00:00<?, ?B/s]
19%|#9 | 7.99M/41.5M [00:00<00:00, 49.5MB/s]
39%|###8 | 16.0M/41.5M [00:00<00:00, 50.2MB/s]
55%|#####4 | 22.7M/41.5M [00:00<00:00, 57.1MB/s]
69%|######8 | 28.4M/41.5M [00:00<00:00, 54.8MB/s]
81%|########1 | 33.8M/41.5M [00:00<00:00, 50.5MB/s]
93%|#########3| 38.7M/41.5M [00:00<00:00, 42.5MB/s]
100%|##########| 41.5M/41.5M [00:00<00:00, 47.1MB/s]
diff --git a/docs/_sources/how_to/compile_models/from_pytorch.rst.txt b/docs/_sources/how_to/compile_models/from_pytorch.rst.txt
index 776f8a480..2dc6121c7 100644
--- a/docs/_sources/how_to/compile_models/from_pytorch.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_pytorch.rst.txt
@@ -94,7 +94,7 @@ Load a pretrained PyTorch model
.. code-block:: none
Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /workspace/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
-
0%| | 0.00/44.7M [00:00<?, ?B/s]
34%|###3 | 15.0M/44.7M [00:00<00:00, 157MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 235MB/s]
+
0%| | 0.00/44.7M [00:00<?, ?B/s]
39%|###8 | 17.2M/44.7M [00:00<00:00, 180MB/s]
94%|#########3| 41.8M/44.7M [00:00<00:00, 226MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 222MB/s]
diff --git a/docs/_sources/how_to/compile_models/from_tensorflow.rst.txt b/docs/_sources/how_to/compile_models/from_tensorflow.rst.txt
index 0c4adc0f5..8c7ab1e54 100644
--- a/docs/_sources/how_to/compile_models/from_tensorflow.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_tensorflow.rst.txt
@@ -423,7 +423,7 @@ Run the corresponding model on tensorflow
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 5.350 seconds)
+ **Total running time of the script:** ( 1 minutes 5.653 seconds)
.. _sphx_glr_download_how_to_compile_models_from_tensorflow.py:
diff --git a/docs/_sources/how_to/compile_models/sg_execution_times.rst.txt b/docs/_sources/how_to/compile_models/sg_execution_times.rst.txt
index 6fcc32b1b..e73280607 100644
--- a/docs/_sources/how_to/compile_models/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/compile_models/sg_execution_times.rst.txt
@@ -5,26 +5,26 @@
Computation times
=================
-**05:01.964** total execution time for **how_to_compile_models** files:
+**05:03.518** total execution time for **how_to_compile_models** files:
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:05.350 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:05.653 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``) | 01:01.980 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``) | 01:03.157 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``) | 00:39.257 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``) | 00:38.880 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``) | 00:27.045 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``) | 00:27.838 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``) | 00:25.306 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``) | 00:25.177 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``) | 00:24.765 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``) | 00:24.905 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``) | 00:21.073 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``) | 00:21.325 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``) | 00:19.155 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``) | 00:19.160 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``) | 00:15.638 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``) | 00:14.963 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``) | 00:02.396 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``) | 00:02.460 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/deploy_models/deploy_model_on_android.rst.txt b/docs/_sources/how_to/deploy_models/deploy_model_on_android.rst.txt
index 97d10bfef..491b985da 100644
--- a/docs/_sources/how_to/deploy_models/deploy_model_on_android.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_model_on_android.rst.txt
@@ -441,7 +441,7 @@ Execute on TVM
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 15.2743 15.2643 15.4321 15.1730 0.0696
+ 15.8805 15.8494 16.1064 15.6990 0.1340
diff --git a/docs/_sources/how_to/deploy_models/deploy_object_detection_pytorch.rst.txt b/docs/_sources/how_to/deploy_models/deploy_object_detection_pytorch.rst.txt
index 4d42e1a34..6de89ca52 100644
--- a/docs/_sources/how_to/deploy_models/deploy_object_detection_pytorch.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_object_detection_pytorch.rst.txt
@@ -123,7 +123,7 @@ Load pre-trained maskrcnn from torchvision and do tracing
.. code-block:: none
Downloading: "https://download.pytorch.org/models/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth" to /workspace/.cache/torch/hub/checkpoints/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth
-
0%| | 0.00/170M [00:00<?, ?B/s]
2%|2 | 3.87M/170M [00:00<00:04, 40.4MB/s]
5%|4 | 7.73M/170M [00:00<00:04, 38.1MB/s]
7%|6 | 11.4M/170M [00:00<00:05, 33.0MB/s]
10%|9 | 16.6M/170M [00:00<00:03, 40.6MB/s]
13%|#2 | 21.5M/170M [00:00<00:03, 44.1MB/s]
15%|#5 | 25.8M/170M [00:00<00:03, 43.3MB/s]
18%|#7 | 30.0M/170M [00:00<00:03, 37.7MB/s]
21%|##1 | 36.0M/170M [00:00<00:03, 44.5MB/s]
24%|##3 | 40.4M/170M [00:01<00:03, 44.4MB/s]
26%|##6 | 44.8M/170M [00:01<00:02, 44.1MB/s]
30%|##9 | 50.1M/170M [00:01<00:02, 47.5MB/s]
32%|###2 | 55.0M/170M [00:01<00:02, 48.6MB/s]
35%|###5 | 59.7M/170M [00:01<00:02, 47.1MB/s]
38%|###8 | 64.7M/170M [00:01<00:02, 48.4MB/s]
41%|####1 | 70.3M/170M [00:01<00:02, 51.5MB/s]
44%|####4 | 75.3M/170M [00:01<00:01, 50.8MB/s]
48%|####7 | 81.0M/170M [00:01<00:01, 53.4MB/
s]
51%|##### | 86.3M/170M [00:01<00:01, 54.0MB/s]
54%|#####4 | 91.9M/170M [00:02<00:01, 55.4MB/s]
57%|#####7 | 97.3M/170M [00:02<00:01, 55.6MB/s]
60%|###### | 103M/170M [00:02<00:01, 53.8MB/s]
63%|######3 | 108M/170M [00:02<00:01, 50.0MB/s]
67%|######7 | 114M/170M [00:02<00:01, 53.7MB/s]
70%|####### | 120M/170M [00:02<00:00, 56.0MB/s]
74%|#######4 | 126M/170M [00:02<00:00, 58.1MB/s]
77%|#######7 | 131M/170M [00:02<00:00, 57.7MB/s]
81%|######## | 137M/170M [00:02<00:00, 58.8MB/s]
84%|########4 | 143M/170M [00:02<00:00, 57.3MB/s]
87%|########7 | 148M/170M [00:03<00:00, 56.8MB/s]
91%|######### | 154M/170M [00:03<00:00, 52.7MB/s]
94%|#########3| 159M/170M [00:03<00:00, 43.4MB/s]
97%|#########6| 164M/170M [00:03<00:00, 46.5MB/s]
100%|##########| 170M/170M [00:03<00:00, 49.5MB/s]
+
0%| | 0.00/170M [00:00<?, ?B/s]
2%|2 | 3.56M/170M [00:00<00:04, 36.4MB/s]
4%|4 | 7.03M/170M [00:00<00:05, 31.4MB/s]
6%|6 | 10.4M/170M [00:00<00:05, 32.6MB/s]
8%|8 | 13.6M/170M [00:00<00:05, 32.4MB/s]
10%|9 | 16.9M/170M [00:00<00:04, 32.9MB/s]
12%|#2 | 20.9M/170M [00:00<00:04, 35.9MB/s]
14%|#4 | 24.4M/170M [00:00<00:04, 35.6MB/s]
17%|#6 | 28.2M/170M [00:00<00:04, 37.1MB/s]
19%|#9 | 32.5M/170M [00:00<00:03, 39.5MB/s]
21%|##1 | 36.3M/170M [00:01<00:03, 39.6MB/s]
24%|##3 | 40.1M/170M [00:01<00:03, 38.7MB/s]
26%|##5 | 43.8M/170M [00:01<00:03, 37.0MB/s]
28%|##7 | 47.4M/170M [00:01<00:03, 35.4MB/s]
30%|##9 | 50.8M/170M [00:01<00:04, 28.0MB/s]
32%|###1 | 53.7M/170M [00:01<00:05, 24.3MB/s]
33%|###3 | 56.6M/170M [00:01<00:04, 25.8MB/s]
35%|###5 | 60.0M/170M [00:01<00:04, 27.9MB/
s]
37%|###6 | 62.8M/170M [00:02<00:04, 24.9MB/s]
39%|###8 | 66.0M/170M [00:02<00:04, 26.5MB/s]
41%|####1 | 70.2M/170M [00:02<00:03, 30.8MB/s]
43%|####3 | 73.3M/170M [00:02<00:03, 31.2MB/s]
45%|####4 | 76.4M/170M [00:02<00:03, 28.9MB/s]
47%|####6 | 79.3M/170M [00:02<00:03, 27.8MB/s]
48%|####8 | 82.0M/170M [00:02<00:03, 27.7MB/s]
50%|####9 | 84.9M/170M [00:02<00:03, 27.2MB/s]
52%|#####1 | 87.5M/170M [00:03<00:05, 17.0MB/s]
53%|#####2 | 89.6M/170M [00:03<00:04, 17.2MB/s]
54%|#####3 | 91.6M/170M [00:03<00:04, 17.0MB/s]
55%|#####5 | 93.5M/170M [00:03<00:04, 17.4MB/s]
56%|#####6 | 95.3M/170M [00:03<00:04, 17.8MB/s]
58%|#####8 | 98.5M/170M [00:03<00:03, 21.7MB/s]
60%|#####9 | 102M/170M [00:03<00:02, 24.5MB/s]
61%|######1 | 104M/170M [00:04<00:02, 23.0MB/s]
63%|######3 | 108M/170M [00:04<00:02, 27.2MB/s]
65%|######5 | 111M/170M [00:04<00:02,
28.3MB/s]
67%|######7 | 114M/170M [00:04<00:02, 28.1MB/s]
69%|######9 | 117M/170M [00:04<00:01, 29.7MB/s]
72%|#######1 | 122M/170M [00:04<00:01, 34.5MB/s]
74%|#######3 | 125M/170M [00:04<00:01, 35.8MB/s]
76%|#######6 | 130M/170M [00:04<00:01, 38.0MB/s]
79%|#######9 | 134M/170M [00:04<00:00, 42.0MB/s]
82%|########1 | 139M/170M [00:04<00:00, 44.3MB/s]
85%|########4 | 144M/170M [00:05<00:00, 42.7MB/s]
88%|########7 | 149M/170M [00:05<00:00, 46.3MB/s]
90%|######### | 153M/170M [00:05<00:00, 45.2MB/s]
93%|#########2| 158M/170M [00:05<00:00, 34.6MB/s]
95%|#########4| 161M/170M [00:05<00:00, 35.4MB/s]
97%|#########7| 165M/170M [00:05<00:00, 33.1MB/s]
99%|#########9| 168M/170M [00:05<00:00, 31.8MB/s]
100%|##########| 170M/170M [00:05<00:00, 30.4MB/s]
/usr/local/lib/python3.7/dist-packages/torch/nn/functional.py:3878: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
for i in range(dim)
/usr/local/lib/python3.7/dist-packages/torchvision/models/detection/anchor_utils.py:127: UserWarning: __floordiv__ is deprecated, and its behavior will change in a future version of pytorch. It currently rounds toward 0 (like the 'trunc' function NOT 'floor'). This results in incorrect rounding for negative values. To keep the current behavior, use torch.div(a, b, rounding_mode='trunc'), or for actual floor division, use torch.div(a, b, rounding_mode='floor').
@@ -295,7 +295,7 @@ Get boxes with score larger than 0.9
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 2 minutes 53.079 seconds)
+ **Total running time of the script:** ( 2 minutes 59.602 seconds)
.. _sphx_glr_download_how_to_deploy_models_deploy_object_detection_pytorch.py:
diff --git a/docs/_sources/how_to/deploy_models/deploy_prequantized.rst.txt b/docs/_sources/how_to/deploy_models/deploy_prequantized.rst.txt
index 7f2e3662d..4213ed462 100644
--- a/docs/_sources/how_to/deploy_models/deploy_prequantized.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_prequantized.rst.txt
@@ -232,7 +232,7 @@ training. Other models require a full post training calibration.
.. code-block:: none
Downloading: "https://download.pytorch.org/models/mobilenet_v2-b0353104.pth" to /workspace/.cache/torch/hub/checkpoints/mobilenet_v2-b0353104.pth
-
0%| | 0.00/13.6M [00:00<?, ?B/s]
34%|###4 | 4.61M/13.6M [00:00<00:00, 48.2MB/s]
68%|######7 | 9.21M/13.6M [00:00<00:00, 45.5MB/s]
100%|##########| 13.6M/13.6M [00:00<00:00, 60.0MB/s]
+
0%| | 0.00/13.6M [00:00<?, ?B/s]
18%|#8 | 2.46M/13.6M [00:00<00:00, 25.6MB/s]
42%|####2 | 5.70M/13.6M [00:00<00:00, 30.5MB/s]
100%|##########| 13.6M/13.6M [00:00<00:00, 58.4MB/s]
@@ -412,7 +412,7 @@ Here we give an example of how to measure performance of TVM compiled models.
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 88.6839 88.6130 89.7298 88.4377 0.2284
+ 90.1142 90.0253 95.4225 89.6971 0.5840
@@ -461,7 +461,7 @@ TODO
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 7.246 seconds)
+ **Total running time of the script:** ( 1 minutes 8.445 seconds)
.. _sphx_glr_download_how_to_deploy_models_deploy_prequantized.py:
diff --git a/docs/_sources/how_to/deploy_models/deploy_prequantized_tflite.rst.txt b/docs/_sources/how_to/deploy_models/deploy_prequantized_tflite.rst.txt
index c0edd81f9..7347dbc03 100644
--- a/docs/_sources/how_to/deploy_models/deploy_prequantized_tflite.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_prequantized_tflite.rst.txt
@@ -439,7 +439,7 @@ Here we give an example of how to measure performance of TVM compiled models.
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 115.3626 114.9331 121.7323 113.8598 1.2658
+ 120.0408 120.0760 124.6104 117.7518 0.8415
@@ -476,7 +476,7 @@ Here we give an example of how to measure performance of TVM compiled models.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 49.373 seconds)
+ **Total running time of the script:** ( 1 minutes 55.942 seconds)
.. _sphx_glr_download_how_to_deploy_models_deploy_prequantized_tflite.py:
diff --git a/docs/_sources/how_to/deploy_models/deploy_quantized.rst.txt b/docs/_sources/how_to/deploy_models/deploy_quantized.rst.txt
index 1d0570eb7..86cc86d2e 100644
--- a/docs/_sources/how_to/deploy_models/deploy_quantized.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_quantized.rst.txt
@@ -255,7 +255,7 @@ We create a Relay VM to build and execute the model.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 30.173 seconds)
+ **Total running time of the script:** ( 1 minutes 30.164 seconds)
.. _sphx_glr_download_how_to_deploy_models_deploy_quantized.py:
diff --git a/docs/_sources/how_to/deploy_models/deploy_ssd_gluoncv.rst.txt b/docs/_sources/how_to/deploy_models/deploy_ssd_gluoncv.rst.txt
index bea77359d..0e1530cdd 100644
--- a/docs/_sources/how_to/deploy_models/deploy_ssd_gluoncv.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_ssd_gluoncv.rst.txt
@@ -158,7 +158,7 @@ Convert and compile model for CPU.
data: None
input_sym_arg_type = in_param.infer_type()[0]
Downloading /workspace/.mxnet/models/ssd_512_resnet50_v1_voc-9c8b225a.zip from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/ssd_512_resnet50_v1_voc-9c8b225a.zip...
-
0%| | 0/132723 [00:00<?, ?KB/s]
5%|4 | 6034/132723 [00:00<00:02, 60335.17KB/s]
11%|# | 14118/132723 [00:00<00:01, 72389.67KB/s]
16%|#6 | 21357/132723 [00:00<00:01, 62265.24KB/s]
22%|##2 | 29437/132723 [00:00<00:01, 68918.80KB/s]
28%|##8 | 37516/132723 [00:00<00:01, 72944.53KB/s]
34%|###4 | 45650/132723 [00:00<00:01, 75677.02KB/s]
41%|#### | 53837/132723 [00:00<00:01, 77645.25KB/s]
47%|####6 | 62018/132723 [00:00<00:00, 78945.51KB/s]
53%|#####2 | 70209/132723 [00:00<00:00, 79856.88KB/s]
59%|#####9 | 78366/132723 [00:01<00:00, 80379.82KB/s]
65%|######5 | 86526/132723 [00:01<00:00, 80746.30KB/s]
71%|#######1 | 94639/132723 [00:01<00:00, 80861.19KB/s]
77%|#######7 | 102806/132723 [00:01<00:00, 81103.66KB/s]
84%|########3 | 111024/132723 [00:01<00:00, 81424.16KB/s]
90%|########9 | 119173/132723 [00:01<00:00, 81358.71KB/s]
96%|########
#5| 127358/132723 [00:01<00:00, 81504.18KB/s]
100%|##########| 132723/132723 [00:01<00:00, 77995.44KB/s]
+
0%| | 0/132723 [00:00<?, ?KB/s]
4%|4 | 5598/132723 [00:00<00:02, 55961.24KB/s]
10%|9 | 13009/132723 [00:00<00:01, 66632.37KB/s]
15%|#5 | 20526/132723 [00:00<00:01, 70527.47KB/s]
21%|##1 | 28024/132723 [00:00<00:01, 72283.52KB/s]
27%|##6 | 35378/132723 [00:00<00:01, 72734.74KB/s]
32%|###2 | 42898/132723 [00:00<00:01, 73569.98KB/s]
38%|###8 | 50465/132723 [00:00<00:01, 74245.79KB/s]
44%|####3 | 57943/132723 [00:00<00:01, 74408.86KB/s]
49%|####9 | 65642/132723 [00:00<00:00, 75213.35KB/s]
55%|#####5 | 73378/132723 [00:01<00:00, 75873.43KB/s]
61%|######1 | 81241/132723 [00:01<00:00, 76703.52KB/s]
67%|######7 | 89158/132723 [00:01<00:00, 77450.97KB/s]
73%|#######3 | 97020/132723 [00:01<00:00, 77803.33KB/s]
79%|#######9 | 104852/132723 [00:01<00:00, 77956.81KB/s]
85%|########4 | 112684/132723 [00:01<00:00, 78064.84KB/s]
91%|#########
| 120552/132723 [00:01<00:00, 78246.25KB/s]
97%|#########6| 128402/132723 [00:01<00:00, 78319.20KB/s]
100%|##########| 132723/132723 [00:01<00:00, 75591.62KB/s]
@@ -241,7 +241,7 @@ Display result
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 2 minutes 31.414 seconds)
+ **Total running time of the script:** ( 2 minutes 34.457 seconds)
.. _sphx_glr_download_how_to_deploy_models_deploy_ssd_gluoncv.py:
diff --git a/docs/_sources/how_to/deploy_models/sg_execution_times.rst.txt b/docs/_sources/how_to/deploy_models/sg_execution_times.rst.txt
index d8c6b4290..3ed74419d 100644
--- a/docs/_sources/how_to/deploy_models/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/deploy_models/sg_execution_times.rst.txt
@@ -5,24 +5,24 @@
Computation times
=================
-**11:02.831** total execution time for **how_to_deploy_models** files:
+**11:22.889** total execution time for **how_to_deploy_models** files:
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 02:53.079 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 02:59.602 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``) | 02:31.414 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``) | 02:34.457 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``) | 01:49.373 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``) | 01:55.942 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``) | 01:30.173 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``) | 01:30.164 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``) | 01:07.246 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``) | 01:08.445 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``) | 00:28.854 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``) | 00:30.138 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``) | 00:21.467 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``) | 00:22.342 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``) | 00:21.220 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``) | 00:21.792 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_deploy_models_deploy_sparse.py` (``deploy_sparse.py``) | 00:00.006 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/extend_tvm/bring_your_own_datatypes.rst.txt b/docs/_sources/how_to/extend_tvm/bring_your_own_datatypes.rst.txt
index 6034e6e5e..65f4a0d7f 100644
--- a/docs/_sources/how_to/extend_tvm/bring_your_own_datatypes.rst.txt
+++ b/docs/_sources/how_to/extend_tvm/bring_your_own_datatypes.rst.txt
@@ -476,7 +476,7 @@ First let us define two helper functions to get the mobilenet model and a cat im
.. code-block:: none
- Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zip1efdd2e7-3622-4514-a4ad-5a86bdf60c5b from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
+ Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zip93f8e8a8-e264-497e-8f53-3afeffa873ab from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
diff --git a/docs/_sources/how_to/extend_tvm/sg_execution_times.rst.txt b/docs/_sources/how_to/extend_tvm/sg_execution_times.rst.txt
index a2f863ffe..22779fbc6 100644
--- a/docs/_sources/how_to/extend_tvm/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/extend_tvm/sg_execution_times.rst.txt
@@ -5,14 +5,14 @@
Computation times
=================
-**00:40.045** total execution time for **how_to_extend_tvm** files:
+**00:40.364** total execution time for **how_to_extend_tvm** files:
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_bring_your_own_datatypes.py` (``bring_your_own_datatypes.py``) | 00:37.056 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_bring_your_own_datatypes.py` (``bring_your_own_datatypes.py``) | 00:37.285 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``) | 00:02.104 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``) | 00:02.140 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_infra.py` (``use_pass_infra.py``) | 00:00.877 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_infra.py` (``use_pass_infra.py``) | 00:00.932 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_low_level_custom_pass.py` (``low_level_custom_pass.py``) | 00:00.008 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_low_level_custom_pass.py` (``low_level_custom_pass.py``) | 00:00.007 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/extend_tvm/use_pass_instrument.rst.txt b/docs/_sources/how_to/extend_tvm/use_pass_instrument.rst.txt
index d24cc5fb5..d837b5ea7 100644
--- a/docs/_sources/how_to/extend_tvm/use_pass_instrument.rst.txt
+++ b/docs/_sources/how_to/extend_tvm/use_pass_instrument.rst.txt
@@ -216,10 +216,10 @@ profile the execution time of each passes.
.. code-block:: none
Printing results of timing profile...
- InferType: 6544us [6544us] (45.82%; 45.82%)
- FoldScaleAxis: 7737us [5us] (54.18%; 54.18%)
- FoldConstant: 7732us [1600us] (54.14%; 99.94%)
- InferType: 6132us [6132us] (42.94%; 79.31%)
+ InferType: 6623us [6623us] (45.78%; 45.78%)
+ FoldScaleAxis: 7842us [5us] (54.22%; 54.22%)
+ FoldConstant: 7837us [1623us] (54.18%; 99.93%)
+ InferType: 6213us [6213us] (42.96%; 79.29%)
@@ -258,10 +258,10 @@ Refer to following sections and :py:func:`tvm.instrument.pass_instrument` for th
.. code-block:: none
Printing results of timing profile...
- InferType: 6253us [6253us] (44.97%; 44.97%)
- FoldScaleAxis: 7652us [4us] (55.03%; 55.03%)
- FoldConstant: 7647us [1572us] (55.00%; 99.94%)
- InferType: 6076us [6076us] (43.70%; 79.45%)
+ InferType: 6269us [6269us] (44.39%; 44.39%)
+ FoldScaleAxis: 7855us [4us] (55.61%; 55.61%)
+ FoldConstant: 7850us [1648us] (55.58%; 99.94%)
+ InferType: 6202us [6202us] (43.91%; 79.00%)
diff --git a/docs/_sources/how_to/optimize_operators/opt_conv_cuda.rst.txt b/docs/_sources/how_to/optimize_operators/opt_conv_cuda.rst.txt
index 9c50a8c26..57540d271 100644
--- a/docs/_sources/how_to/optimize_operators/opt_conv_cuda.rst.txt
+++ b/docs/_sources/how_to/optimize_operators/opt_conv_cuda.rst.txt
@@ -340,7 +340,7 @@ latency of convolution.
.. code-block:: none
- Convolution: 44.051947 ms
+ Convolution: 54.123335 ms
diff --git a/docs/_sources/how_to/optimize_operators/opt_conv_tensorcore.rst.txt b/docs/_sources/how_to/optimize_operators/opt_conv_tensorcore.rst.txt
index f068cf885..9b91e44b9 100644
--- a/docs/_sources/how_to/optimize_operators/opt_conv_tensorcore.rst.txt
+++ b/docs/_sources/how_to/optimize_operators/opt_conv_tensorcore.rst.txt
@@ -671,7 +671,7 @@ be able to run on our build server
.. code-block:: none
- conv2d with tensor core: 10.977927 ms
+ conv2d with tensor core: 7.243840 ms
diff --git a/docs/_sources/how_to/optimize_operators/opt_gemm.rst.txt b/docs/_sources/how_to/optimize_operators/opt_gemm.rst.txt
index a640df506..21391167f 100644
--- a/docs/_sources/how_to/optimize_operators/opt_gemm.rst.txt
+++ b/docs/_sources/how_to/optimize_operators/opt_gemm.rst.txt
@@ -143,8 +143,8 @@ Then we write a baseline implementation, the simplest way to write a matrix mult
.. code-block:: none
- Numpy running time: 0.017506
- Baseline: 3.333864
+ Numpy running time: 0.017938
+ Baseline: 3.440580
@@ -239,7 +239,7 @@ fill 32 * 32 * sizeof(float) which is 4KB in the cache whose total size is 32KB
.. code-block:: none
- Opt1: 0.294581
+ Opt1: 0.293736
@@ -342,7 +342,7 @@ In this tutorial, we chose to vectorize the inner loop row data since it is cach
.. code-block:: none
- Opt2: 0.328113
+ Opt2: 0.328085
@@ -438,7 +438,7 @@ the access pattern for A matrix is more cache friendly.
.. code-block:: none
- Opt3: 0.111140
+ Opt3: 0.113581
@@ -563,7 +563,7 @@ flattening.
.. code-block:: none
- Opt4: 0.107377
+ Opt4: 0.109142
@@ -685,7 +685,7 @@ write to C when all the block results are ready.
.. code-block:: none
- Opt5: 0.108128
+ Opt5: 0.111236
@@ -810,7 +810,7 @@ Furthermore, we can also utilize multi-core processors to do the thread-level pa
.. code-block:: none
- Opt6: 0.144638
+ Opt6: 0.146516
diff --git a/docs/_sources/how_to/optimize_operators/sg_execution_times.rst.txt b/docs/_sources/how_to/optimize_operators/sg_execution_times.rst.txt
index b0ca969e6..2764d877e 100644
--- a/docs/_sources/how_to/optimize_operators/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/optimize_operators/sg_execution_times.rst.txt
@@ -5,12 +5,12 @@
Computation times
=================
-**00:33.822** total execution time for **how_to_optimize_operators** files:
+**00:34.483** total execution time for **how_to_optimize_operators** files:
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``) | 00:31.449 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``) | 00:32.049 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.288 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.330 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``) | 00:01.086 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``) | 00:01.104 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/tune_with_autoscheduler/sg_execution_times.rst.txt b/docs/_sources/how_to/tune_with_autoscheduler/sg_execution_times.rst.txt
index cd42583bf..b07e5b777 100644
--- a/docs/_sources/how_to/tune_with_autoscheduler/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/tune_with_autoscheduler/sg_execution_times.rst.txt
@@ -5,18 +5,18 @@
Computation times
=================
-**05:58.933** total execution time for **how_to_tune_with_autoscheduler** files:
+**06:10.469** total execution time for **how_to_tune_with_autoscheduler** files:
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_conv2d_layer_cuda.py` (``tune_conv2d_layer_cuda.py``) | 03:17.057 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_conv2d_layer_cuda.py` (``tune_conv2d_layer_cuda.py``) | 03:25.235 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``) | 01:20.723 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``) | 01:22.299 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``) | 00:45.656 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``) | 00:46.433 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``) | 00:18.498 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``) | 00:19.276 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``) | 00:08.611 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``) | 00:08.687 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``) | 00:08.388 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``) | 00:08.539 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/tune_with_autoscheduler/tune_conv2d_layer_cuda.rst.txt b/docs/_sources/how_to/tune_with_autoscheduler/tune_conv2d_layer_cuda.rst.txt
index a40e255ca..d7384361c 100644
--- a/docs/_sources/how_to/tune_with_autoscheduler/tune_conv2d_layer_cuda.rst.txt
+++ b/docs/_sources/how_to/tune_with_autoscheduler/tune_conv2d_layer_cuda.rst.txt
@@ -242,53 +242,888 @@ cooperative fetching, unrolling and operator fusion.
preflattened_buffer_map = {data_1: data_3: Buffer(data_2, float32, [1, 512, 7, 7], []), kernel_1: kernel_3: Buffer(kernel_2, float32, [512, 512, 3, 3], []), bias_1: bias_3: Buffer(bias_2, float32, [1, 512, 1, 1], []), compute_1: compute_3: Buffer(compute_2, float32, [1, 512, 7, 7], [])} {
attr [IterVar(blockIdx.x: int32, (nullptr), "ThreadIndex", "blockIdx.x")] "thread_extent" = 16;
allocate(conv2d_nchw: Pointer(local float32), float32, [8]), storage_scope = local;
- allocate(pad_temp.shared: Pointer(shared float32), float32, [1008]), storage_scope = shared;
- allocate(kernel.shared: Pointer(shared float32), float32, [1536]), storage_scope = shared;
+ allocate(pad_temp.shared: Pointer(shared float32), float32, [1568]), storage_scope = shared;
+ allocate(kernel.shared: Pointer(shared float32), float32, [1024]), storage_scope = shared;
attr [IterVar(threadIdx.x: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196 {
- for (ff.outer.inner.init: int32, 0, 2) {
- for (ff.inner.init: int32, 0, 2) {
- let cse_var_1: int32 = ((ff.outer.inner.init*2) + ff.inner.init)
- {
- conv2d_nchw_1: Buffer(conv2d_nchw, float32, [16], [], scope="local", align=16)[cse_var_1] = 0f32
- conv2d_nchw_1[(cse_var_1 + 4)] = 0f32
- }
- }
- }
- for (rc.outer.outer: int32, 0, 32) {
+ conv2d_nchw_1: Buffer(conv2d_nchw, float32, [4], [], scope="local", align=8)[0] = 0f32
+ conv2d_nchw_1[2] = 0f32
+ conv2d_nchw_1[4] = 0f32
+ conv2d_nchw_1[6] = 0f32
+ conv2d_nchw_1[1] = 0f32
+ conv2d_nchw_1[3] = 0f32
+ conv2d_nchw_1[5] = 0f32
+ conv2d_nchw_1[7] = 0f32
+ for (rc.outer.outer: int32, 0, 16) {
for (ry.outer.outer: int32, 0, 3) {
- for (ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer: int32, 0, 6) {
+ let cse_var_2: int32 = (rc.outer.outer*288)
+ let cse_var_1: int32 = (ry.outer.outer*3)
+ {
attr [IterVar(threadIdx.x_1: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
- if @tir.likely((((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer*7) + floordiv(threadIdx.x_1, 28)) < 36), dtype=bool) {
- pad_temp.shared_1: Buffer(pad_temp.shared, float32, [1008], [], scope="shared")[((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer*196) + threadIdx.x_1)] = @tir.if_then_else(((((1 <= (floordiv(floormod(((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer*7) + threadIdx.x_1), 63), 9) + ry.outer.outer)) && ((floordiv(floormod(((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer*7) + threadIdx.x_1), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod(((ax0.ax1.fused.ax2.fused.ax3.fused.out [...]
- }
- }
- for (ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1: int32, 0, 8) {
+ pad_temp.shared_1: Buffer(pad_temp.shared, float32, [1568], [], scope="shared")[threadIdx.x_1] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 196)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 188)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 392)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 384)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 588)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 580)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 784)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 776)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 980)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 972)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1176)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1168)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1372)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1364)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_2: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
- if @tir.likely((((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1*49) + floordiv(threadIdx.x_2, 4)) < 384), dtype=bool) {
- kernel.shared_1: Buffer(kernel.shared, float32, [1536], [], scope="shared")[((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1*196) + threadIdx.x_2)] = kernel[((((((blockIdx.x*147456) + (floordiv(((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1*49) + floordiv(threadIdx.x_2, 4)), 12)*4608)) + (rc.outer.outer*144)) + (floordiv(floormod(((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1*4) + threadIdx.x_2), 48), 3)*9)) + (ry.outer.outer*3)) + floormod((threadIdx.x_2 + ax0.ax1 [...]
+ kernel.shared_1: Buffer(kernel.shared, float32, [1024], [], scope="shared")[threadIdx.x_2] = kernel[(((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 32)*4608)) + cse_var_2) + (floormod(threadIdx.x_2, 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 196)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 196), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 4), 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 392)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 392), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 8), 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 588)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 588), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 12), 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 784)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 784), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 16), 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ if @tir.likely((threadIdx.x_2 < 44), dtype=bool) {
+ kernel.shared_1[(threadIdx.x_2 + 980)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 980), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 20), 32)*9)) + cse_var_1)]
}
- }
- for (rc.outer.inner: int32, 0, 2) {
- for (ff.outer.inner: int32, 0, 2) {
- for (rc.inner: int32, 0, 8) {
- for (rx.inner: int32, 0, 3) {
- for (ff.inner: int32, 0, 2) {
- let cse_var_3: int32 = ((ff.outer.inner*2) + ff.inner)
- let cse_var_2: int32 = (cse_var_3 + 4)
- {
- conv2d_nchw_1[cse_var_3] = (conv2d_nchw_1[cse_var_3] + (pad_temp.shared_1[(((((rc.outer.inner*504) + (rc.inner*63)) + (floordiv(floormod(threadIdx.x, 49), 7)*9)) + rx.inner) + floormod(threadIdx.x, 7))]*kernel.shared_1[((((((floordiv(threadIdx.x, 49)*192) + (ff.outer.inner*96)) + (ff.inner*48)) + (rc.outer.inner*24)) + (rc.inner*3)) + rx.inner)]))
- conv2d_nchw_1[cse_var_2] = (conv2d_nchw_1[cse_var_2] + (pad_temp.shared_1[(((((rc.outer.inner*504) + (rc.inner*63)) + (floordiv(floormod(threadIdx.x, 49), 7)*9)) + rx.inner) + floormod(threadIdx.x, 7))]*kernel.shared_1[(((((((floordiv(threadIdx.x, 49)*192) + (ff.outer.inner*96)) + (ff.inner*48)) + (rc.outer.inner*24)) + (rc.inner*3)) + rx.inner) + 768)]))
- }
- }
- }
- }
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[(floordiv(threadIdx.x, 49)*64)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 256)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 512)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 768)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 32)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 288)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 544)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 800)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 1)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 257)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 513)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 769)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 33)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 289)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 545)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 801)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 2)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 258)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 514)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 770)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 34)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 290)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 546)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 802)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 3)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 259)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 515)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 771)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 35)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 291)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 547)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 803)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 4)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 260)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 516)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 772)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 36)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 292)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 548)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 804)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 5)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 261)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 517)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 773)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 37)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 293)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 549)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 805)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 6)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 262)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 518)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 774)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 38)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 294)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 550)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 806)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 7)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 263)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 519)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 775)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 39)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 295)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 551)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 807)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 8)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 264)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 520)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 776)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 40)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 296)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 552)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 808)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 9)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 265)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 521)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 777)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 41)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 297)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 553)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 809)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 10)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 266)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 522)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 778)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 42)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 298)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 554)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 810)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 11)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 267)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 523)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 779)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 43)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 299)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 555)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 811)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 12)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 268)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 524)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 780)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 44)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 300)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 556)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 812)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 13)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 269)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 525)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 781)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 45)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 301)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 557)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 813)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 14)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 270)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 526)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 782)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 46)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 302)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 558)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 814)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 15)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 271)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 527)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 783)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 47)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 303)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 559)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 815)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 16)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 272)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 528)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 784)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 48)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 304)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 560)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 816)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 17)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 273)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 529)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 785)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 49)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 305)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 561)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 817)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 18)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 274)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 530)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 786)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 50)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 306)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 562)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 818)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 19)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 275)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 531)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 787)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 51)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 307)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 563)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 819)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 20)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 276)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 532)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 788)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 52)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 308)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 564)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 820)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 21)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 277)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 533)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 789)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 53)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 309)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 565)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 821)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 22)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 278)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 534)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 790)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 54)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 310)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 566)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 822)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 23)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 279)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 535)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 791)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 55)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 311)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 567)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 823)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 24)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 280)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 536)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 792)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 56)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 312)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 568)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 824)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 25)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 281)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 537)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 793)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 57)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 313)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 569)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 825)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 26)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 282)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 538)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 794)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 58)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 314)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 570)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 826)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 27)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 283)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 539)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 795)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 59)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 315)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 571)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 827)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 28)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 284)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 540)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 796)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 60)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 316)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 572)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 828)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 29)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 285)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 541)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 797)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 61)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 317)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 573)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 829)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 30)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 286)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 542)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 798)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 62)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 318)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 574)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 830)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 31)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 287)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 543)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 799)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 63)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 319)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 575)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 831)]))
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[threadIdx.x_1] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) - 7)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 196)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 189)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 392)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 385)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 588)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 581)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 784)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 777)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 980)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 973)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1176)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1169)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1372)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1365)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[threadIdx.x_2] = kernel[((((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 32)*4608)) + cse_var_2) + (floormod(threadIdx.x_2, 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 196)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 196), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 4), 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 392)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 392), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 8), 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 588)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 588), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 12), 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 784)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 784), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 16), 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ if @tir.likely((threadIdx.x_2 < 44), dtype=bool) {
+ kernel.shared_1[(threadIdx.x_2 + 980)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 980), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 20), 32)*9)) + cse_var_1) + 1)]
}
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[(floordiv(threadIdx.x, 49)*64)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 256)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 512)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 768)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 32)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 288)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 544)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 800)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 1)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 257)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 513)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 769)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 33)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 289)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 545)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 801)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 2)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 258)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 514)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 770)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 34)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 290)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 546)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 802)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 3)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 259)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 515)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 771)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 35)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 291)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 547)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 803)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 4)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 260)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 516)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 772)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 36)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 292)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 548)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 804)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 5)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 261)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 517)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 773)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 37)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 293)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 549)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 805)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 6)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 262)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 518)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 774)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 38)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 294)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 550)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 806)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 7)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 263)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 519)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 775)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 39)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 295)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 551)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 807)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 8)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 264)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 520)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 776)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 40)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 296)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 552)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 808)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 9)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 265)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 521)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 777)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 41)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 297)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 553)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 809)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 10)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 266)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 522)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 778)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 42)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 298)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 554)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 810)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 11)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 267)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 523)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 779)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 43)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 299)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 555)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 811)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 12)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 268)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 524)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 780)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 44)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 300)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 556)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 812)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 13)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 269)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 525)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 781)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 45)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 301)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 557)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 813)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 14)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 270)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 526)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 782)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 46)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 302)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 558)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 814)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 15)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 271)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 527)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 783)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 47)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 303)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 559)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 815)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 16)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 272)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 528)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 784)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 48)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 304)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 560)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 816)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 17)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 273)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 529)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 785)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 49)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 305)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 561)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 817)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 18)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 274)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 530)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 786)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 50)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 306)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 562)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 818)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 19)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 275)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 531)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 787)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 51)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 307)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 563)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 819)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 20)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 276)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 532)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 788)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 52)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 308)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 564)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 820)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 21)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 277)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 533)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 789)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 53)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 309)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 565)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 821)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 22)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 278)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 534)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 790)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 54)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 310)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 566)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 822)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 23)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 279)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 535)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 791)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 55)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 311)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 567)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 823)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 24)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 280)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 536)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 792)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 56)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 312)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 568)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 824)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 25)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 281)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 537)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 793)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 57)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 313)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 569)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 825)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 26)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 282)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 538)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 794)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 58)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 314)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 570)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 826)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 27)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 283)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 539)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 795)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 59)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 315)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 571)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 827)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 28)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 284)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 540)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 796)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 60)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 316)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 572)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 828)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 29)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 285)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 541)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 797)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 61)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 317)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 573)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 829)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 30)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 286)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 542)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 798)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 62)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 318)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 574)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 830)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 31)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 287)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 543)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 799)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 63)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 319)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 575)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 831)]))
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[threadIdx.x_1] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) - 6)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 196)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 190)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 392)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 386)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 588)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 582)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 784)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 778)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 980)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 974)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1176)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1170)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1372)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1366)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[threadIdx.x_2] = kernel[((((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 32)*4608)) + cse_var_2) + (floormod(threadIdx.x_2, 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 196)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 196), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 4), 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 392)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 392), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 8), 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 588)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 588), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 12), 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 784)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 784), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 16), 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ if @tir.likely((threadIdx.x_2 < 44), dtype=bool) {
+ kernel.shared_1[(threadIdx.x_2 + 980)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 980), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 20), 32)*9)) + cse_var_1) + 2)]
+ }
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[(floordiv(threadIdx.x, 49)*64)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 256)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 512)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 768)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 32)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 288)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 544)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 800)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 1)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 257)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 513)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 769)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 33)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 289)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 545)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 801)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 2)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 258)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 514)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 770)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 34)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 290)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 546)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 802)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 3)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 259)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 515)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 771)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 35)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 291)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 547)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 803)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 4)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 260)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 516)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 772)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 36)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 292)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 548)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 804)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 5)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 261)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 517)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 773)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 37)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 293)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 549)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 805)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 6)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 262)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 518)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 774)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 38)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 294)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 550)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 806)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 7)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 263)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 519)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 775)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 39)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 295)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 551)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 807)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 8)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 264)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 520)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 776)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 40)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 296)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 552)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 808)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 9)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 265)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 521)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 777)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 41)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 297)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 553)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 809)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 10)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 266)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 522)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 778)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 42)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 298)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 554)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 810)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 11)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 267)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 523)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 779)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 43)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 299)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 555)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 811)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 12)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 268)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 524)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 780)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 44)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 300)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 556)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 812)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 13)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 269)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 525)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 781)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 45)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 301)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 557)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 813)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 14)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 270)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 526)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 782)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 46)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 302)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 558)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 814)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 15)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 271)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 527)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 783)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 47)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 303)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 559)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 815)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 16)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 272)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 528)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 784)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 48)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 304)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 560)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 816)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 17)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 273)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 529)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 785)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 49)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 305)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 561)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 817)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 18)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 274)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 530)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 786)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 50)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 306)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 562)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 818)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 19)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 275)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 531)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 787)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 51)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 307)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 563)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 819)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 20)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 276)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 532)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 788)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 52)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 308)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 564)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 820)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 21)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 277)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 533)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 789)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 53)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 309)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 565)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 821)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 22)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 278)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 534)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 790)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 54)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 310)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 566)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 822)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 23)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 279)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 535)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 791)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 55)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 311)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 567)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 823)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 24)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 280)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 536)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 792)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 56)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 312)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 568)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 824)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 25)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 281)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 537)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 793)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 57)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 313)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 569)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 825)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 26)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 282)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 538)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 794)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 58)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 314)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 570)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 826)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 27)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 283)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 539)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 795)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 59)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 315)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 571)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 827)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 28)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 284)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 540)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 796)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 60)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 316)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 572)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 828)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 29)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 285)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 541)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 797)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 61)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 317)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 573)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 829)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 30)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 286)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 542)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 798)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 62)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 318)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 574)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 830)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 31)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 287)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 543)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 799)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 63)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 319)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 575)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 831)]))
}
}
}
- for (i1.inner: int32, 0, 4) {
- compute[((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*196)) + (i1.inner*49)) + floormod(threadIdx.x, 49))] = max((conv2d_nchw_1[i1.inner] + bias[(((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*4)) + i1.inner)]), 0f32)
- compute[(((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*196)) + (i1.inner*49)) + floormod(threadIdx.x, 49)) + 784)] = max((conv2d_nchw_1[(i1.inner + 4)] + bias[((((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*4)) + i1.inner) + 16)]), 0f32)
+ for (i1.inner: int32, 0, 2) {
+ compute[((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*98)) + (i1.inner*49)) + floormod(threadIdx.x, 49))] = max((conv2d_nchw_1[i1.inner] + bias[(((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*2)) + i1.inner)]), 0f32)
+ compute[(((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*98)) + (i1.inner*49)) + floormod(threadIdx.x, 49)) + 392)] = max((conv2d_nchw_1[(i1.inner + 2)] + bias[((((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*2)) + i1.inner) + 8)]), 0f32)
+ compute[(((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*98)) + (i1.inner*49)) + floormod(threadIdx.x, 49)) + 784)] = max((conv2d_nchw_1[(i1.inner + 4)] + bias[((((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*2)) + i1.inner) + 16)]), 0f32)
+ compute[(((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*98)) + (i1.inner*49)) + floormod(threadIdx.x, 49)) + 1176)] = max((conv2d_nchw_1[(i1.inner + 6)] + bias[((((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*2)) + i1.inner) + 24)]), 0f32)
}
}
}
@@ -343,7 +1178,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 0.366 ms
+ Execution time of this operator: 0.243 ms
@@ -391,10 +1226,10 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
conv2d_nchw_nn_o_o_i, conv2d_nchw_nn_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_i, factor=1)
conv2d_nchw_nn_o_o_o_i, conv2d_nchw_nn_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_i, factor=1)
conv2d_nchw_nn_o_o_o_o, conv2d_nchw_nn_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_o_i, factor=1)
- conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=2)
+ conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=1)
conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=2)
conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=4)
- conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=2)
+ conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=4)
conv2d_nchw_yy_o_i, conv2d_nchw_yy_i = s[conv2d_nchw].split(conv2d_nchw_yy, factor=1)
conv2d_nchw_yy_o_o_i, conv2d_nchw_yy_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_i, factor=1)
conv2d_nchw_yy_o_o_o_i, conv2d_nchw_yy_o_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_o_i, factor=7)
@@ -403,19 +1238,19 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
conv2d_nchw_xx_o_o_i, conv2d_nchw_xx_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_i, factor=1)
conv2d_nchw_xx_o_o_o_i, conv2d_nchw_xx_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_i, factor=7)
conv2d_nchw_xx_o_o_o_o, conv2d_nchw_xx_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_o_i, factor=1)
- conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=8)
- conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=2)
+ conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=1)
+ conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=32)
conv2d_nchw_ry_o_i, conv2d_nchw_ry_i = s[conv2d_nchw].split(conv2d_nchw_ry, factor=1)
conv2d_nchw_ry_o_o, conv2d_nchw_ry_o_i = s[conv2d_nchw].split(conv2d_nchw_ry_o_i, factor=1)
- conv2d_nchw_rx_o_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=3)
+ conv2d_nchw_rx_o_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=1)
conv2d_nchw_rx_o_o, conv2d_nchw_rx_o_i = s[conv2d_nchw].split(conv2d_nchw_rx_o_i, factor=1)
s[conv2d_nchw].reorder(conv2d_nchw_nn_o_o_o_o, conv2d_nchw_ff_o_o_o_o, conv2d_nchw_yy_o_o_o_o, conv2d_nchw_xx_o_o_o_o, conv2d_nchw_nn_o_o_o_i, conv2d_nchw_ff_o_o_o_i, conv2d_nchw_yy_o_o_o_i, conv2d_nchw_xx_o_o_o_i, conv2d_nchw_nn_o_o_i, conv2d_nchw_ff_o_o_i, conv2d_nchw_yy_o_o_i, conv2d_nchw_xx_o_o_i, conv2d_nchw_rc_o_o, conv2d_nchw_ry_o_o, conv2d_nchw_rx_o_o, conv2d_nchw_rc_o_i, conv2d_nchw_ry_o_i, conv2d_nchw_rx_o_i, conv2d_nchw_nn_o_i, conv2d_nchw_ff_o_i, conv2d_nchw_yy_o_i, conv2 [...]
compute_i0_o_i, compute_i0_i = s[compute].split(compute_i0, factor=1)
compute_i0_o_o_i, compute_i0_o_i = s[compute].split(compute_i0_o_i, factor=1)
compute_i0_o_o_o, compute_i0_o_o_i = s[compute].split(compute_i0_o_o_i, factor=1)
- compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=4)
+ compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=2)
compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=4)
- compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=2)
+ compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=4)
compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=7)
compute_i2_o_o_o, compute_i2_o_o_i = s[compute].split(compute_i2_o_o_i, factor=1)
@@ -447,7 +1282,7 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
s[pad_temp_shared].vectorize(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=196)
s[pad_temp_shared].bind(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
- s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 0)
+ s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 1024)
s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
@@ -467,45 +1302,847 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
#endif
extern "C" __global__ void __launch_bounds__(196) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
float conv2d_nchw[8];
- __shared__ float pad_temp_shared[1008];
- __shared__ float kernel_shared[1536];
- for (int ff_outer_inner_init = 0; ff_outer_inner_init < 2; ++ff_outer_inner_init) {
- for (int ff_inner_init = 0; ff_inner_init < 2; ++ff_inner_init) {
- conv2d_nchw[((ff_outer_inner_init * 2) + ff_inner_init)] = 0.000000e+00f;
- conv2d_nchw[(((ff_outer_inner_init * 2) + ff_inner_init) + 4)] = 0.000000e+00f;
- }
- }
- for (int rc_outer_outer = 0; rc_outer_outer < 32; ++rc_outer_outer) {
+ __shared__ float pad_temp_shared[1568];
+ __shared__ float kernel_shared[1024];
+ conv2d_nchw[0] = 0.000000e+00f;
+ conv2d_nchw[2] = 0.000000e+00f;
+ conv2d_nchw[4] = 0.000000e+00f;
+ conv2d_nchw[6] = 0.000000e+00f;
+ conv2d_nchw[1] = 0.000000e+00f;
+ conv2d_nchw[3] = 0.000000e+00f;
+ conv2d_nchw[5] = 0.000000e+00f;
+ conv2d_nchw[7] = 0.000000e+00f;
+ for (int rc_outer_outer = 0; rc_outer_outer < 16; ++rc_outer_outer) {
for (int ry_outer_outer = 0; ry_outer_outer < 3; ++ry_outer_outer) {
__syncthreads();
- for (int ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer = 0; ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer < 6; ++ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer) {
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) / 28)) < 36) {
- pad_temp_shared[((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 196) + ((int)threadIdx.x))] = (((((1 <= (((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + ((int)threadIdx.x)) % 63) / 9) + ry_outer_outer)) && ((((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + ((int)threadIdx.x)) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + ((int)threadIdx.x)) % 9))) && ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * [...]
- }
+ pad_temp_shared[((int)threadIdx.x)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 196)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 188)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 392)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 384)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 588)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 580)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 776)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 980)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 972)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1176)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1168)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1372)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1364)] : 0.000000e+00f);
+ kernel_shared[((int)threadIdx.x)] = kernel[(((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) >> 5) * 4608)) + (rc_outer_outer * 288)) + ((((int)threadIdx.x) & 31) * 9)) + (ry_outer_outer * 3))];
+ kernel_shared[(((int)threadIdx.x) + 196)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 196) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 4) & 31) * 9)) + (ry_outer_outer * 3))];
+ kernel_shared[(((int)threadIdx.x) + 392)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 392) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 8) & 31) * 9)) + (ry_outer_outer * 3))];
+ kernel_shared[(((int)threadIdx.x) + 588)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 588) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 12) & 31) * 9)) + (ry_outer_outer * 3))];
+ kernel_shared[(((int)threadIdx.x) + 784)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 784) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 16) & 31) * 9)) + (ry_outer_outer * 3))];
+ if (((int)threadIdx.x) < 44) {
+ kernel_shared[(((int)threadIdx.x) + 980)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 980) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 20) & 31) * 9)) + (ry_outer_outer * 3))];
}
- for (int ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 = 0; ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 < 8; ++ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1) {
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 * 49) + (((int)threadIdx.x) >> 2)) < 384) {
- kernel_shared[((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 * 196) + ((int)threadIdx.x))] = kernel[((((((((int)blockIdx.x) * 147456) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 * 49) + (((int)threadIdx.x) >> 2)) / 12) * 4608)) + (rc_outer_outer * 144)) + (((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 * 4) + ((int)threadIdx.x)) % 48) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1) % 3))];
- }
+ __syncthreads();
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[((((int)threadIdx.x) / 49) * 64)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 256)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 512)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 768)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 32)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 288)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 544)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 800)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 1)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 257)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 513)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 769)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 33)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 289)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 545)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 801)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 2)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 258)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 514)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 770)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 34)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 290)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 546)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 802)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 3)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 259)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 515)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 771)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 35)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 291)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 547)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 803)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 4)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 260)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 516)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 772)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 36)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 292)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 548)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 804)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 5)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 261)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 517)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 773)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 37)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 293)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 549)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 805)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 6)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 262)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 518)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 774)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 38)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 294)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 550)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 806)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 7)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 263)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 519)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 775)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 39)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 295)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 551)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 807)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 8)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 264)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 520)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 776)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 40)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 296)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 552)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 808)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 9)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 265)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 521)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 777)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 41)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 297)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 553)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 809)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 10)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 266)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 522)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 778)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 42)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 298)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 554)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 810)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 11)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 267)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 523)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 779)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 43)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 299)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 555)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 811)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 12)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 268)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 524)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 780)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 44)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 300)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 556)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 812)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 13)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 269)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 525)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 781)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 45)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 301)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 557)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 813)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 14)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 270)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 526)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 782)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 46)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 302)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 558)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 814)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 15)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 271)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 527)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 783)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 47)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 303)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 559)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 815)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 16)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 272)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 528)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 784)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 48)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 304)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 560)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 816)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 17)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 273)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 529)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 785)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 49)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 305)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 561)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 817)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 18)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 274)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 530)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 786)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 50)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 306)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 562)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 818)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 19)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 275)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 531)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 787)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 51)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 307)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 563)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 819)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 20)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 276)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 532)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 788)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 52)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 308)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 564)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 820)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 21)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 277)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 533)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 789)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 53)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 309)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 565)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 821)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 22)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 278)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 534)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 790)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 54)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 310)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 566)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 822)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 23)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 279)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 535)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 791)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 55)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 311)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 567)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 823)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 24)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 280)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 536)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 792)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 56)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 312)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 568)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 824)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 25)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 281)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 537)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 793)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 57)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 313)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 569)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 825)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 26)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 282)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 538)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 794)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 58)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 314)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 570)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 826)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 27)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 283)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 539)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 795)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 59)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 315)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 571)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 827)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 28)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 284)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 540)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 796)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 60)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 316)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 572)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 828)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 29)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 285)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 541)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 797)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 61)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 317)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 573)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 829)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 30)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 286)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 542)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 798)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 62)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 318)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 574)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 830)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 31)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 287)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 543)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 799)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 63)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 319)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 575)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 831)]));
+ __syncthreads();
+ pad_temp_shared[((int)threadIdx.x)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) - 7)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 196)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 189)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 392)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 385)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 588)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 581)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 777)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 980)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 973)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1176)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1169)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1372)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1365)] : 0.000000e+00f);
+ kernel_shared[((int)threadIdx.x)] = kernel[((((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) >> 5) * 4608)) + (rc_outer_outer * 288)) + ((((int)threadIdx.x) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ kernel_shared[(((int)threadIdx.x) + 196)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 196) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 4) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ kernel_shared[(((int)threadIdx.x) + 392)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 392) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 8) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ kernel_shared[(((int)threadIdx.x) + 588)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 588) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 12) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ kernel_shared[(((int)threadIdx.x) + 784)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 784) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 16) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ if (((int)threadIdx.x) < 44) {
+ kernel_shared[(((int)threadIdx.x) + 980)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 980) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 20) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
}
__syncthreads();
- for (int rc_outer_inner = 0; rc_outer_inner < 2; ++rc_outer_inner) {
- for (int ff_outer_inner = 0; ff_outer_inner < 2; ++ff_outer_inner) {
- for (int rc_inner = 0; rc_inner < 8; ++rc_inner) {
- for (int rx_inner = 0; rx_inner < 3; ++rx_inner) {
- for (int ff_inner = 0; ff_inner < 2; ++ff_inner) {
- conv2d_nchw[((ff_outer_inner * 2) + ff_inner)] = (conv2d_nchw[((ff_outer_inner * 2) + ff_inner)] + (pad_temp_shared[(((((rc_outer_inner * 504) + (rc_inner * 63)) + (((((int)threadIdx.x) % 49) / 7) * 9)) + rx_inner) + (((int)threadIdx.x) % 7))] * kernel_shared[(((((((((int)threadIdx.x) / 49) * 192) + (ff_outer_inner * 96)) + (ff_inner * 48)) + (rc_outer_inner * 24)) + (rc_inner * 3)) + rx_inner)]));
- conv2d_nchw[(((ff_outer_inner * 2) + ff_inner) + 4)] = (conv2d_nchw[(((ff_outer_inner * 2) + ff_inner) + 4)] + (pad_temp_shared[(((((rc_outer_inner * 504) + (rc_inner * 63)) + (((((int)threadIdx.x) % 49) / 7) * 9)) + rx_inner) + (((int)threadIdx.x) % 7))] * kernel_shared[((((((((((int)threadIdx.x) / 49) * 192) + (ff_outer_inner * 96)) + (ff_inner * 48)) + (rc_outer_inner * 24)) + (rc_inner * 3)) + rx_inner) + 768)]));
- }
- }
- }
- }
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[((((int)threadIdx.x) / 49) * 64)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 256)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 512)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 768)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 32)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 288)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 544)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 800)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 1)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 257)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 513)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 769)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 33)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 289)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 545)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 801)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 2)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 258)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 514)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 770)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 34)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 290)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 546)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 802)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 3)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 259)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 515)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 771)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 35)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 291)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 547)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 803)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 4)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 260)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 516)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 772)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 36)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 292)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 548)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 804)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 5)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 261)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 517)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 773)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 37)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 293)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 549)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 805)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 6)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 262)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 518)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 774)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 38)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 294)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 550)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 806)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 7)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 263)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 519)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 775)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 39)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 295)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 551)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 807)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 8)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 264)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 520)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 776)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 40)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 296)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 552)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 808)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 9)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 265)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 521)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 777)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 41)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 297)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 553)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 809)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 10)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 266)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 522)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 778)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 42)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 298)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 554)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 810)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 11)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 267)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 523)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 779)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 43)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 299)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 555)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 811)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 12)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 268)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 524)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 780)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 44)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 300)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 556)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 812)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 13)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 269)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 525)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 781)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 45)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 301)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 557)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 813)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 14)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 270)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 526)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 782)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 46)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 302)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 558)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 814)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 15)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 271)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 527)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 783)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 47)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 303)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 559)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 815)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 16)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 272)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 528)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 784)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 48)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 304)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 560)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 816)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 17)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 273)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 529)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 785)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 49)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 305)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 561)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 817)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 18)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 274)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 530)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 786)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 50)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 306)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 562)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 818)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 19)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 275)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 531)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 787)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 51)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 307)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 563)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 819)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 20)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 276)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 532)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 788)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 52)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 308)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 564)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 820)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 21)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 277)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 533)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 789)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 53)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 309)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 565)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 821)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 22)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 278)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 534)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 790)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 54)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 310)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 566)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 822)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 23)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 279)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 535)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 791)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 55)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 311)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 567)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 823)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 24)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 280)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 536)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 792)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 56)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 312)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 568)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 824)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 25)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 281)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 537)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 793)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 57)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 313)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 569)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 825)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 26)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 282)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 538)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 794)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 58)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 314)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 570)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 826)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 27)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 283)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 539)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 795)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 59)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 315)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 571)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 827)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 28)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 284)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 540)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 796)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 60)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 316)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 572)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 828)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 29)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 285)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 541)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 797)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 61)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 317)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 573)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 829)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 30)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 286)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 542)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 798)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 62)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 318)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 574)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 830)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 31)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 287)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 543)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 799)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 63)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 319)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 575)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 831)]));
+ __syncthreads();
+ pad_temp_shared[((int)threadIdx.x)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) - 6)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 196)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 190)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 392)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 386)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 588)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 582)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 778)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 980)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 974)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1176)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1170)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1372)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1366)] : 0.000000e+00f);
+ kernel_shared[((int)threadIdx.x)] = kernel[((((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) >> 5) * 4608)) + (rc_outer_outer * 288)) + ((((int)threadIdx.x) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ kernel_shared[(((int)threadIdx.x) + 196)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 196) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 4) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ kernel_shared[(((int)threadIdx.x) + 392)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 392) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 8) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ kernel_shared[(((int)threadIdx.x) + 588)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 588) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 12) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ kernel_shared[(((int)threadIdx.x) + 784)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 784) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 16) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ if (((int)threadIdx.x) < 44) {
+ kernel_shared[(((int)threadIdx.x) + 980)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 980) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 20) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
}
+ __syncthreads();
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[((((int)threadIdx.x) / 49) * 64)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 256)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 512)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 768)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 32)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 288)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 544)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 800)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 1)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 257)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 513)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 769)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 33)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 289)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 545)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 801)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 2)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 258)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 514)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 770)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 34)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 290)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 546)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 802)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 3)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 259)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 515)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 771)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 35)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 291)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 547)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 803)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 4)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 260)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 516)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 772)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 36)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 292)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 548)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 804)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 5)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 261)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 517)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 773)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 37)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 293)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 549)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 805)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 6)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 262)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 518)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 774)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 38)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 294)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 550)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 806)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 7)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 263)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 519)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 775)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 39)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 295)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 551)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 807)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 8)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 264)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 520)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 776)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 40)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 296)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 552)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 808)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 9)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 265)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 521)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 777)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 41)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 297)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 553)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 809)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 10)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 266)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 522)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 778)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 42)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 298)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 554)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 810)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 11)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 267)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 523)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 779)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 43)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 299)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 555)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 811)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 12)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 268)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 524)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 780)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 44)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 300)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 556)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 812)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 13)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 269)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 525)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 781)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 45)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 301)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 557)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 813)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 14)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 270)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 526)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 782)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 46)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 302)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 558)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 814)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 15)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 271)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 527)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 783)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 47)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 303)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 559)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 815)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 16)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 272)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 528)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 784)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 48)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 304)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 560)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 816)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 17)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 273)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 529)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 785)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 49)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 305)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 561)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 817)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 18)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 274)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 530)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 786)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 50)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 306)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 562)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 818)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 19)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 275)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 531)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 787)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 51)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 307)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 563)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 819)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 20)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 276)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 532)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 788)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 52)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 308)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 564)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 820)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 21)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 277)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 533)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 789)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 53)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 309)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 565)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 821)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 22)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 278)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 534)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 790)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 54)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 310)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 566)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 822)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 23)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 279)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 535)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 791)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 55)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 311)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 567)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 823)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 24)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 280)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 536)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 792)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 56)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 312)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 568)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 824)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 25)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 281)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 537)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 793)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 57)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 313)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 569)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 825)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 26)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 282)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 538)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 794)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 58)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 314)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 570)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 826)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 27)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 283)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 539)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 795)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 59)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 315)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 571)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 827)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 28)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 284)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 540)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 796)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 60)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 316)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 572)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 828)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 29)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 285)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 541)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 797)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 61)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 317)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 573)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 829)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 30)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 286)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 542)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 798)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 62)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 318)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 574)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 830)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 31)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 287)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 543)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 799)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 63)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 319)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 575)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 831)]));
}
}
- for (int i1_inner = 0; i1_inner < 4; ++i1_inner) {
- compute[((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 196)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49))] = max((conv2d_nchw[i1_inner] + bias[(((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 4)) + i1_inner)]), 0.000000e+00f);
- compute[(((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 196)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49)) + 784)] = max((conv2d_nchw[(i1_inner + 4)] + bias[((((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 4)) + i1_inner) + 16)]), 0.000000e+00f);
+ for (int i1_inner = 0; i1_inner < 2; ++i1_inner) {
+ compute[((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 98)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49))] = max((conv2d_nchw[i1_inner] + bias[(((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 2)) + i1_inner)]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 98)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49)) + 392)] = max((conv2d_nchw[(i1_inner + 2)] + bias[((((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 2)) + i1_inner) + 8)]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 98)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49)) + 784)] = max((conv2d_nchw[(i1_inner + 4)] + bias[((((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 2)) + i1_inner) + 16)]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 98)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49)) + 1176)] = max((conv2d_nchw[(i1_inner + 6)] + bias[((((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 2)) + i1_inner) + 24)]), 0.000000e+00f);
}
}
@@ -567,7 +2204,7 @@ In the example below we resume the status and do more 5 trials.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 3 minutes 17.057 seconds)
+ **Total running time of the script:** ( 3 minutes 25.235 seconds)
.. _sphx_glr_download_how_to_tune_with_autoscheduler_tune_conv2d_layer_cuda.py:
diff --git a/docs/_sources/how_to/tune_with_autoscheduler/tune_network_cuda.rst.txt b/docs/_sources/how_to/tune_with_autoscheduler/tune_network_cuda.rst.txt
index efa998eac..a24d1722e 100644
--- a/docs/_sources/how_to/tune_with_autoscheduler/tune_network_cuda.rst.txt
+++ b/docs/_sources/how_to/tune_with_autoscheduler/tune_network_cuda.rst.txt
@@ -647,7 +647,7 @@ so we can read the log file and load the best schedules.
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 9.8031 9.8079 9.8367 9.7646 0.0296
+ 9.9558 9.9761 10.0060 9.8852 0.0513
diff --git a/docs/_sources/how_to/tune_with_autoscheduler/tune_network_x86.rst.txt b/docs/_sources/how_to/tune_with_autoscheduler/tune_network_x86.rst.txt
index 5e2634af0..50f5c3e81 100644
--- a/docs/_sources/how_to/tune_with_autoscheduler/tune_network_x86.rst.txt
+++ b/docs/_sources/how_to/tune_with_autoscheduler/tune_network_x86.rst.txt
@@ -666,7 +666,7 @@ so we can read the log file and load the best schedules.
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 741.7649 742.4460 742.8278 740.0208 1.2431
+ 754.8454 754.7219 755.2403 754.5740 0.2857
@@ -694,7 +694,7 @@ Other Tips
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 20.723 seconds)
+ **Total running time of the script:** ( 1 minutes 22.299 seconds)
.. _sphx_glr_download_how_to_tune_with_autoscheduler_tune_network_x86.py:
diff --git a/docs/_sources/how_to/tune_with_autoscheduler/tune_sparse_x86.rst.txt b/docs/_sources/how_to/tune_with_autoscheduler/tune_sparse_x86.rst.txt
index cb501fa73..548f0ac86 100644
--- a/docs/_sources/how_to/tune_with_autoscheduler/tune_sparse_x86.rst.txt
+++ b/docs/_sources/how_to/tune_with_autoscheduler/tune_sparse_x86.rst.txt
@@ -397,77 +397,103 @@ layout transformation, parallelization, vectorization, unrolling, and operator f
placeholder_4: Buffer(placeholder_14: Pointer(float32), float32, [65536], []),
compute: Buffer(compute_2: Pointer(float32), float32, [65536], [])}
buffer_map = {placeholder_5: placeholder, placeholder_6: placeholder_1, placeholder_7: placeholder_2, placeholder_8: placeholder_3, placeholder_9: placeholder_4, compute_1: compute}
- preflattened_buffer_map = {placeholder_6: placeholder_15: Buffer(placeholder_11, float32, [4916, 16, 1], []), placeholder_9: placeholder_16: Buffer(placeholder_14, float32, [128, 512], []), placeholder_7: placeholder_17: Buffer(placeholder_12, int32, [4916], []), placeholder_5: placeholder_18: Buffer(placeholder_10, float32, [128, 256], []), placeholder_8: placeholder_19: Buffer(placeholder_13, int32, [33], []), compute_1: compute_3: Buffer(compute_2, float32, [128, 512], [])} {
- for (i0.outer.i1.outer.fused: int32, 0, 128) "parallel" {
- allocate(compute_4: Pointer(global float32), float32, [512]), storage_scope = global {
- for (nb_j.inner: int32, 0, 2) {
- for (i.inner.init: int32, 0, 16) {
- let cse_var_1: int32 = ((i.inner.init*32) + (nb_j.inner*16))
- {
- compute_5: Buffer(compute_4, float32, [512], [])[cse_var_1] = 0f32
- compute_5[(cse_var_1 + 1)] = 0f32
- compute_5[(cse_var_1 + 2)] = 0f32
- compute_5[(cse_var_1 + 3)] = 0f32
- compute_5[(cse_var_1 + 4)] = 0f32
- compute_5[(cse_var_1 + 5)] = 0f32
- compute_5[(cse_var_1 + 6)] = 0f32
- compute_5[(cse_var_1 + 7)] = 0f32
- compute_5[(cse_var_1 + 8)] = 0f32
- compute_5[(cse_var_1 + 9)] = 0f32
- compute_5[(cse_var_1 + 10)] = 0f32
- compute_5[(cse_var_1 + 11)] = 0f32
- compute_5[(cse_var_1 + 12)] = 0f32
- compute_5[(cse_var_1 + 13)] = 0f32
- compute_5[(cse_var_1 + 14)] = 0f32
- compute_5[(cse_var_1 + 15)] = 0f32
- }
+ preflattened_buffer_map = {placeholder_5: placeholder_15: Buffer(placeholder_10, float32, [128, 256], []), placeholder_6: placeholder_16: Buffer(placeholder_11, float32, [4916, 16, 1], []), placeholder_7: placeholder_17: Buffer(placeholder_12, int32, [4916], []), placeholder_8: placeholder_18: Buffer(placeholder_13, int32, [33], []), placeholder_9: placeholder_19: Buffer(placeholder_14, float32, [128, 512], []), compute_1: compute_3: Buffer(compute_2, float32, [128, 512], [])} {
+ allocate(compute_4: Pointer(global float32), float32, [2048]), storage_scope = global;
+ for (i1.outer: int32, 0, 32) {
+ for (i.outer.inner: int32, 0, 4) {
+ for (i.inner.init: int32, 0, 32) {
+ let cse_var_1: int32 = ((i.outer.inner*512) + (i.inner.init*16))
+ {
+ compute_5: Buffer(compute_4, float32, [2048], [])[cse_var_1] = 0f32
+ compute_5[(cse_var_1 + 1)] = 0f32
+ compute_5[(cse_var_1 + 2)] = 0f32
+ compute_5[(cse_var_1 + 3)] = 0f32
+ compute_5[(cse_var_1 + 4)] = 0f32
+ compute_5[(cse_var_1 + 5)] = 0f32
+ compute_5[(cse_var_1 + 6)] = 0f32
+ compute_5[(cse_var_1 + 7)] = 0f32
+ compute_5[(cse_var_1 + 8)] = 0f32
+ compute_5[(cse_var_1 + 9)] = 0f32
+ compute_5[(cse_var_1 + 10)] = 0f32
+ compute_5[(cse_var_1 + 11)] = 0f32
+ compute_5[(cse_var_1 + 12)] = 0f32
+ compute_5[(cse_var_1 + 13)] = 0f32
+ compute_5[(cse_var_1 + 14)] = 0f32
+ compute_5[(cse_var_1 + 15)] = 0f32
}
- for (elem_idx: int32, 0, let cse_var_2: int32 = ((floormod(i0.outer.i1.outer.fused, 16)*2) + nb_j.inner) in (placeholder_3[(cse_var_2 + 1)] - placeholder_3[cse_var_2])) {
- for (i.inner: int32, 0, 16) {
- let cse_var_21: int32 = (elem_idx*16)
- let cse_var_20: int32 = ((i.inner*32) + (nb_j.inner*16))
- let cse_var_19: int32 = ((floormod(i0.outer.i1.outer.fused, 16)*2) + nb_j.inner)
- let cse_var_18: int32 = ((floordiv(i0.outer.i1.outer.fused, 16)*4096) + (i.inner*256))
- let cse_var_17: int32 = (cse_var_20 + 9)
- let cse_var_16: int32 = (cse_var_20 + 8)
- let cse_var_15: int32 = (cse_var_20 + 7)
- let cse_var_14: int32 = (cse_var_20 + 6)
- let cse_var_13: int32 = (cse_var_20 + 5)
- let cse_var_12: int32 = (cse_var_20 + 4)
- let cse_var_11: int32 = (cse_var_20 + 3)
- let cse_var_10: int32 = (cse_var_20 + 2)
- let cse_var_9: int32 = (cse_var_20 + 15)
- let cse_var_8: int32 = (cse_var_20 + 14)
- let cse_var_7: int32 = (cse_var_20 + 13)
- let cse_var_6: int32 = (cse_var_20 + 12)
- let cse_var_5: int32 = (cse_var_20 + 11)
- let cse_var_4: int32 = (cse_var_20 + 10)
- let cse_var_3: int32 = (cse_var_20 + 1)
- {
- compute_5[cse_var_20] = (compute_5[cse_var_20] + (placeholder_1[((placeholder_3[cse_var_19]*16) + cse_var_21)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_3] = (compute_5[cse_var_3] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 1)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_10] = (compute_5[cse_var_10] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 2)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_11] = (compute_5[cse_var_11] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 3)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_12] = (compute_5[cse_var_12] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 4)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_13] = (compute_5[cse_var_13] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 5)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_14] = (compute_5[cse_var_14] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 6)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_15] = (compute_5[cse_var_15] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 7)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_16] = (compute_5[cse_var_16] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 8)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_17] = (compute_5[cse_var_17] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 9)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_4] = (compute_5[cse_var_4] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 10)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_5] = (compute_5[cse_var_5] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 11)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_6] = (compute_5[cse_var_6] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 12)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_7] = (compute_5[cse_var_7] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 13)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_8] = (compute_5[cse_var_8] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 14)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_9] = (compute_5[cse_var_9] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 15)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- }
+ }
+ for (elem_idx: int32, 0, (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])) {
+ for (i.inner: int32, 0, 32) {
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_2: int32 = ((i.outer.inner*512) + (i.inner*16))
+ compute_5[cse_var_2] = (compute_5[cse_var_2] + (placeholder_1[((placeholder_3[i1.outer]*16) + (elem_idx*16))]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_3: int32 = (((i.outer.inner*512) + (i.inner*16)) + 1)
+ compute_5[cse_var_3] = (compute_5[cse_var_3] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 1)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_4: int32 = (((i.outer.inner*512) + (i.inner*16)) + 2)
+ compute_5[cse_var_4] = (compute_5[cse_var_4] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 2)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_5: int32 = (((i.outer.inner*512) + (i.inner*16)) + 3)
+ compute_5[cse_var_5] = (compute_5[cse_var_5] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 3)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_6: int32 = (((i.outer.inner*512) + (i.inner*16)) + 4)
+ compute_5[cse_var_6] = (compute_5[cse_var_6] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 4)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_7: int32 = (((i.outer.inner*512) + (i.inner*16)) + 5)
+ compute_5[cse_var_7] = (compute_5[cse_var_7] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 5)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_8: int32 = (((i.outer.inner*512) + (i.inner*16)) + 6)
+ compute_5[cse_var_8] = (compute_5[cse_var_8] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 6)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_9: int32 = (((i.outer.inner*512) + (i.inner*16)) + 7)
+ compute_5[cse_var_9] = (compute_5[cse_var_9] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 7)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_10: int32 = (((i.outer.inner*512) + (i.inner*16)) + 8)
+ compute_5[cse_var_10] = (compute_5[cse_var_10] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 8)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_11: int32 = (((i.outer.inner*512) + (i.inner*16)) + 9)
+ compute_5[cse_var_11] = (compute_5[cse_var_11] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 9)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_12: int32 = (((i.outer.inner*512) + (i.inner*16)) + 10)
+ compute_5[cse_var_12] = (compute_5[cse_var_12] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 10)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_13: int32 = (((i.outer.inner*512) + (i.inner*16)) + 11)
+ compute_5[cse_var_13] = (compute_5[cse_var_13] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 11)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_14: int32 = (((i.outer.inner*512) + (i.inner*16)) + 12)
+ compute_5[cse_var_14] = (compute_5[cse_var_14] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 12)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_15: int32 = (((i.outer.inner*512) + (i.inner*16)) + 13)
+ compute_5[cse_var_15] = (compute_5[cse_var_15] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 13)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_16: int32 = (((i.outer.inner*512) + (i.inner*16)) + 14)
+ compute_5[cse_var_16] = (compute_5[cse_var_16] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 14)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_17: int32 = (((i.outer.inner*512) + (i.inner*16)) + 15)
+ compute_5[cse_var_17] = (compute_5[cse_var_17] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 15)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
}
}
}
- for (i0.inner: int32, 0, 16) {
- let cse_var_22: int32 = (((floordiv(i0.outer.i1.outer.fused, 16)*8192) + (i0.inner*512)) + (floormod(i0.outer.i1.outer.fused, 16)*32))
- compute[ramp(cse_var_22, 1, 32)] = max((compute_5[ramp((i0.inner*32), 1, 32)] + placeholder_4[ramp(cse_var_22, 1, 32)]), broadcast(0f32, 32))
- }
+ }
+ for (i0.inner: int32, 0, 128) {
+ let cse_var_18: int32 = ((i0.inner*512) + (i1.outer*16))
+ compute[ramp(cse_var_18, 1, 16)] = max((compute_5[ramp((i0.inner*16), 1, 16)] + placeholder_4[ramp(cse_var_18, 1, 16)]), broadcast(0f32, 16))
}
}
}
@@ -522,7 +548,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 1.651 ms
+ Execution time of this operator: 1.721 ms
diff --git a/docs/_sources/how_to/tune_with_autotvm/sg_execution_times.rst.txt b/docs/_sources/how_to/tune_with_autotvm/sg_execution_times.rst.txt
index b130fe35c..c00c052f4 100644
--- a/docs/_sources/how_to/tune_with_autotvm/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/tune_with_autotvm/sg_execution_times.rst.txt
@@ -5,16 +5,16 @@
Computation times
=================
-**00:44.846** total execution time for **how_to_tune_with_autotvm** files:
+**00:46.348** total execution time for **how_to_tune_with_autotvm** files:
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_conv2d_cuda.py` (``tune_conv2d_cuda.py``) | 00:44.812 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_conv2d_cuda.py` (``tune_conv2d_cuda.py``) | 00:46.314 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_x86.py` (``tune_relay_x86.py``) | 00:00.019 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_x86.py` (``tune_relay_x86.py``) | 00:00.020 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_cuda.py` (``tune_relay_cuda.py``) | 00:00.005 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_mobile_gpu.py` (``tune_relay_mobile_gpu.py``) | 00:00.005 | 0.0 MB |
-+--------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_arm.py` (``tune_relay_arm.py``) | 00:00.005 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_mobile_gpu.py` (``tune_relay_mobile_gpu.py``) | 00:00.005 | 0.0 MB |
++--------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/tune_with_autotvm/tune_conv2d_cuda.rst.txt b/docs/_sources/how_to/tune_with_autotvm/tune_conv2d_cuda.rst.txt
index b64c3127e..43ece9f50 100644
--- a/docs/_sources/how_to/tune_with_autotvm/tune_conv2d_cuda.rst.txt
+++ b/docs/_sources/how_to/tune_with_autotvm/tune_conv2d_cuda.rst.txt
@@ -1156,8 +1156,8 @@ for this template
TimeoutError
[('tile_f', [-1, 2, 1, 64]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4909501
- No: 9 GFLOPS: 190.31/190.31 result: MeasureResult(costs=(0.0012164683333333334,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.792853593826294, timestamp=1662973591.8327475) [('tile_f', [-1, 1, 4, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5072689
- No: 10 GFLOPS: 0.00/190.31 result: Traceback (most recent call last):
+ No: 9 GFLOPS: 219.45/219.45 result: MeasureResult(costs=(0.0010549379172413794,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.907240867614746, timestamp=1663013704.4100664) [('tile_f', [-1, 1, 4, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5072689
+ No: 10 GFLOPS: 0.00/219.45 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1280,8 +1280,8 @@ for this template
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 4, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 64, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5092711
- No: 11 GFLOPS: 260.87/260.87 result: MeasureResult(costs=(0.0008874141546961326,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7072300910949707, timestamp=1662973592.7237387) [('tile_f', [-1, 8, 2, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4264713
- No: 12 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+ No: 11 GFLOPS: 259.35/259.35 result: MeasureResult(costs=(0.0008926184475138122,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6609346866607666, timestamp=1663013705.284125) [('tile_f', [-1, 8, 2, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4264713
+ No: 12 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1404,7 +1404,7 @@ for this template
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 128, 1, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 256]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,183542
- No: 13 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+ No: 13 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1527,7 +1527,7 @@ for this template
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 8, 8]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 64]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2482196
- No: 14 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+ No: 14 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1650,9 +1650,9 @@ for this template
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 64, 1, 4]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10306226
- No: 15 GFLOPS: 5.46/260.87 result: MeasureResult(costs=(0.042424332,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7882673740386963, timestamp=1662973597.2073998) [('tile_f', [-1, 2, 2, 8]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5330964
- No: 16 GFLOPS: 3.35/260.87 result: MeasureResult(costs=(0.06920571675,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.458357572555542, timestamp=1662973598.4397767) [('tile_f', [-1, 8, 4, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2140058
- No: 17 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+ No: 15 GFLOPS: 5.27/259.35 result: MeasureResult(costs=(0.0439689325,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.8793528079986572, timestamp=1663013709.848698) [('tile_f', [-1, 2, 2, 8]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5330964
+ No: 16 GFLOPS: 3.34/259.35 result: MeasureResult(costs=(0.06932482525,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.540909290313721, timestamp=1663013711.0937493) [('tile_f', [-1, 8, 4, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2140058
+ No: 17 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 142, in build
res = future.result()
File "/usr/lib/python3.7/concurrent/futures/_base.py", line 435, in result
@@ -1670,8 +1670,8 @@ for this template
TimeoutError
[('tile_f', [-1, 2, 2, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 16]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10195251
- No: 18 GFLOPS: 26.14/260.87 result: MeasureResult(costs=(0.008855842416666667,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.1080987453460693, timestamp=1662973609.323898) [('tile_f', [-1, 4, 8, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6068603
- No: 19 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+ No: 18 GFLOPS: 28.28/259.35 result: MeasureResult(costs=(0.008186122,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.3027231693267822, timestamp=1663013722.1563756) [('tile_f', [-1, 4, 8, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6068603
+ No: 19 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1794,7 +1794,7 @@ for this template
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 16, 4, 8]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6956993
- No: 20 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+ No: 20 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1973,7 +1973,7 @@ and measure running time.
Best config:
[('tile_f', [-1, 8, 2, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4264713
Finish loading 20 records
- Time cost of this operator: 0.001305
+ Time cost of this operator: 0.001259
diff --git a/docs/_sources/how_to/work_with_microtvm/micro_autotune.rst.txt b/docs/_sources/how_to/work_with_microtvm/micro_autotune.rst.txt
index e87fc4008..cb3f8121d 100644
--- a/docs/_sources/how_to/work_with_microtvm/micro_autotune.rst.txt
+++ b/docs/_sources/how_to/work_with_microtvm/micro_autotune.rst.txt
@@ -329,10 +329,10 @@ Timing the untuned program
########## Build without Autotuning ##########
Node Name Ops Time(us) Time(%) Shape Inputs Outputs Measurements(us)
--------- --- -------- ------- ----- ------ ------- ----------------
- tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 305.2 98.66 (1, 2, 10, 10, 3) 2 1 [305.2]
- tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.215 1.039 (1, 6, 10, 10) 1 1 [3.215]
- tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.932 0.301 (1, 1, 10, 10, 3) 1 1 [0.932]
- Total_time - 309.347 - - - - -
+ tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 311.4 98.634 (1, 2, 10, 10, 3) 2 1 [311.4]
+ tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.171 1.005 (1, 6, 10, 10) 1 1 [3.171]
+ tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 1.142 0.362 (1, 1, 10, 10, 3) 1 1 [1.142]
+ Total_time - 315.713 - - - - -
@@ -398,10 +398,10 @@ Timing the tuned program
########## Build with Autotuning ##########
Node Name Ops Time(us) Time(%) Shape Inputs Outputs Measurements(us)
--------- --- -------- ------- ----- ------ ------- ----------------
- tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 128.7 97.884 (1, 6, 10, 10, 1) 2 1 [128.7]
- tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.806 1.373 (1, 6, 10, 10) 1 1 [1.806]
- tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.976 0.743 (1, 1, 10, 10, 3) 1 1 [0.976]
- Total_time - 131.482 - - - - -
+ tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 218.0 98.617 (1, 1, 10, 10, 6) 2 1 [218.0]
+ tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 2.208 0.999 (1, 6, 10, 10) 1 1 [2.208]
+ tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.849 0.384 (1, 3, 10, 10, 1) 1 1 [0.849]
+ Total_time - 221.057 - - - - -
diff --git a/docs/_sources/how_to/work_with_microtvm/micro_train.rst.txt b/docs/_sources/how_to/work_with_microtvm/micro_train.rst.txt
index 5d8eb9b8d..988240029 100644
--- a/docs/_sources/how_to/work_with_microtvm/micro_train.rst.txt
+++ b/docs/_sources/how_to/work_with_microtvm/micro_train.rst.txt
@@ -225,7 +225,7 @@ take about **2 minutes** to download the Stanford Cars, while COCO 2017 validati
.. code-block:: none
- '/tmp/tmpkzbcrnpu/images/random'
+ '/tmp/tmp75inl8ex/images/random'
@@ -325,8 +325,8 @@ objects to other stuff? We can display some examples from our datasets using ``m
.. code-block:: none
- /tmp/tmpkzbcrnpu/images/target contains 8144 images
- /tmp/tmpkzbcrnpu/images/random contains 5000 images
+ /tmp/tmp75inl8ex/images/target contains 8144 images
+ /tmp/tmp75inl8ex/images/random contains 5000 images
@@ -501,13 +501,13 @@ the time on our validation set).
.. code-block:: none
Epoch 1/3
- 328/328 - 46s - loss: 0.2312 - accuracy: 0.9217 - val_loss: 0.1355 - val_accuracy: 0.9562 - 46s/epoch - 139ms/step
+ 328/328 - 47s - loss: 0.2088 - accuracy: 0.9299 - val_loss: 0.1318 - val_accuracy: 0.9558 - 47s/epoch - 142ms/step
Epoch 2/3
- 328/328 - 42s - loss: 0.0931 - accuracy: 0.9651 - val_loss: 0.1191 - val_accuracy: 0.9622 - 42s/epoch - 129ms/step
+ 328/328 - 43s - loss: 0.1011 - accuracy: 0.9610 - val_loss: 0.1258 - val_accuracy: 0.9630 - 43s/epoch - 132ms/step
Epoch 3/3
- 328/328 - 42s - loss: 0.0649 - accuracy: 0.9750 - val_loss: 0.2244 - val_accuracy: 0.9316 - 42s/epoch - 129ms/step
+ 328/328 - 43s - loss: 0.0675 - accuracy: 0.9744 - val_loss: 0.1085 - val_accuracy: 0.9630 - 43s/epoch - 131ms/step
- <keras.callbacks.History object at 0x7f6154ba6390>
+ <keras.callbacks.History object at 0x7fda749133d0>
@@ -871,7 +871,7 @@ Arduino tutorial for how to do that `on GitHub <https://github.com/guberti/tvm-a
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 4 minutes 36.292 seconds)
+ **Total running time of the script:** ( 4 minutes 48.525 seconds)
.. _sphx_glr_download_how_to_work_with_microtvm_micro_train.py:
diff --git a/docs/_sources/how_to/work_with_microtvm/sg_execution_times.rst.txt b/docs/_sources/how_to/work_with_microtvm/sg_execution_times.rst.txt
index 90fe6bb62..ce1800d74 100644
--- a/docs/_sources/how_to/work_with_microtvm/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/work_with_microtvm/sg_execution_times.rst.txt
@@ -5,16 +5,16 @@
Computation times
=================
-**05:27.627** total execution time for **how_to_work_with_microtvm** files:
+**05:41.583** total execution time for **how_to_work_with_microtvm** files:
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``) | 04:36.292 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``) | 04:48.525 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``) | 00:40.411 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``) | 00:41.646 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``) | 00:07.799 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``) | 00:08.116 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``) | 00:03.124 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``) | 00:03.295 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_microtvm_micro_ethosu.py` (``micro_ethosu.py``) | 00:00.001 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt b/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt
index 74203acdf..5eea14377 100644
--- a/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt
@@ -5,14 +5,14 @@
Computation times
=================
-**00:38.996** total execution time for **how_to_work_with_relay** files:
+**00:42.559** total execution time for **how_to_work_with_relay** files:
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_using_pipeline_executor.py` (``using_pipeline_executor.py``) | 00:30.095 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_pipeline_executor.py` (``using_pipeline_executor.py``) | 00:31.307 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``) | 00:07.682 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``) | 00:09.811 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``) | 00:01.212 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``) | 00:01.435 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_relay_using_relay_viz.py` (``using_relay_viz.py``) | 00:00.007 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/work_with_schedules/intrin_math.rst.txt b/docs/_sources/how_to/work_with_schedules/intrin_math.rst.txt
index b54c1916b..164cde00d 100644
--- a/docs/_sources/how_to/work_with_schedules/intrin_math.rst.txt
+++ b/docs/_sources/how_to/work_with_schedules/intrin_math.rst.txt
@@ -261,7 +261,7 @@ The following example customizes CUDA lowering rule for :code:`exp`.
.. code-block:: none
- <function my_cuda_math_rule at 0x7f60f3ae8950>
+ <function my_cuda_math_rule at 0x7fd9f52db710>
diff --git a/docs/_sources/how_to/work_with_schedules/sg_execution_times.rst.txt b/docs/_sources/how_to/work_with_schedules/sg_execution_times.rst.txt
index 8298d1337..274faae45 100644
--- a/docs/_sources/how_to/work_with_schedules/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/work_with_schedules/sg_execution_times.rst.txt
@@ -5,20 +5,20 @@
Computation times
=================
-**00:04.082** total execution time for **how_to_work_with_schedules** files:
+**00:07.923** total execution time for **how_to_work_with_schedules** files:
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``) | 00:02.023 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``) | 00:05.711 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``) | 00:00.897 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``) | 00:00.963 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``) | 00:00.496 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``) | 00:00.542 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``) | 00:00.489 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``) | 00:00.527 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_extern_op.py` (``extern_op.py``) | 00:00.096 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_extern_op.py` (``extern_op.py``) | 00:00.097 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_schedule_primitives.py` (``schedule_primitives.py``) | 00:00.041 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_schedule_primitives.py` (``schedule_primitives.py``) | 00:00.042 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_schedules_tedd.py` (``tedd.py``) | 00:00.026 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/work_with_schedules/tensorize.rst.txt b/docs/_sources/how_to/work_with_schedules/tensorize.rst.txt
index c6d94253a..1e37f7639 100644
--- a/docs/_sources/how_to/work_with_schedules/tensorize.rst.txt
+++ b/docs/_sources/how_to/work_with_schedules/tensorize.rst.txt
@@ -347,7 +347,7 @@ The importing needs to happen before the tensorized GEMV being executed.
C: Buffer(C_2: Pointer(float32), float32, [524288], [])}
buffer_map = {A_1: A, B_1: B, C_1: C}
preflattened_buffer_map = {A_1: A_3: Buffer(A_2, float32, [1024, 64], []), B_1: B_3: Buffer(B_2, float32, [512, 64], []), C_1: C_3: Buffer(C_2, float32, [1024, 512], [])} {
- attr [IterVar(i: int32, (nullptr), "DataPar", "")] "pragma_import_llvm" = "; ModuleID = '/tmp/tmphvuuz9sb/input0.cc'\nsource_filename = \"/tmp/tmphvuuz9sb/input0.cc\"\ntarget datalayout = \"e-m:e-i64:64-f80:128-n8:16:32:64-S128\"\ntarget triple = \"x86_64-pc-linux-gnu\"\n\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local i32 @gemv_update(float*, float*, float*, i32, i32, i32) #0 {\n %7 = alloca float*, align 8\n %8 = alloca float*, align 8\n %9 = alloca floa [...]
+ attr [IterVar(i: int32, (nullptr), "DataPar", "")] "pragma_import_llvm" = "; ModuleID = '/tmp/tmpozmw_jum/input0.cc'\nsource_filename = \"/tmp/tmpozmw_jum/input0.cc\"\ntarget datalayout = \"e-m:e-i64:64-f80:128-n8:16:32:64-S128\"\ntarget triple = \"x86_64-pc-linux-gnu\"\n\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local i32 @gemv_update(float*, float*, float*, i32, i32, i32) #0 {\n %7 = alloca float*, align 8\n %8 = alloca float*, align 8\n %9 = alloca floa [...]
for (i, 0, 1024) {
for (j.outer: int32, 0, 32) {
@tir.call_extern("gemv_update", @tir.tvm_access_ptr(@tir.type_annotation(, dtype=float32), C_2, ((i*512) + (j.outer*16)), 16, 2, dtype=handle), @tir.tvm_access_ptr(@tir.type_annotation(, dtype=float32), A_2, (i*64), 64, 1, dtype=handle), @tir.tvm_access_ptr(@tir.type_annotation(, dtype=float32), B_2, (j.outer*1024), 1024, 1, dtype=handle), 16, 64, 64, dtype=int32)
diff --git a/docs/_sources/topic/vta/tutorials/autotvm/sg_execution_times.rst.txt b/docs/_sources/topic/vta/tutorials/autotvm/sg_execution_times.rst.txt
index c3f64e05c..9095db2f5 100644
--- a/docs/_sources/topic/vta/tutorials/autotvm/sg_execution_times.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/autotvm/sg_execution_times.rst.txt
@@ -5,10 +5,10 @@
Computation times
=================
-**00:20.636** total execution time for **topic_vta_tutorials_autotvm** files:
+**00:21.045** total execution time for **topic_vta_tutorials_autotvm** files:
+---------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:20.630 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:21.039 | 0.0 MB |
+---------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_alu_vta.py` (``tune_alu_vta.py``) | 00:00.006 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_alu_vta.py` (``tune_alu_vta.py``) | 00:00.007 | 0.0 MB |
+---------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/topic/vta/tutorials/frontend/deploy_classification.rst.txt b/docs/_sources/topic/vta/tutorials/frontend/deploy_classification.rst.txt
index c16c0724c..80607250f 100644
--- a/docs/_sources/topic/vta/tutorials/frontend/deploy_classification.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/frontend/deploy_classification.rst.txt
@@ -291,7 +291,7 @@ The compilation steps are:
DeprecationWarning,
/workspace/vta/tutorials/frontend/deploy_classification.py:213: DeprecationWarning: legacy graph executor behavior of producing json / lib / params will be removed in the next release. Please see documents of tvm.contrib.graph_executor.GraphModule for the new recommended usage.
relay_prog, target=tvm.target.Target(target, host=env.target_host), params=params
- resnet18_v1 inference graph built in 21.88s!
+ resnet18_v1 inference graph built in 22.33s!
diff --git a/docs/_sources/topic/vta/tutorials/frontend/deploy_detection.rst.txt b/docs/_sources/topic/vta/tutorials/frontend/deploy_detection.rst.txt
index ed1618588..4c0c7190c 100644
--- a/docs/_sources/topic/vta/tutorials/frontend/deploy_detection.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/frontend/deploy_detection.rst.txt
@@ -335,7 +335,7 @@ The compilation steps are:
"target_host parameter is going to be deprecated. "
/workspace/python/tvm/relay/build_module.py:348: DeprecationWarning: Please use input parameter mod (tvm.IRModule) instead of deprecated parameter mod (tvm.relay.function.Function)
DeprecationWarning,
- yolov3-tiny inference graph built in 15.71s!
+ yolov3-tiny inference graph built in 15.98s!
diff --git a/docs/_sources/topic/vta/tutorials/frontend/sg_execution_times.rst.txt b/docs/_sources/topic/vta/tutorials/frontend/sg_execution_times.rst.txt
index c6e5f6b9b..df5f9c718 100644
--- a/docs/_sources/topic/vta/tutorials/frontend/sg_execution_times.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/frontend/sg_execution_times.rst.txt
@@ -5,10 +5,10 @@
Computation times
=================
-**01:29.603** total execution time for **topic_vta_tutorials_frontend** files:
+**01:32.211** total execution time for **topic_vta_tutorials_frontend** files:
+------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``) | 00:48.030 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``) | 00:49.579 | 0.0 MB |
+------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:41.573 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:42.632 | 0.0 MB |
+------------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/topic/vta/tutorials/optimize/sg_execution_times.rst.txt b/docs/_sources/topic/vta/tutorials/optimize/sg_execution_times.rst.txt
index 89a1b9ae0..063024ea0 100644
--- a/docs/_sources/topic/vta/tutorials/optimize/sg_execution_times.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/optimize/sg_execution_times.rst.txt
@@ -5,10 +5,10 @@
Computation times
=================
-**00:02.918** total execution time for **topic_vta_tutorials_optimize** files:
+**00:03.026** total execution time for **topic_vta_tutorials_optimize** files:
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``) | 00:02.560 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``) | 00:02.628 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.358 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.398 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/topic/vta/tutorials/sg_execution_times.rst.txt b/docs/_sources/topic/vta/tutorials/sg_execution_times.rst.txt
index f15a04d64..6b6bacb30 100644
--- a/docs/_sources/topic/vta/tutorials/sg_execution_times.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/sg_execution_times.rst.txt
@@ -5,10 +5,10 @@
Computation times
=================
-**00:00.645** total execution time for **topic_vta_tutorials** files:
+**00:00.733** total execution time for **topic_vta_tutorials** files:
+---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.344 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.385 | 0.0 MB |
+---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.301 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.348 | 0.0 MB |
+---------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt b/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
index cc3c23b92..20ebaed52 100644
--- a/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
+++ b/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
@@ -203,6 +203,13 @@ trials, we can load the best schedule from the log file and apply it.
+.. rst-class:: sphx-glr-script-out
+
+ .. code-block:: none
+
+ *E
+
+
@@ -326,7 +333,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 93.984 ms
+ Execution time of this operator: 94.066 ms
@@ -444,7 +451,7 @@ operations.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 15.309 seconds)
+ **Total running time of the script:** ( 1 minutes 2.104 seconds)
.. _sphx_glr_download_tutorial_auto_scheduler_matmul_x86.py:
diff --git a/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt b/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt
index a59677d84..82ff556bf 100644
--- a/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt
+++ b/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt
@@ -462,16 +462,16 @@ reduce variance, we take 5 measurements and average them.
waiting for device...
device available
Get devices for measurement successfully!
- No: 1 GFLOPS: 9.90/9.90 result: MeasureResult(costs=(0.0271277448,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.566582441329956, timestamp=1662972387.717942) [('tile_y', [-1, 1]), ('tile_x', [-1, 256])],None,80
- No: 2 GFLOPS: 2.71/9.90 result: MeasureResult(costs=(0.0990925466,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7294437885284424, timestamp=1662972389.4650137) [('tile_y', [-1, 4]), ('tile_x', [-1, 8])],None,32
- No: 3 GFLOPS: 11.97/11.97 result: MeasureResult(costs=(0.022425194,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5510766506195068, timestamp=1662972390.502887) [('tile_y', [-1, 64]), ('tile_x', [-1, 32])],None,56
- No: 4 GFLOPS: 1.81/11.97 result: MeasureResult(costs=(0.1483099452,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.484706401824951, timestamp=1662972393.035191) [('tile_y', [-1, 1]), ('tile_x', [-1, 4])],None,20
- No: 5 GFLOPS: 3.69/11.97 result: MeasureResult(costs=(0.0727039434,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2947072982788086, timestamp=1662972394.4600983) [('tile_y', [-1, 256]), ('tile_x', [-1, 16])],None,48
- No: 6 GFLOPS: 1.88/11.97 result: MeasureResult(costs=(0.1427382328,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.413706064224243, timestamp=1662972397.4361706) [('tile_y', [-1, 512]), ('tile_x', [-1, 4])],None,29
- No: 7 GFLOPS: 0.80/11.97 result: MeasureResult(costs=(0.336255298,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.50844144821167, timestamp=1662972403.5031362) [('tile_y', [-1, 512]), ('tile_x', [-1, 2])],None,19
- No: 8 GFLOPS: 10.31/11.97 result: MeasureResult(costs=(0.026032006200000003,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5535826683044434, timestamp=1662972404.0802963) [('tile_y', [-1, 4]), ('tile_x', [-1, 64])],None,62
- No: 9 GFLOPS: 1.86/11.97 result: MeasureResult(costs=(0.1444488502,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.4097604751586914, timestamp=1662972406.6085155) [('tile_y', [-1, 2]), ('tile_x', [-1, 2])],None,11
- No: 10 GFLOPS: 2.67/11.97 result: MeasureResult(costs=(0.100651721,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7215051651000977, timestamp=1662972408.380948) [('tile_y', [-1, 4]), ('tile_x', [-1, 4])],None,22
+ No: 1 GFLOPS: 8.74/8.74 result: MeasureResult(costs=(0.030710399399999998,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6231119632720947, timestamp=1663012481.2453396) [('tile_y', [-1, 1]), ('tile_x', [-1, 256])],None,80
+ No: 2 GFLOPS: 2.89/8.74 result: MeasureResult(costs=(0.09291998679999999,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6363613605499268, timestamp=1663012483.420194) [('tile_y', [-1, 4]), ('tile_x', [-1, 8])],None,32
+ No: 3 GFLOPS: 11.89/11.89 result: MeasureResult(costs=(0.0225735932,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5437366962432861, timestamp=1663012483.9427207) [('tile_y', [-1, 64]), ('tile_x', [-1, 32])],None,56
+ No: 4 GFLOPS: 1.85/11.89 result: MeasureResult(costs=(0.14512044419999998,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.439370632171631, timestamp=1663012486.953823) [('tile_y', [-1, 1]), ('tile_x', [-1, 4])],None,20
+ No: 5 GFLOPS: 3.72/11.89 result: MeasureResult(costs=(0.07225399,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2970874309539795, timestamp=1663012488.3742313) [('tile_y', [-1, 256]), ('tile_x', [-1, 16])],None,48
+ No: 6 GFLOPS: 1.70/11.89 result: MeasureResult(costs=(0.1574418042,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.6412782669067383, timestamp=1663012491.0577068) [('tile_y', [-1, 512]), ('tile_x', [-1, 4])],None,29
+ No: 7 GFLOPS: 0.87/11.89 result: MeasureResult(costs=(0.30774615680000006,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.062021970748901, timestamp=1663012496.6790211) [('tile_y', [-1, 512]), ('tile_x', [-1, 2])],None,19
+ No: 8 GFLOPS: 10.63/11.89 result: MeasureResult(costs=(0.0252591224,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5901744365692139, timestamp=1663012497.290008) [('tile_y', [-1, 4]), ('tile_x', [-1, 64])],None,62
+ No: 9 GFLOPS: 1.66/11.89 result: MeasureResult(costs=(0.1620853014,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.6944057941436768, timestamp=1663012500.1043165) [('tile_y', [-1, 2]), ('tile_x', [-1, 2])],None,11
+ No: 10 GFLOPS: 2.47/11.89 result: MeasureResult(costs=(0.10860995640000001,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.8438212871551514, timestamp=1663012502.0047426) [('tile_y', [-1, 4]), ('tile_x', [-1, 4])],None,22
diff --git a/docs/_sources/tutorial/autotvm_relay_x86.rst.txt b/docs/_sources/tutorial/autotvm_relay_x86.rst.txt
index a35b4713d..96332c03e 100644
--- a/docs/_sources/tutorial/autotvm_relay_x86.rst.txt
+++ b/docs/_sources/tutorial/autotvm_relay_x86.rst.txt
@@ -327,7 +327,7 @@ standard deviation.
.. code-block:: none
- {'mean': 501.64646821969654, 'median': 501.2921533503686, 'std': 1.4021516095849}
+ {'mean': 510.73006595000606, 'median': 511.101387650001, 'std': 1.6193648739236104}
@@ -563,30 +563,30 @@ the tuning data to.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
-
[Task 1/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 1/25] Current/Best: 17.86/ 17.86 GFLOPS | Progress: (4/20) | 6.23 s
[Task 1/25] Current/Best: 6.18/ 17.86 GFLOPS | Progress: (8/20) | 9.20 s
[Task 1/25] Current/Best: 11.36/ 22.62 GFLOPS | Progress: (12/20) | 11.63 s
[Task 1/25] Current/Best: 16.78/ 22.62 GFLOPS | Progress: (16/20) | 13.30 s
[Task 1/25] Current/Best: 11.50/ 23.90 GFLOPS | Progress: (20/20) | 15.07 s Done.
-
[Task 2/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 2/25] Current/Best: 12.37/ 12.47 GFLOPS | Progress: (4/20) | 3.74 s
[Task 2/25] Current/Best: 12.71/ 18.76 GFLOPS | Progress: (8/20) | 5.07 s
[Task 2/25] Current/Best: 21.17/ 21.17 GFLOPS | Progress: (12/20) | 6.38 s
[Task 2/25] Current/Best: 11.37/ 21.17 GFLOPS | Progress: (16/20) | 7.62 s
[Task 2/25] Current/Best: 18.46/ 21.17 GFLOPS | Progress: (20/20) | 9.17 s Done.
-
[Task 3/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 3/25] Current/Best: 1.65/ 10.29 GFLOPS | Progress: (4/20) | 5.82 s
[Task 3/25] Current/Best: 15.64/ 17.17 GFLOPS | Progress: (8/20) | 7.75 s
[Task 3/25] Current/Best: 15.31/ 17.17 GFLOPS | Progress: (12/20) | 9.47 s
[Task 3/25] Current/Best: 6.87/ 23.72 GFLOPS | Progress: (16/20) | 11.45 s
[Task 3/25] Current/Best: 11.24/ 23.72 GFLOPS | Progress: (20/20) | 15.95 s Done.
-
[Task 4/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 4/25] Current/Best: 9.23/ 18.78 GFLOPS | Progress: (4/20) | 2.41 s
[Task 4/25] Current/Best: 6.68/ 18.78 GFLOPS | Progress: (8/20) | 6.69 s
[Task 4/25] Current/Best: 21.67/ 21.67 GFLOPS | Progress: (12/20) | 11.15 s
[Task 4/25] Current/Best: 15.41/ 21.67 GFLOPS | Progress: (16/20) | 13.38 s
[Task 4/25] Current/Best: 12.44/ 21.67 GFLOPS | Progress: (20/20) | 15.38 s Done.
-
[Task 5/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 5/25] Current/Best: 9.16/ 10.01 GFLOPS | Progress: (4/20) | 2.60 s
[Task 5/25] Current/Best: 11.30/ 11.30 GFLOPS | Progress: (8/20) | 4.73 s
[Task 5/25] Current/Best: 11.89/ 18.34 GFLOPS | Progress: (12/20) | 7.86 s
[Task 5/25] Current/Best: 11.71/ 21.70 GFLOPS | Progress: (16/20) | 9.29 s
[Task 5/25] Current/Best: 12.20/ 21.70 GFLOPS | Progress: (20/20) | 11.16 s Done.
-
[Task 6/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 6/25] Current/Best: 12.15/ 20.26 GFLOPS | Progress: (4/20) | 3.93 s
[Task 6/25] Current/Best: 19.09/ 20.26 GFLOPS | Progress: (8/20) | 5.69 s
[Task 6/25] Current/Best: 13.40/ 20.26 GFLOPS | Progress: (12/20) | 7.66 s
[Task 6/25] Current/Best: 19.91/ 20.26 GFLOPS | Progress: (16/20) | 9.89 s
[Task 6/25] Current/Best: 3.68/ 20.26 GFLOPS | Progress: (20/20) | 12.46 s Done.
-
[Task 7/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 7/25] Current/Best: 9.95/ 12.27 GFLOPS | Progress: (4/20) | 3.66 s
[Task 7/25] Current/Best: 19.68/ 20.29 GFLOPS | Progress: (8/20) | 5.18 s
[Task 7/25] Current/Best: 16.37/ 20.29 GFLOPS | Progress: (12/20) | 7.06 s
[Task 7/25] Current/Best: 12.33/ 20.29 GFLOPS | Progress: (16/20) | 9.14 s
[Task 7/25] Current/Best: 6.13/ 20.29 GFLOPS | Progress: (20/20) | 11.63 s Done.
-
[Task 8/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 8/25] Current/Best: 9.73/ 13.29 GFLOPS | Progress: (4/20) | 2.90 s
[Task 8/25] Current/Best: 9.40/ 13.29 GFLOPS | Progress: (8/20) | 7.64 s
[Task 8/25] Current/Best: 12.86/ 13.29 GFLOPS | Progress: (12/20) | 13.65 s
[Task 8/25] Current/Best: 19.24/ 19.24 GFLOPS | Progress: (16/20) | 15.74 s
[Task 8/25] Current/Best: 19.43/ 19.43 GFLOPS | Progress: (20/20) | 22.28 s Done.
-
[Task 9/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 9/25] Current/Best: 14.52/ 14.52 GFLOPS | Progress: (4/20) | 11.95 s
[Task 9/25] Current/Best: 21.30/ 21.30 GFLOPS | Progress: (8/20) | 13.74 s
[Task 9/25] Current/Best: 8.13/ 21.30 GFLOPS | Progress: (12/20) | 16.07 s
[Task 9/25] Current/Best: 18.26/ 21.30 GFLOPS | Progress: (16/20) | 18.70 s
[Task 9/25] Current/Best: 9.22/ 21.30 GFLOPS | Progress: (20/20) | 26.05 s
[Task 10/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 10/25] Current/Best: 18.47/ 18.47 GFLOPS | Progress: (4/20) | 2.56 s
[Task 10/25] Current/Best: 15.72/ 18.47 GFLOPS | Progress: (8/20) | 4.13 s
[Task 10/25] Current/Best: 11.45/ 19.09 GFLOPS | Progress: (12/20) | 5.64 s
[Task 10/25] Current/Best: 19.43/ 20.57 GFLOPS | Progress: (16/20) | 6.74 s
[Task 10/25] Current/Best: 8.64/ 20.57 GFLOPS | Progress: (20/20
) | 8.29 s Done.
-
[Task 11/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 11/25] Current/Best: 11.07/ 18.54 GFLOPS | Progress: (4/20) | 3.32 s
[Task 11/25] Current/Best: 15.19/ 18.54 GFLOPS | Progress: (8/20) | 6.06 s
[Task 11/25] Current/Best: 16.14/ 18.54 GFLOPS | Progress: (12/20) | 8.09 s
[Task 11/25] Current/Best: 12.05/ 20.61 GFLOPS | Progress: (16/20) | 10.91 s
[Task 11/25] Current/Best: 18.94/ 20.61 GFLOPS | Progress: (20/20) | 12.95 s Done.
-
[Task 12/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 12/25] Current/Best: 7.80/ 18.13 GFLOPS | Progress: (4/20) | 5.29 s
[Task 12/25] Current/Best: 5.06/ 18.13 GFLOPS | Progress: (8/20) | 8.97 s
[Task 12/25] Current/Best: 19.10/ 19.10 GFLOPS | Progress: (12/20) | 10.97 s
[Task 12/25] Current/Best: 15.46/ 19.10 GFLOPS | Progress: (16/20) | 13.75 s
[Task 12/25] Current/Best: 15.32/ 19.10 GFLOPS | Progress: (20/20) | 15.68 s Done.
-
[Task 13/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 13/25] Current/Best: 8.70/ 17.48 GFLOPS | Progress: (4/20) | 3.64 s
[Task 13/25] Current/Best: 15.78/ 20.87 GFLOPS | Progress: (8/20) | 6.06 s
[Task 13/25] Current/Best: 19.06/ 21.80 GFLOPS | Progress: (12/20) | 8.92 s
[Task 13/25] Current/Best: 12.46/ 21.80 GFLOPS | Progress: (16/20) | 12.28 s
[Task 13/25] Current/Best: 17.75/ 21.80 GFLOPS | Progress: (20/20) | 14.56 s Done.
-
[Task 14/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 14/25] Current/Best: 12.32/ 13.53 GFLOPS | Progress: (4/20) | 3.26 s
[Task 14/25] Current/Best: 6.17/ 13.53 GFLOPS | Progress: (8/20) | 5.45 s
[Task 14/25] Current/Best: 19.57/ 19.57 GFLOPS | Progress: (12/20) | 8.02 s
[Task 14/25] Current/Best: 16.32/ 19.57 GFLOPS | Progress: (16/20) | 9.65 s Done.
-
[Task 14/25] Current/Best: 17.26/ 19.57 GFLOPS | Progress: (20/20) | 11.36 s
[Task 15/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 15/25] Current/Best: 15.87/ 17.49 GFLOPS | Progress: (4/20) | 2.69 s
[Task 15/25] Current/Best: 12.87/ 18.02 GFLOPS | Progress: (8/20) | 3.99 s
[Task 15/25] Current/Best: 10.08/ 21.93 GFLOPS | Progress: (12/20) | 6.08 s
[Task 15/25] Current/Best: 20.47/ 21.93 GFLOPS | Progress: (16/20) | 9.00 s
[Task 15/25] Current/Best: 9.61/ 21.93 GFLOPS | Progress: (20/20) | 9.97 s
[Task 16/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 16/25] Current/Best: 19.20/ 19.20 GFLOPS | Progress: (4/20) | 2.99 s
[Task 16/25] Current/Best: 3.05/ 19.20 GFLOPS | Progress: (8/20) | 4.62 s
[Task 16/25] Current/Best: 18.20/ 19.63 GFLOPS | Progress: (12/20) | 5.84 s
[Task 16/25] Current/Best: 18.42/ 19.63 GFLOPS | Progress: (16/20) |
7.19 s
[Task 16/25] Current/Best: 10.13/ 21.71 GFLOPS | Progress: (20/20) | 9.21 s Done.
-
[Task 17/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 17/25] Current/Best: 12.75/ 16.46 GFLOPS | Progress: (4/20) | 4.66 s
[Task 17/25] Current/Best: 12.84/ 23.14 GFLOPS | Progress: (8/20) | 7.47 s
[Task 17/25] Current/Best: 16.78/ 23.14 GFLOPS | Progress: (12/20) | 9.55 s
[Task 17/25] Current/Best: 16.70/ 23.14 GFLOPS | Progress: (16/20) | 11.66 s
[Task 17/25] Current/Best: 10.13/ 23.14 GFLOPS | Progress: (20/20) | 13.75 s Done.
-
[Task 18/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 18/25] Current/Best: 10.30/ 17.21 GFLOPS | Progress: (4/20) | 3.67 s
[Task 18/25] Current/Best: 10.72/ 17.21 GFLOPS | Progress: (8/20) | 7.08 s
[Task 18/25] Current/Best: 19.64/ 19.64 GFLOPS | Progress: (12/20) | 9.00 s
[Task 18/25] Current/Best: 10.09/ 19.64 GFLOPS | Progress: (16/20) | 12.59 s
[Task 18/25] Current/Best: 21.09/ 21.09 GFLOPS | Progress: (20/20) | 14.11 s Done.
-
[Task 19/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 19/25] Current/Best: 7.37/ 20.19 GFLOPS | Progress: (4/20) | 5.88 s
[Task 19/25] Current/Best: 2.73/ 20.19 GFLOPS | Progress: (8/20) | 9.11 s
[Task 19/25] Current/Best: 19.46/ 21.34 GFLOPS | Progress: (12/20) | 11.92 s
[Task 19/25] Current/Best: 13.92/ 21.57 GFLOPS | Progress: (16/20) | 14.77 s
[Task 19/25] Current/Best: 2.74/ 23.04 GFLOPS | Progress: (20/20) | 17.58 s Done.
-
[Task 20/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 20/25] Current/Best: 9.04/ 15.42 GFLOPS | Progress: (4/20) | 3.27 s Done.
+
[Task 1/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 1/25] Current/Best: 17.53/ 17.53 GFLOPS | Progress: (4/20) | 6.34 s
[Task 1/25] Current/Best: 6.11/ 17.53 GFLOPS | Progress: (8/20) | 9.39 s
[Task 1/25] Current/Best: 11.16/ 22.20 GFLOPS | Progress: (12/20) | 11.90 s
[Task 1/25] Current/Best: 16.50/ 22.37 GFLOPS | Progress: (16/20) | 13.58 s
[Task 1/25] Current/Best: 11.34/ 23.67 GFLOPS | Progress: (20/20) | 15.35 s Done.
+
[Task 2/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 2/25] Current/Best: 12.18/ 12.51 GFLOPS | Progress: (4/20) | 3.75 s
[Task 2/25] Current/Best: 12.45/ 17.75 GFLOPS | Progress: (8/20) | 5.04 s
[Task 2/25] Current/Best: 20.72/ 20.72 GFLOPS | Progress: (12/20) | 6.35 s
[Task 2/25] Current/Best: 11.56/ 20.72 GFLOPS | Progress: (16/20) | 7.62 s
[Task 2/25] Current/Best: 18.22/ 20.72 GFLOPS | Progress: (20/20) | 9.17 s Done.
+
[Task 3/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 3/25] Current/Best: 1.63/ 10.13 GFLOPS | Progress: (4/20) | 5.85 s
[Task 3/25] Current/Best: 15.39/ 16.76 GFLOPS | Progress: (8/20) | 7.79 s
[Task 3/25] Current/Best: 15.06/ 16.76 GFLOPS | Progress: (12/20) | 9.51 s
[Task 3/25] Current/Best: 6.84/ 23.06 GFLOPS | Progress: (16/20) | 11.48 s
[Task 3/25] Current/Best: 11.09/ 23.06 GFLOPS | Progress: (20/20) | 16.09 s Done.
+
[Task 4/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 4/25] Current/Best: 8.88/ 18.39 GFLOPS | Progress: (4/20) | 2.40 s
[Task 4/25] Current/Best: 6.58/ 18.39 GFLOPS | Progress: (8/20) | 6.72 s
[Task 4/25] Current/Best: 21.09/ 21.09 GFLOPS | Progress: (12/20) | 11.20 s
[Task 4/25] Current/Best: 16.16/ 21.09 GFLOPS | Progress: (16/20) | 13.42 s
[Task 4/25] Current/Best: 12.70/ 21.09 GFLOPS | Progress: (20/20) | 15.38 s Done.
+
[Task 5/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 5/25] Current/Best: 8.43/ 9.71 GFLOPS | Progress: (4/20) | 2.61 s
[Task 5/25] Current/Best: 11.09/ 11.19 GFLOPS | Progress: (8/20) | 4.69 s
[Task 5/25] Current/Best: 11.76/ 18.12 GFLOPS | Progress: (12/20) | 7.79 s
[Task 5/25] Current/Best: 11.51/ 22.53 GFLOPS | Progress: (16/20) | 9.23 s
[Task 5/25] Current/Best: 12.10/ 22.53 GFLOPS | Progress: (20/20) | 11.10 s Done.
+
[Task 6/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 6/25] Current/Best: 12.00/ 19.89 GFLOPS | Progress: (4/20) | 3.99 s
[Task 6/25] Current/Best: 18.94/ 19.89 GFLOPS | Progress: (8/20) | 5.77 s
[Task 6/25] Current/Best: 13.26/ 19.89 GFLOPS | Progress: (12/20) | 7.75 s
[Task 6/25] Current/Best: 19.69/ 19.89 GFLOPS | Progress: (16/20) | 10.01 s
[Task 6/25] Current/Best: 3.69/ 19.89 GFLOPS | Progress: (20/20) | 12.58 s Done.
+
[Task 7/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 7/25] Current/Best: 9.79/ 12.12 GFLOPS | Progress: (4/20) | 3.71 s
[Task 7/25] Current/Best: 19.64/ 19.64 GFLOPS | Progress: (8/20) | 5.26 s
[Task 7/25] Current/Best: 13.18/ 19.64 GFLOPS | Progress: (12/20) | 7.20 s
[Task 7/25] Current/Best: 12.11/ 20.12 GFLOPS | Progress: (16/20) | 9.29 s
[Task 7/25] Current/Best: 6.14/ 20.51 GFLOPS | Progress: (20/20) | 11.81 s Done.
+
[Task 8/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 8/25] Current/Best: 9.58/ 13.53 GFLOPS | Progress: (4/20) | 2.95 s
[Task 8/25] Current/Best: 9.20/ 13.53 GFLOPS | Progress: (8/20) | 7.73 s
[Task 8/25] Current/Best: 12.67/ 13.53 GFLOPS | Progress: (12/20) | 13.81 s
[Task 8/25] Current/Best: 18.94/ 18.94 GFLOPS | Progress: (16/20) | 15.95 s
[Task 8/25] Current/Best: 18.71/ 18.94 GFLOPS | Progress: (20/20) | 22.47 s Done.
+
[Task 9/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 9/25] Current/Best: 14.37/ 14.37 GFLOPS | Progress: (4/20) | 11.95 s
[Task 9/25] Current/Best: 21.17/ 21.17 GFLOPS | Progress: (8/20) | 13.70 s
[Task 9/25] Current/Best: 8.02/ 21.17 GFLOPS | Progress: (12/20) | 16.08 s
[Task 9/25] Current/Best: 17.92/ 21.17 GFLOPS | Progress: (16/20) | 18.71 s
[Task 9/25] Current/Best: 9.05/ 21.17 GFLOPS | Progress: (20/20) | 26.43 s
[Task 10/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 10/25] Current/Best: 18.08/ 18.08 GFLOPS | Progress: (4/20) | 2.56 s
[Task 10/25] Current/Best: 15.64/ 18.08 GFLOPS | Progress: (8/20) | 4.19 s
[Task 10/25] Current/Best: 11.45/ 19.02 GFLOPS | Progress: (12/20) | 5.73 s
[Task 10/25] Current/Best: 19.15/ 20.20 GFLOPS | Progress: (16/20) | 6.83 s
[Task 10/25] Current/Best: 8.72/ 20.20 GFLOPS | Progress: (20/20
) | 8.37 s Done.
+
[Task 11/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 11/25] Current/Best: 10.83/ 18.18 GFLOPS | Progress: (4/20) | 3.39 s
[Task 11/25] Current/Best: 14.89/ 18.18 GFLOPS | Progress: (8/20) | 6.18 s
[Task 11/25] Current/Best: 15.94/ 18.18 GFLOPS | Progress: (12/20) | 8.24 s
[Task 11/25] Current/Best: 11.88/ 20.13 GFLOPS | Progress: (16/20) | 11.06 s
[Task 11/25] Current/Best: 18.60/ 20.42 GFLOPS | Progress: (20/20) | 13.14 s Done.
+
[Task 12/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 12/25] Current/Best: 7.81/ 17.97 GFLOPS | Progress: (4/20) | 5.30 s
[Task 12/25] Current/Best: 5.06/ 17.97 GFLOPS | Progress: (8/20) | 9.01 s
[Task 12/25] Current/Best: 18.80/ 18.80 GFLOPS | Progress: (12/20) | 11.03 s
[Task 12/25] Current/Best: 15.34/ 18.80 GFLOPS | Progress: (16/20) | 13.84 s
[Task 12/25] Current/Best: 15.16/ 18.80 GFLOPS | Progress: (20/20) | 15.80 s Done.
+
[Task 13/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 13/25] Current/Best: 8.36/ 17.25 GFLOPS | Progress: (4/20) | 3.68 s
[Task 13/25] Current/Best: 15.11/ 20.55 GFLOPS | Progress: (8/20) | 6.14 s
[Task 13/25] Current/Best: 18.81/ 21.57 GFLOPS | Progress: (12/20) | 9.01 s
[Task 13/25] Current/Best: 12.21/ 21.57 GFLOPS | Progress: (16/20) | 12.40 s
[Task 13/25] Current/Best: 17.11/ 21.57 GFLOPS | Progress: (20/20) | 14.73 s Done.
+
[Task 14/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 14/25] Current/Best: 12.10/ 13.10 GFLOPS | Progress: (4/20) | 3.38 s
[Task 14/25] Current/Best: 6.06/ 13.19 GFLOPS | Progress: (8/20) | 5.59 s
[Task 14/25] Current/Best: 19.44/ 19.44 GFLOPS | Progress: (12/20) | 8.17 s
[Task 14/25] Current/Best: 15.77/ 19.44 GFLOPS | Progress: (16/20) | 9.83 s Done.
+
[Task 14/25] Current/Best: 16.74/ 19.44 GFLOPS | Progress: (20/20) | 11.55 s
[Task 15/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 15/25] Current/Best: 15.63/ 17.27 GFLOPS | Progress: (4/20) | 2.73 s
[Task 15/25] Current/Best: 12.69/ 17.90 GFLOPS | Progress: (8/20) | 4.07 s
[Task 15/25] Current/Best: 9.86/ 21.49 GFLOPS | Progress: (12/20) | 6.13 s
[Task 15/25] Current/Best: 20.03/ 21.49 GFLOPS | Progress: (16/20) | 9.58 s
[Task 15/25] Current/Best: 9.53/ 21.49 GFLOPS | Progress: (20/20) | 10.60 s
[Task 16/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 16/25] Current/Best: 17.66/ 17.66 GFLOPS | Progress: (4/20) | 2.99 s
[Task 16/25] Current/Best: 2.99/ 17.66 GFLOPS | Progress: (8/20) | 4.62 s
[Task 16/25] Current/Best: 18.60/ 19.44 GFLOPS | Progress: (12/20) | 5.84 s
[Task 16/25] Current/Best: 17.98/ 19.44 GFLOPS | Progress: (16/20) |
7.18 s
[Task 16/25] Current/Best: 10.04/ 21.47 GFLOPS | Progress: (20/20) | 9.22 s Done.
+
[Task 17/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 17/25] Current/Best: 12.73/ 16.00 GFLOPS | Progress: (4/20) | 4.72 s
[Task 17/25] Current/Best: 12.68/ 22.03 GFLOPS | Progress: (8/20) | 7.58 s
[Task 17/25] Current/Best: 16.48/ 22.03 GFLOPS | Progress: (12/20) | 9.71 s
[Task 17/25] Current/Best: 16.44/ 22.03 GFLOPS | Progress: (16/20) | 11.89 s
[Task 17/25] Current/Best: 10.00/ 22.03 GFLOPS | Progress: (20/20) | 14.01 s Done.
+
[Task 18/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 18/25] Current/Best: 10.93/ 16.74 GFLOPS | Progress: (4/20) | 3.70 s
[Task 18/25] Current/Best: 10.53/ 16.74 GFLOPS | Progress: (8/20) | 7.14 s
[Task 18/25] Current/Best: 19.37/ 19.37 GFLOPS | Progress: (12/20) | 9.10 s
[Task 18/25] Current/Best: 10.20/ 19.37 GFLOPS | Progress: (16/20) | 12.65 s
[Task 18/25] Current/Best: 20.53/ 20.53 GFLOPS | Progress: (20/20) | 14.17 s Done.
+
[Task 19/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 19/25] Current/Best: 7.13/ 19.86 GFLOPS | Progress: (4/20) | 6.03 s
[Task 19/25] Current/Best: 2.69/ 19.86 GFLOPS | Progress: (8/20) | 9.32 s
[Task 19/25] Current/Best: 18.93/ 20.33 GFLOPS | Progress: (12/20) | 12.16 s
[Task 19/25] Current/Best: 12.85/ 21.33 GFLOPS | Progress: (16/20) | 15.09 s
[Task 19/25] Current/Best: 2.70/ 22.59 GFLOPS | Progress: (20/20) | 17.93 s Done.
+
[Task 20/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 20/25] Current/Best: 9.31/ 15.25 GFLOPS | Progress: (4/20) | 3.27 s Done.
Done.
-
[Task 20/25] Current/Best: 10.01/ 15.42 GFLOPS | Progress: (8/20) | 6.70 s
[Task 20/25] Current/Best: 2.35/ 15.42 GFLOPS | Progress: (12/20) | 10.57 s
[Task 20/25] Current/Best: 11.21/ 15.42 GFLOPS | Progress: (16/20) | 14.07 s
[Task 20/25] Current/Best: 11.41/ 22.32 GFLOPS | Progress: (20/20) | 16.15 s
[Task 21/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 21/25] Current/Best: 6.44/ 17.89 GFLOPS | Progress: (4/20) | 3.18 s
[Task 21/25] Current/Best: 14.88/ 17.89 GFLOPS | Progress: (8/20) | 4.71 s
[Task 21/25] Current/Best: 1.63/ 17.89 GFLOPS | Progress: (12/20) | 6.83 s
[Task 21/25] Current/Best: 16.13/ 17.89 GFLOPS | Progress: (16/20) | 10.24 s
[Task 21/25] Current/Best: 4.50/ 17.89 GFLOPS | Progress: (20/20) | 17.10 s
[Task 22/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 22/25] Current/Best: 2.74/ 16.86 GFLOPS | Progress: (4/20
) | 2.66 s
[Task 22/25] Current/Best: 8.89/ 21.48 GFLOPS | Progress: (8/20) | 4.63 s
[Task 22/25] Current/Best: 20.11/ 21.48 GFLOPS | Progress: (12/20) | 6.89 s
[Task 22/25] Current/Best: 15.63/ 21.48 GFLOPS | Progress: (16/20) | 8.92 s
[Task 22/25] Current/Best: 12.67/ 21.48 GFLOPS | Progress: (20/20) | 10.64 s Done.
-
[Task 23/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 23/25] Current/Best: 16.97/ 20.43 GFLOPS | Progress: (4/20) | 3.24 s
[Task 23/25] Current/Best: 13.64/ 20.43 GFLOPS | Progress: (8/20) | 6.57 s
[Task 23/25] Current/Best: 20.99/ 22.17 GFLOPS | Progress: (12/20) | 8.36 s
[Task 23/25] Current/Best: 6.66/ 22.17 GFLOPS | Progress: (16/20) | 15.27 s
[Task 23/25] Current/Best: 7.79/ 22.17 GFLOPS | Progress: (20/20) | 19.41 s Done.
-
[Task 24/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 24/25] Current/Best: 8.27/ 8.27 GFLOPS | Progress: (4/20) | 11.78 s
[Task 24/25] Current/Best: 2.08/ 8.27 GFLOPS | Progress: (8/20) | 22.77 s
[Task 24/25] Current/Best: 4.05/ 8.27 GFLOPS | Progress: (12/20) | 34.30 s Done.
-
[Task 24/25] Current/Best: 5.49/ 8.93 GFLOPS | Progress: (16/20) | 39.52 s
[Task 24/25] Current/Best: 3.02/ 8.93 GFLOPS | Progress: (20/20) | 45.37 s Done.
-
[Task 25/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 25/25] Current/Best: 1.57/ 2.73 GFLOPS | Progress: (4/20) | 11.60 s
[Task 25/25] Current/Best: 5.99/ 8.15 GFLOPS | Progress: (8/20) | 22.88 s
[Task 25/25] Current/Best: 6.19/ 8.15 GFLOPS | Progress: (12/20) | 34.31 s
[Task 25/25] Current/Best: 5.93/ 8.65 GFLOPS | Progress: (16/20) | 36.16 s
[Task 25/25] Current/Best: 2.89/ 8.88 GFLOPS | Progress: (20/20) | 46.84 s
+
[Task 20/25] Current/Best: 9.72/ 15.25 GFLOPS | Progress: (8/20) | 6.76 s
[Task 20/25] Current/Best: 2.32/ 15.25 GFLOPS | Progress: (12/20) | 10.80 s
[Task 20/25] Current/Best: 10.83/ 15.25 GFLOPS | Progress: (16/20) | 14.49 s
[Task 20/25] Current/Best: 11.17/ 21.85 GFLOPS | Progress: (20/20) | 16.59 s
[Task 21/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 21/25] Current/Best: 6.36/ 17.74 GFLOPS | Progress: (4/20) | 3.22 s
[Task 21/25] Current/Best: 14.58/ 17.74 GFLOPS | Progress: (8/20) | 4.80 s
[Task 21/25] Current/Best: 1.61/ 17.74 GFLOPS | Progress: (12/20) | 6.94 s
[Task 21/25] Current/Best: 15.89/ 17.74 GFLOPS | Progress: (16/20) | 10.37 s
[Task 21/25] Current/Best: 4.47/ 17.74 GFLOPS | Progress: (20/20) | 17.36 s
[Task 22/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 22/25] Current/Best: 2.70/ 16.77 GFLOPS | Progress: (4/20
) | 2.68 s
[Task 22/25] Current/Best: 8.70/ 20.23 GFLOPS | Progress: (8/20) | 4.66 s
[Task 22/25] Current/Best: 19.87/ 20.23 GFLOPS | Progress: (12/20) | 6.94 s
[Task 22/25] Current/Best: 15.43/ 20.23 GFLOPS | Progress: (16/20) | 8.99 s
[Task 22/25] Current/Best: 13.53/ 20.23 GFLOPS | Progress: (20/20) | 10.74 s Done.
+
[Task 23/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 23/25] Current/Best: 16.47/ 19.70 GFLOPS | Progress: (4/20) | 3.26 s
[Task 23/25] Current/Best: 13.72/ 19.86 GFLOPS | Progress: (8/20) | 6.63 s
[Task 23/25] Current/Best: 20.70/ 21.54 GFLOPS | Progress: (12/20) | 8.41 s
[Task 23/25] Current/Best: 6.61/ 21.54 GFLOPS | Progress: (16/20) | 15.40 s
[Task 23/25] Current/Best: 7.59/ 21.54 GFLOPS | Progress: (20/20) | 19.63 s Done.
+
[Task 24/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 24/25] Current/Best: 8.28/ 8.28 GFLOPS | Progress: (4/20) | 11.79 s
[Task 24/25] Current/Best: 3.19/ 8.28 GFLOPS | Progress: (8/20) | 23.03 s
[Task 24/25] Current/Best: 3.59/ 8.28 GFLOPS | Progress: (12/20) | 33.76 s Done.
+
[Task 24/25] Current/Best: 5.67/ 8.70 GFLOPS | Progress: (16/20) | 39.20 s
[Task 24/25] Current/Best: 2.95/ 8.70 GFLOPS | Progress: (20/20) | 45.05 s Done.
+
[Task 25/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 25/25] Current/Best: 1.55/ 2.70 GFLOPS | Progress: (4/20) | 11.58 s
[Task 25/25] Current/Best: 5.71/ 7.89 GFLOPS | Progress: (8/20) | 22.85 s
[Task 25/25] Current/Best: 5.86/ 7.89 GFLOPS | Progress: (12/20) | 34.23 s
[Task 25/25] Current/Best: 5.82/ 8.39 GFLOPS | Progress: (16/20) | 36.08 s
[Task 25/25] Current/Best: 2.88/ 8.39 GFLOPS | Progress: (20/20) | 46.76 s
@@ -690,8 +690,8 @@ Verify that the optimized model runs and produces the same results:
.. code-block:: none
- class='n02123045 tabby, tabby cat' with probability=0.621104
- class='n02123159 tiger cat' with probability=0.356378
+ class='n02123045 tabby, tabby cat' with probability=0.621105
+ class='n02123159 tiger cat' with probability=0.356377
class='n02124075 Egyptian cat' with probability=0.019712
class='n02129604 tiger, Panthera tigris' with probability=0.001215
class='n04040759 radiator' with probability=0.000262
@@ -748,8 +748,8 @@ improvement in comparing the optimized model to the unoptimized model.
.. code-block:: none
- optimized: {'mean': 403.80653754022205, 'median': 403.70924704911886, 'std': 0.5867979772696545}
- unoptimized: {'mean': 501.64646821969654, 'median': 501.2921533503686, 'std': 1.4021516095849}
+ optimized: {'mean': 405.86647802999323, 'median': 405.71762745000797, 'std': 0.6242720580978915}
+ unoptimized: {'mean': 510.73006595000606, 'median': 511.101387650001, 'std': 1.6193648739236104}
@@ -772,7 +772,7 @@ profiling/benchmarking.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 10 minutes 7.511 seconds)
+ **Total running time of the script:** ( 10 minutes 24.692 seconds)
.. _sphx_glr_download_tutorial_autotvm_relay_x86.py:
diff --git a/docs/_sources/tutorial/cross_compilation_and_rpc.rst.txt b/docs/_sources/tutorial/cross_compilation_and_rpc.rst.txt
index e62344bea..2d2b829f3 100644
--- a/docs/_sources/tutorial/cross_compilation_and_rpc.rst.txt
+++ b/docs/_sources/tutorial/cross_compilation_and_rpc.rst.txt
@@ -282,7 +282,7 @@ device and returns the measured cost. Network overhead is excluded.
.. code-block:: none
- 1.295e-07 secs/op
+ 1.246e-07 secs/op
diff --git a/docs/_sources/tutorial/intro_topi.rst.txt b/docs/_sources/tutorial/intro_topi.rst.txt
index 4a86b730d..e70d7d5a2 100644
--- a/docs/_sources/tutorial/intro_topi.rst.txt
+++ b/docs/_sources/tutorial/intro_topi.rst.txt
@@ -263,7 +263,7 @@ As you can see, scheduled stages of computation have been accumulated and we can
.. code-block:: none
- [stage(a, placeholder(a, 0x1ff933a0)), stage(b, placeholder(b, 0x886c770)), stage(T_add, compute(T_add, body=[(a[ax0, ax1, ax2] + b[ax1, ax2])], axis=[iter_var(ax0, range(min=0, ext=100)), iter_var(ax1, range(min=0, ext=10)), iter_var(ax2, range(min=0, ext=10))], reduce_axis=[], tag=broadcast, attrs={})), stage(T_multiply, compute(T_multiply, body=[(a[ax0, ax1, ax2]*b[ax1, ax2])], axis=[iter_var(ax0, range(min=0, ext=100)), iter_var(ax1, range(min=0, ext=10)), iter_var(ax2, range(min [...]
+ [stage(a, placeholder(a, 0xd8fe4c0)), stage(b, placeholder(b, 0x19dd8a00)), stage(T_add, compute(T_add, body=[(a[ax0, ax1, ax2] + b[ax1, ax2])], axis=[iter_var(ax0, range(min=0, ext=100)), iter_var(ax1, range(min=0, ext=10)), iter_var(ax2, range(min=0, ext=10))], reduce_axis=[], tag=broadcast, attrs={})), stage(T_multiply, compute(T_multiply, body=[(a[ax0, ax1, ax2]*b[ax1, ax2])], axis=[iter_var(ax0, range(min=0, ext=100)), iter_var(ax1, range(min=0, ext=10)), iter_var(ax2, range(min [...]
diff --git a/docs/_sources/tutorial/sg_execution_times.rst.txt b/docs/_sources/tutorial/sg_execution_times.rst.txt
index 90e4925a7..980e04cf8 100644
--- a/docs/_sources/tutorial/sg_execution_times.rst.txt
+++ b/docs/_sources/tutorial/sg_execution_times.rst.txt
@@ -5,32 +5,32 @@
Computation times
=================
-**13:19.324** total execution time for **tutorial** files:
+**13:24.169** total execution time for **tutorial** files:
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``) | 10:07.511 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``) | 10:24.692 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 01:15.309 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 01:02.104 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``) | 01:00.733 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``) | 01:00.762 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``) | 00:30.058 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``) | 00:30.965 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``) | 00:24.116 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``) | 00:24.276 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``) | 00:00.748 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``) | 00:00.694 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``) | 00:00.693 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``) | 00:00.525 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.149 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.144 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_introduction.py` (``introduction.py``) | 00:00.004 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_introduction.py` (``introduction.py``) | 00:00.005 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_tutorial_uma.py` (``uma.py``) | 00:00.001 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
+| :ref:`sphx_glr_tutorial_tvmc_command_line_driver.py` (``tvmc_command_line_driver.py``) | 00:00.001 | 0.0 MB |
++------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_tutorial_install.py` (``install.py``) | 00:00.001 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_tutorial_tvmc_python.py` (``tvmc_python.py``) | 00:00.001 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tvmc_command_line_driver.py` (``tvmc_command_line_driver.py``) | 00:00.001 | 0.0 MB |
-+------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/tutorial/tensor_expr_get_started.rst.txt b/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
index 28e98bb80..607d1c7aa 100644
--- a/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
+++ b/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
@@ -301,7 +301,7 @@ helper function to run a profile of the TVM generated code.
.. code-block:: none
- Numpy running time: 0.000007
+ Numpy running time: 0.000008
naive: 0.000007
@@ -460,7 +460,7 @@ factor to be the number of threads on your CPU.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
- vector: 0.000024
+ vector: 0.000025
@main = primfn(A_1: handle, B_1: handle, C_1: handle) -> ()
attr = {"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True}
buffers = {A: Buffer(A_2: Pointer(float32), float32, [(stride: int32*n: int32)], [], type="auto"),
@@ -512,10 +512,10 @@ We can now compare the different schedules
.. code-block:: none
Operator Timing Performance
- numpy 7.263150182552636e-06 1.0
- naive 6.6703e-06 0.9183756128330148
- parallel 6.0228999999999995e-06 0.8292407355789041
- vector 2.3852799999999996e-05 3.284084646535138
+ numpy 7.505210000999796e-06 1.0
+ naive 6.7073e-06 0.8936858527751385
+ parallel 6.0147e-06 0.8014032917398395
+ vector 2.46064e-05 3.2785758155630678
@@ -936,7 +936,7 @@ matrix multiplication.
.. code-block:: none
- Numpy running time: 0.017319
+ Numpy running time: 0.017868
@@ -996,7 +996,7 @@ optimizations.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
- none: 3.444058
+ none: 3.433973
@@ -1101,7 +1101,7 @@ schedule.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
- blocking: 0.292449
+ blocking: 0.292093
@@ -1199,7 +1199,7 @@ already cache friendly from our previous optimizations.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
- vectorization: 0.326932
+ vectorization: 0.331134
@main = primfn(A_1: handle, B_1: handle, C_1: handle) -> ()
attr = {"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True}
buffers = {A: Buffer(A_2: Pointer(float32), float32, [1048576], []),
@@ -1275,7 +1275,7 @@ more cache friendly.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
- loop permutation: 0.112778
+ loop permutation: 0.115847
@main = primfn(A_1: handle, B_1: handle, C_1: handle) -> ()
attr = {"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True}
buffers = {A: Buffer(A_2: Pointer(float32), float32, [1048576], []),
@@ -1376,7 +1376,7 @@ optimized schedule.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
- array packing: 0.106524
+ array packing: 0.108878
@main = primfn(A_1: handle, B_1: handle, C_1: handle) -> ()
attr = {"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True}
buffers = {A: Buffer(A_2: Pointer(float32), float32, [1048576], []),
@@ -1471,7 +1471,7 @@ to `C` when all the block results are ready.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
- block caching: 0.108353
+ block caching: 0.111419
@main = primfn(A_1: handle, B_1: handle, C_1: handle) -> ()
attr = {"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True}
buffers = {A: Buffer(A_2: Pointer(float32), float32, [1048576], []),
@@ -1559,7 +1559,7 @@ of thread-level parallelization.
/workspace/python/tvm/driver/build_module.py:267: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
"target_host parameter is going to be deprecated. "
- parallelization: 0.143483
+ parallelization: 0.146707
@main = primfn(A_1: handle, B_1: handle, C_1: handle) -> ()
attr = {"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True}
buffers = {A: Buffer(A_2: Pointer(float32), float32, [1048576], []),
@@ -1640,13 +1640,13 @@ working, we can compare the results.
.. code-block:: none
Operator Timing Performance
- none 3.4440584045000002 1.0
- blocking 0.29244923429999997 0.08491413325566324
- vectorization 0.32693231959999997 0.09492647371276597
- loop permutation 0.1127775773 0.03274554727429855
- array packing 0.10652408660000001 0.030929814215930787
- block caching 0.10835264 0.031460744062419686
- parallelization 0.14348306900000002 0.04166104407884759
+ none 3.4339734321999997 1.0
+ blocking 0.29209253020000003 0.0850596360068135
+ vectorization 0.3311338981 0.09642878858496487
+ loop permutation 0.11584685980000001 0.0337355142919035
+ array packing 0.10887823559999998 0.031706196262050394
+ block caching 0.1114187959 0.032446027349902574
+ parallelization 0.1467066874 0.04272213815760692
@@ -1688,7 +1688,7 @@ the computation for specific platforms.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 0.733 seconds)
+ **Total running time of the script:** ( 1 minutes 0.762 seconds)
.. _sphx_glr_download_tutorial_tensor_expr_get_started.py:
diff --git a/docs/commit_hash b/docs/commit_hash
index fc6c0e5f9..8408b8fd7 100644
--- a/docs/commit_hash
+++ b/docs/commit_hash
@@ -1 +1 @@
-4c863fc115ee463284f20b5ee37c973ac0ed5d9a
+b22b872da800b0b44feeca67e808319e21b840a2
diff --git a/docs/how_to/compile_models/from_darknet.html b/docs/how_to/compile_models/from_darknet.html
index 52b62065e..72ad160f3 100644
--- a/docs/how_to/compile_models/from_darknet.html
+++ b/docs/how_to/compile_models/from_darknet.html
@@ -574,7 +574,7 @@ class:['truck 0.9266'] left:471 top:83 right:689 bottom:169
class:['bicycle 0.9984'] left:111 top:113 right:577 bottom:447
</pre></div>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 1.980 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 3.157 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-compile-models-from-darknet-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/7716f96385bd5abb6e822041e285be54/from_darknet.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">from_darknet.py</span></code></a></p>
diff --git a/docs/how_to/compile_models/from_keras.html b/docs/how_to/compile_models/from_keras.html
index e76115b41..85558974f 100644
--- a/docs/how_to/compile_models/from_keras.html
+++ b/docs/how_to/compile_models/from_keras.html
@@ -493,7 +493,7 @@ pip install -U tensorflow --user
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Relay top-1 id: 285, class name: Egyptian cat
1/1 [==============================] - ETA: 0s
-1/1 [==============================] - 1s 897ms/step
+1/1 [==============================] - 1s 916ms/step
Keras top-1 id: 285, class name: Egyptian cat
</pre></div>
</div>
diff --git a/docs/how_to/compile_models/from_mxnet.html b/docs/how_to/compile_models/from_mxnet.html
index 608441ae4..ddfee778b 100644
--- a/docs/how_to/compile_models/from_mxnet.html
+++ b/docs/how_to/compile_models/from_mxnet.html
@@ -427,7 +427,7 @@ to download the full example code</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"x"</span><span class="p">,</span> <a href="https://docs.python.org/3/library/stdtypes.html#tuple" title="builtins.tuple" class="sphx-glr-backref-module-builtins sphx-glr-backref-type-py-class sphx-glr-backref-instance"><span class="n">x</span><span class="o">.</span><span class="n">shape</span></a><span class="p">)</span>
</pre></div>
</div>
-<img src="../../_images/sphx_glr_from_mxnet_001.png" srcset="../../_images/sphx_glr_from_mxnet_001.png" alt="from mxnet" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zip621d8ff1-749e-4f6f-ba07-09560314ab54 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip...
+<img src="../../_images/sphx_glr_from_mxnet_001.png" srcset="../../_images/sphx_glr_from_mxnet_001.png" alt="from mxnet" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zip19c5496d-5bd0-456d-86d0-04121012ec1c from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip...
x (1, 3, 224, 224)
</pre></div>
</div>
diff --git a/docs/how_to/compile_models/from_oneflow.html b/docs/how_to/compile_models/from_oneflow.html
index 38e07af2b..1ce5bcb78 100644
--- a/docs/how_to/compile_models/from_oneflow.html
+++ b/docs/how_to/compile_models/from_oneflow.html
@@ -435,13 +435,13 @@ Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdo
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading: "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/ResNet/resnet18.zip" to /workspace/.oneflow/flowvision_cache/resnet18.zip
0%| | 0.00/41.5M [00:00<?, ?B/s]
- 20%|#9 | 8.12M/41.5M [00:00<00:00, 85.1MB/s]
- 39%|###9 | 16.2M/41.5M [00:00<00:00, 71.5MB/s]
- 56%|#####5 | 23.2M/41.5M [00:00<00:00, 67.7MB/s]
- 72%|#######1 | 29.7M/41.5M [00:00<00:00, 62.8MB/s]
- 86%|########6 | 35.8M/41.5M [00:00<00:00, 55.7MB/s]
- 99%|#########9| 41.2M/41.5M [00:00<00:00, 53.2MB/s]
-100%|##########| 41.5M/41.5M [00:00<00:00, 59.5MB/s]
+ 19%|#9 | 7.99M/41.5M [00:00<00:00, 49.5MB/s]
+ 39%|###8 | 16.0M/41.5M [00:00<00:00, 50.2MB/s]
+ 55%|#####4 | 22.7M/41.5M [00:00<00:00, 57.1MB/s]
+ 69%|######8 | 28.4M/41.5M [00:00<00:00, 54.8MB/s]
+ 81%|########1 | 33.8M/41.5M [00:00<00:00, 50.5MB/s]
+ 93%|#########3| 38.7M/41.5M [00:00<00:00, 42.5MB/s]
+100%|##########| 41.5M/41.5M [00:00<00:00, 47.1MB/s]
</pre></div>
</div>
</div>
diff --git a/docs/how_to/compile_models/from_pytorch.html b/docs/how_to/compile_models/from_pytorch.html
index c7064ea8e..b8e024703 100644
--- a/docs/how_to/compile_models/from_pytorch.html
+++ b/docs/how_to/compile_models/from_pytorch.html
@@ -414,8 +414,9 @@ be unstable.</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /workspace/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
0%| | 0.00/44.7M [00:00<?, ?B/s]
- 34%|###3 | 15.0M/44.7M [00:00<00:00, 157MB/s]
-100%|##########| 44.7M/44.7M [00:00<00:00, 235MB/s]
+ 39%|###8 | 17.2M/44.7M [00:00<00:00, 180MB/s]
+ 94%|#########3| 41.8M/44.7M [00:00<00:00, 226MB/s]
+100%|##########| 44.7M/44.7M [00:00<00:00, 222MB/s]
</pre></div>
</div>
</div>
diff --git a/docs/how_to/compile_models/from_tensorflow.html b/docs/how_to/compile_models/from_tensorflow.html
index 4c67f09a9..923dae192 100644
--- a/docs/how_to/compile_models/from_tensorflow.html
+++ b/docs/how_to/compile_models/from_tensorflow.html
@@ -636,7 +636,7 @@ banana (score = 0.00022)
desk (score = 0.00019)
</pre></div>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 5.350 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 5.653 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-compile-models-from-tensorflow-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/7f1d3d1b878694c201c614c807cdebc8/from_tensorflow.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">from_tensorflow.py</span></code></a></p>
diff --git a/docs/how_to/compile_models/sg_execution_times.html b/docs/how_to/compile_models/sg_execution_times.html
index c9ce0f7df..d0682f971 100644
--- a/docs/how_to/compile_models/sg_execution_times.html
+++ b/docs/how_to/compile_models/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-compile-models-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>05:01.964</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
+<p><strong>05:03.518</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 81%" />
@@ -336,43 +336,43 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="from_tensorflow.html#sphx-glr-how-to-compile-models-from-tensorflow-py"><span class="std std-ref">Compile Tensorflow Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_tensorflow.py</span></code>)</p></td>
-<td><p>01:05.350</p></td>
+<td><p>01:05.653</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="from_darknet.html#sphx-glr-how-to-compile-models-from-darknet-py"><span class="std std-ref">Compile YOLO-V2 and YOLO-V3 in DarkNet Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_darknet.py</span></code>)</p></td>
-<td><p>01:01.980</p></td>
+<td><p>01:03.157</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="from_paddle.html#sphx-glr-how-to-compile-models-from-paddle-py"><span class="std std-ref">Compile PaddlePaddle Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_paddle.py</span></code>)</p></td>
-<td><p>00:39.257</p></td>
+<td><p>00:38.880</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="from_oneflow.html#sphx-glr-how-to-compile-models-from-oneflow-py"><span class="std std-ref">Compile OneFlow Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_oneflow.py</span></code>)</p></td>
-<td><p>00:27.045</p></td>
+<td><p>00:27.838</p></td>
<td><p>0.0 MB</p></td>
</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="from_mxnet.html#sphx-glr-how-to-compile-models-from-mxnet-py"><span class="std std-ref">Compile MXNet Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_mxnet.py</span></code>)</p></td>
-<td><p>00:25.306</p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="from_tflite.html#sphx-glr-how-to-compile-models-from-tflite-py"><span class="std std-ref">Compile TFLite Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_tflite.py</span></code>)</p></td>
+<td><p>00:25.177</p></td>
<td><p>0.0 MB</p></td>
</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="from_tflite.html#sphx-glr-how-to-compile-models-from-tflite-py"><span class="std std-ref">Compile TFLite Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_tflite.py</span></code>)</p></td>
-<td><p>00:24.765</p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="from_mxnet.html#sphx-glr-how-to-compile-models-from-mxnet-py"><span class="std std-ref">Compile MXNet Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_mxnet.py</span></code>)</p></td>
+<td><p>00:24.905</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="from_coreml.html#sphx-glr-how-to-compile-models-from-coreml-py"><span class="std std-ref">Compile CoreML Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_coreml.py</span></code>)</p></td>
-<td><p>00:21.073</p></td>
+<td><p>00:21.325</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="from_pytorch.html#sphx-glr-how-to-compile-models-from-pytorch-py"><span class="std std-ref">Compile PyTorch Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_pytorch.py</span></code>)</p></td>
-<td><p>00:19.155</p></td>
+<td><p>00:19.160</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="from_keras.html#sphx-glr-how-to-compile-models-from-keras-py"><span class="std std-ref">Compile Keras Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_keras.py</span></code>)</p></td>
-<td><p>00:15.638</p></td>
+<td><p>00:14.963</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="from_onnx.html#sphx-glr-how-to-compile-models-from-onnx-py"><span class="std std-ref">Compile ONNX Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_onnx.py</span></code>)</p></td>
-<td><p>00:02.396</p></td>
+<td><p>00:02.460</p></td>
<td><p>0.0 MB</p></td>
</tr>
</tbody>
diff --git a/docs/how_to/deploy_models/deploy_model_on_android.html b/docs/how_to/deploy_models/deploy_model_on_android.html
index fa922a7be..86ffd7f8d 100644
--- a/docs/how_to/deploy_models/deploy_model_on_android.html
+++ b/docs/how_to/deploy_models/deploy_model_on_android.html
@@ -653,7 +653,7 @@ to the remote android device.</p>
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 15.2743 15.2643 15.4321 15.1730 0.0696
+ 15.8805 15.8494 16.1064 15.6990 0.1340
</pre></div>
</div>
</div>
diff --git a/docs/how_to/deploy_models/deploy_object_detection_pytorch.html b/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
index 4eb0a04cd..13b30253c 100644
--- a/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
+++ b/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
@@ -436,39 +436,56 @@ be unstable.</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading: "https://download.pytorch.org/models/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth" to /workspace/.cache/torch/hub/checkpoints/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth
0%| | 0.00/170M [00:00<?, ?B/s]
- 2%|2 | 3.87M/170M [00:00<00:04, 40.4MB/s]
- 5%|4 | 7.73M/170M [00:00<00:04, 38.1MB/s]
- 7%|6 | 11.4M/170M [00:00<00:05, 33.0MB/s]
- 10%|9 | 16.6M/170M [00:00<00:03, 40.6MB/s]
- 13%|#2 | 21.5M/170M [00:00<00:03, 44.1MB/s]
- 15%|#5 | 25.8M/170M [00:00<00:03, 43.3MB/s]
- 18%|#7 | 30.0M/170M [00:00<00:03, 37.7MB/s]
- 21%|##1 | 36.0M/170M [00:00<00:03, 44.5MB/s]
- 24%|##3 | 40.4M/170M [00:01<00:03, 44.4MB/s]
- 26%|##6 | 44.8M/170M [00:01<00:02, 44.1MB/s]
- 30%|##9 | 50.1M/170M [00:01<00:02, 47.5MB/s]
- 32%|###2 | 55.0M/170M [00:01<00:02, 48.6MB/s]
- 35%|###5 | 59.7M/170M [00:01<00:02, 47.1MB/s]
- 38%|###8 | 64.7M/170M [00:01<00:02, 48.4MB/s]
- 41%|####1 | 70.3M/170M [00:01<00:02, 51.5MB/s]
- 44%|####4 | 75.3M/170M [00:01<00:01, 50.8MB/s]
- 48%|####7 | 81.0M/170M [00:01<00:01, 53.4MB/s]
- 51%|##### | 86.3M/170M [00:01<00:01, 54.0MB/s]
- 54%|#####4 | 91.9M/170M [00:02<00:01, 55.4MB/s]
- 57%|#####7 | 97.3M/170M [00:02<00:01, 55.6MB/s]
- 60%|###### | 103M/170M [00:02<00:01, 53.8MB/s]
- 63%|######3 | 108M/170M [00:02<00:01, 50.0MB/s]
- 67%|######7 | 114M/170M [00:02<00:01, 53.7MB/s]
- 70%|####### | 120M/170M [00:02<00:00, 56.0MB/s]
- 74%|#######4 | 126M/170M [00:02<00:00, 58.1MB/s]
- 77%|#######7 | 131M/170M [00:02<00:00, 57.7MB/s]
- 81%|######## | 137M/170M [00:02<00:00, 58.8MB/s]
- 84%|########4 | 143M/170M [00:02<00:00, 57.3MB/s]
- 87%|########7 | 148M/170M [00:03<00:00, 56.8MB/s]
- 91%|######### | 154M/170M [00:03<00:00, 52.7MB/s]
- 94%|#########3| 159M/170M [00:03<00:00, 43.4MB/s]
- 97%|#########6| 164M/170M [00:03<00:00, 46.5MB/s]
-100%|##########| 170M/170M [00:03<00:00, 49.5MB/s]
+ 2%|2 | 3.56M/170M [00:00<00:04, 36.4MB/s]
+ 4%|4 | 7.03M/170M [00:00<00:05, 31.4MB/s]
+ 6%|6 | 10.4M/170M [00:00<00:05, 32.6MB/s]
+ 8%|8 | 13.6M/170M [00:00<00:05, 32.4MB/s]
+ 10%|9 | 16.9M/170M [00:00<00:04, 32.9MB/s]
+ 12%|#2 | 20.9M/170M [00:00<00:04, 35.9MB/s]
+ 14%|#4 | 24.4M/170M [00:00<00:04, 35.6MB/s]
+ 17%|#6 | 28.2M/170M [00:00<00:04, 37.1MB/s]
+ 19%|#9 | 32.5M/170M [00:00<00:03, 39.5MB/s]
+ 21%|##1 | 36.3M/170M [00:01<00:03, 39.6MB/s]
+ 24%|##3 | 40.1M/170M [00:01<00:03, 38.7MB/s]
+ 26%|##5 | 43.8M/170M [00:01<00:03, 37.0MB/s]
+ 28%|##7 | 47.4M/170M [00:01<00:03, 35.4MB/s]
+ 30%|##9 | 50.8M/170M [00:01<00:04, 28.0MB/s]
+ 32%|###1 | 53.7M/170M [00:01<00:05, 24.3MB/s]
+ 33%|###3 | 56.6M/170M [00:01<00:04, 25.8MB/s]
+ 35%|###5 | 60.0M/170M [00:01<00:04, 27.9MB/s]
+ 37%|###6 | 62.8M/170M [00:02<00:04, 24.9MB/s]
+ 39%|###8 | 66.0M/170M [00:02<00:04, 26.5MB/s]
+ 41%|####1 | 70.2M/170M [00:02<00:03, 30.8MB/s]
+ 43%|####3 | 73.3M/170M [00:02<00:03, 31.2MB/s]
+ 45%|####4 | 76.4M/170M [00:02<00:03, 28.9MB/s]
+ 47%|####6 | 79.3M/170M [00:02<00:03, 27.8MB/s]
+ 48%|####8 | 82.0M/170M [00:02<00:03, 27.7MB/s]
+ 50%|####9 | 84.9M/170M [00:02<00:03, 27.2MB/s]
+ 52%|#####1 | 87.5M/170M [00:03<00:05, 17.0MB/s]
+ 53%|#####2 | 89.6M/170M [00:03<00:04, 17.2MB/s]
+ 54%|#####3 | 91.6M/170M [00:03<00:04, 17.0MB/s]
+ 55%|#####5 | 93.5M/170M [00:03<00:04, 17.4MB/s]
+ 56%|#####6 | 95.3M/170M [00:03<00:04, 17.8MB/s]
+ 58%|#####8 | 98.5M/170M [00:03<00:03, 21.7MB/s]
+ 60%|#####9 | 102M/170M [00:03<00:02, 24.5MB/s]
+ 61%|######1 | 104M/170M [00:04<00:02, 23.0MB/s]
+ 63%|######3 | 108M/170M [00:04<00:02, 27.2MB/s]
+ 65%|######5 | 111M/170M [00:04<00:02, 28.3MB/s]
+ 67%|######7 | 114M/170M [00:04<00:02, 28.1MB/s]
+ 69%|######9 | 117M/170M [00:04<00:01, 29.7MB/s]
+ 72%|#######1 | 122M/170M [00:04<00:01, 34.5MB/s]
+ 74%|#######3 | 125M/170M [00:04<00:01, 35.8MB/s]
+ 76%|#######6 | 130M/170M [00:04<00:01, 38.0MB/s]
+ 79%|#######9 | 134M/170M [00:04<00:00, 42.0MB/s]
+ 82%|########1 | 139M/170M [00:04<00:00, 44.3MB/s]
+ 85%|########4 | 144M/170M [00:05<00:00, 42.7MB/s]
+ 88%|########7 | 149M/170M [00:05<00:00, 46.3MB/s]
+ 90%|######### | 153M/170M [00:05<00:00, 45.2MB/s]
+ 93%|#########2| 158M/170M [00:05<00:00, 34.6MB/s]
+ 95%|#########4| 161M/170M [00:05<00:00, 35.4MB/s]
+ 97%|#########7| 165M/170M [00:05<00:00, 33.1MB/s]
+ 99%|#########9| 168M/170M [00:05<00:00, 31.8MB/s]
+100%|##########| 170M/170M [00:05<00:00, 30.4MB/s]
/usr/local/lib/python3.7/dist-packages/torch/nn/functional.py:3878: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
for i in range(dim)
/usr/local/lib/python3.7/dist-packages/torchvision/models/detection/anchor_utils.py:127: UserWarning: __floordiv__ is deprecated, and its behavior will change in a future version of pytorch. It currently rounds toward 0 (like the 'trunc' function NOT 'floor'). This results in incorrect rounding for negative values. To keep the current behavior, use torch.div(a, b, rounding_mode='trunc'), or for actual floor division, use torch.div(a, b, rounding_mode='floor').
@@ -566,7 +583,7 @@ torchvision rcnn models.</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Get 9 valid boxes
</pre></div>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 2 minutes 53.079 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 2 minutes 59.602 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-deploy-models-deploy-object-detection-pytorch-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/7795da4b258c8feff986668b95ef57ad/deploy_object_detection_pytorch.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">deploy_object_detection_pytorch.py</span></code></a></p>
diff --git a/docs/how_to/deploy_models/deploy_prequantized.html b/docs/how_to/deploy_models/deploy_prequantized.html
index dcd5d65ec..c2e2f105d 100644
--- a/docs/how_to/deploy_models/deploy_prequantized.html
+++ b/docs/how_to/deploy_models/deploy_prequantized.html
@@ -480,9 +480,9 @@ training. Other models require a full post training calibration.</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading: "https://download.pytorch.org/models/mobilenet_v2-b0353104.pth" to /workspace/.cache/torch/hub/checkpoints/mobilenet_v2-b0353104.pth
0%| | 0.00/13.6M [00:00<?, ?B/s]
- 34%|###4 | 4.61M/13.6M [00:00<00:00, 48.2MB/s]
- 68%|######7 | 9.21M/13.6M [00:00<00:00, 45.5MB/s]
-100%|##########| 13.6M/13.6M [00:00<00:00, 60.0MB/s]
+ 18%|#8 | 2.46M/13.6M [00:00<00:00, 25.6MB/s]
+ 42%|####2 | 5.70M/13.6M [00:00<00:00, 30.5MB/s]
+100%|##########| 13.6M/13.6M [00:00<00:00, 58.4MB/s]
</pre></div>
</div>
</div>
@@ -571,7 +571,7 @@ output values are identical out of 1000 outputs from mobilenet v2.</p>
</div>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 88.6839 88.6130 89.7298 88.4377 0.2284
+ 90.1142 90.0253 95.4225 89.6971 0.5840
</pre></div>
</div>
<div class="admonition note">
@@ -610,7 +610,7 @@ This includes support for the VNNI 8 bit dot product instruction (CascadeLake or
<div class="section" id="deploy-a-quantized-tflite-model">
<h2>Deploy a quantized TFLite Model<a class="headerlink" href="#deploy-a-quantized-tflite-model" title="Permalink to this headline">¶</a></h2>
<p>TODO</p>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 7.246 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 8.445 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-deploy-models-deploy-prequantized-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/fb8217c13f4351224c6cf3aacf1a87fc/deploy_prequantized.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">deploy_prequantized.py</span></code></a></p>
diff --git a/docs/how_to/deploy_models/deploy_prequantized_tflite.html b/docs/how_to/deploy_models/deploy_prequantized_tflite.html
index 6bb66ad9a..64cf25d02 100644
--- a/docs/how_to/deploy_models/deploy_prequantized_tflite.html
+++ b/docs/how_to/deploy_models/deploy_prequantized_tflite.html
@@ -573,7 +573,7 @@ TFLite Top-5 labels: [387 102 386 341 349]
</div>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 115.3626 114.9331 121.7323 113.8598 1.2658
+ 120.0408 120.0760 124.6104 117.7518 0.8415
</pre></div>
</div>
<div class="admonition note">
@@ -601,7 +601,7 @@ network for ARM CPU</span></a>.</p></li>
</ul>
</div></blockquote>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 49.373 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 55.942 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-deploy-models-deploy-prequantized-tflite-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/56691c7a27d45da61d112276334640d3/deploy_prequantized_tflite.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">deploy_prequantized_tflite.py</span></code></a></p>
diff --git a/docs/how_to/deploy_models/deploy_quantized.html b/docs/how_to/deploy_models/deploy_quantized.html
index ecb190e85..0727b87b3 100644
--- a/docs/how_to/deploy_models/deploy_quantized.html
+++ b/docs/how_to/deploy_models/deploy_quantized.html
@@ -509,7 +509,7 @@ for calibration. But the accuracy might be impacted.</p>
DeprecationWarning,
</pre></div>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 30.173 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 30.164 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-deploy-models-deploy-quantized-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/7810ecf51bfc05f7d5e8a400ac3e815d/deploy_quantized.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">deploy_quantized.py</span></code></a></p>
diff --git a/docs/how_to/deploy_models/deploy_ssd_gluoncv.html b/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
index 5e66c5bc2..70076d8a3 100644
--- a/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
+++ b/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
@@ -441,23 +441,24 @@ to your device.</p>
Downloading /workspace/.mxnet/models/ssd_512_resnet50_v1_voc-9c8b225a.zip from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/ssd_512_resnet50_v1_voc-9c8b225a.zip...
0%| | 0/132723 [00:00<?, ?KB/s]
- 5%|4 | 6034/132723 [00:00<00:02, 60335.17KB/s]
- 11%|# | 14118/132723 [00:00<00:01, 72389.67KB/s]
- 16%|#6 | 21357/132723 [00:00<00:01, 62265.24KB/s]
- 22%|##2 | 29437/132723 [00:00<00:01, 68918.80KB/s]
- 28%|##8 | 37516/132723 [00:00<00:01, 72944.53KB/s]
- 34%|###4 | 45650/132723 [00:00<00:01, 75677.02KB/s]
- 41%|#### | 53837/132723 [00:00<00:01, 77645.25KB/s]
- 47%|####6 | 62018/132723 [00:00<00:00, 78945.51KB/s]
- 53%|#####2 | 70209/132723 [00:00<00:00, 79856.88KB/s]
- 59%|#####9 | 78366/132723 [00:01<00:00, 80379.82KB/s]
- 65%|######5 | 86526/132723 [00:01<00:00, 80746.30KB/s]
- 71%|#######1 | 94639/132723 [00:01<00:00, 80861.19KB/s]
- 77%|#######7 | 102806/132723 [00:01<00:00, 81103.66KB/s]
- 84%|########3 | 111024/132723 [00:01<00:00, 81424.16KB/s]
- 90%|########9 | 119173/132723 [00:01<00:00, 81358.71KB/s]
- 96%|#########5| 127358/132723 [00:01<00:00, 81504.18KB/s]
-100%|##########| 132723/132723 [00:01<00:00, 77995.44KB/s]
+ 4%|4 | 5598/132723 [00:00<00:02, 55961.24KB/s]
+ 10%|9 | 13009/132723 [00:00<00:01, 66632.37KB/s]
+ 15%|#5 | 20526/132723 [00:00<00:01, 70527.47KB/s]
+ 21%|##1 | 28024/132723 [00:00<00:01, 72283.52KB/s]
+ 27%|##6 | 35378/132723 [00:00<00:01, 72734.74KB/s]
+ 32%|###2 | 42898/132723 [00:00<00:01, 73569.98KB/s]
+ 38%|###8 | 50465/132723 [00:00<00:01, 74245.79KB/s]
+ 44%|####3 | 57943/132723 [00:00<00:01, 74408.86KB/s]
+ 49%|####9 | 65642/132723 [00:00<00:00, 75213.35KB/s]
+ 55%|#####5 | 73378/132723 [00:01<00:00, 75873.43KB/s]
+ 61%|######1 | 81241/132723 [00:01<00:00, 76703.52KB/s]
+ 67%|######7 | 89158/132723 [00:01<00:00, 77450.97KB/s]
+ 73%|#######3 | 97020/132723 [00:01<00:00, 77803.33KB/s]
+ 79%|#######9 | 104852/132723 [00:01<00:00, 77956.81KB/s]
+ 85%|########4 | 112684/132723 [00:01<00:00, 78064.84KB/s]
+ 91%|######### | 120552/132723 [00:01<00:00, 78246.25KB/s]
+ 97%|#########6| 128402/132723 [00:01<00:00, 78319.20KB/s]
+100%|##########| 132723/132723 [00:01<00:00, 75591.62KB/s]
</pre></div>
</div>
<p>Create TVM runtime and do inference
@@ -500,7 +501,7 @@ Downloading /workspace/.mxnet/models/ssd_512_resnet50_v1_voc-9c8b225a.zip from h
<span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">()</span>
</pre></div>
</div>
-<img src="../../_images/sphx_glr_deploy_ssd_gluoncv_001.png" srcset="../../_images/sphx_glr_deploy_ssd_gluoncv_001.png" alt="deploy ssd gluoncv" class = "sphx-glr-single-img"/><p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 2 minutes 31.414 seconds)</p>
+<img src="../../_images/sphx_glr_deploy_ssd_gluoncv_001.png" srcset="../../_images/sphx_glr_deploy_ssd_gluoncv_001.png" alt="deploy ssd gluoncv" class = "sphx-glr-single-img"/><p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 2 minutes 34.457 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-deploy-models-deploy-ssd-gluoncv-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/cccb17d28e5e8b2e94ea8cd5ec59f6ed/deploy_ssd_gluoncv.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">deploy_ssd_gluoncv.py</span></code></a></p>
diff --git a/docs/how_to/deploy_models/sg_execution_times.html b/docs/how_to/deploy_models/sg_execution_times.html
index e9d4bb628..e2e6dab49 100644
--- a/docs/how_to/deploy_models/sg_execution_times.html
+++ b/docs/how_to/deploy_models/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-deploy-models-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>11:02.831</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
+<p><strong>11:22.889</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 86%" />
@@ -336,35 +336,35 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="deploy_object_detection_pytorch.html#sphx-glr-how-to-deploy-models-deploy-object-detection-pytorch-py"><span class="std std-ref">Compile PyTorch Object Detection Models</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_object_detection_pytorch.py</span></code>)</p></td>
-<td><p>02:53.079</p></td>
+<td><p>02:59.602</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="deploy_ssd_gluoncv.html#sphx-glr-how-to-deploy-models-deploy-ssd-gluoncv-py"><span class="std std-ref">Deploy Single Shot Multibox Detector(SSD) model</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_ssd_gluoncv.py</span></code>)</p></td>
-<td><p>02:31.414</p></td>
+<td><p>02:34.457</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="deploy_prequantized_tflite.html#sphx-glr-how-to-deploy-models-deploy-prequantized-tflite-py"><span class="std std-ref">Deploy a Framework-prequantized Model with TVM - Part 3 (TFLite)</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_prequantized_tflite.py</span></code>)</p></td>
-<td><p>01:49.373</p></td>
+<td><p>01:55.942</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="deploy_quantized.html#sphx-glr-how-to-deploy-models-deploy-quantized-py"><span class="std std-ref">Deploy a Quantized Model on Cuda</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_quantized.py</span></code>)</p></td>
-<td><p>01:30.173</p></td>
+<td><p>01:30.164</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="deploy_prequantized.html#sphx-glr-how-to-deploy-models-deploy-prequantized-py"><span class="std std-ref">Deploy a Framework-prequantized Model with TVM</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_prequantized.py</span></code>)</p></td>
-<td><p>01:07.246</p></td>
+<td><p>01:08.445</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="deploy_model_on_android.html#sphx-glr-how-to-deploy-models-deploy-model-on-android-py"><span class="std std-ref">Deploy the Pretrained Model on Android</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_android.py</span></code>)</p></td>
-<td><p>00:28.854</p></td>
+<td><p>00:30.138</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="deploy_model_on_nano.html#sphx-glr-how-to-deploy-models-deploy-model-on-nano-py"><span class="std std-ref">Deploy the Pretrained Model on Jetson Nano</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_nano.py</span></code>)</p></td>
-<td><p>00:21.467</p></td>
+<td><p>00:22.342</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="deploy_model_on_rasp.html#sphx-glr-how-to-deploy-models-deploy-model-on-rasp-py"><span class="std std-ref">Deploy the Pretrained Model on Raspberry Pi</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_rasp.py</span></code>)</p></td>
-<td><p>00:21.220</p></td>
+<td><p>00:21.792</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="deploy_sparse.html#sphx-glr-how-to-deploy-models-deploy-sparse-py"><span class="std std-ref">Deploy a Hugging Face Pruned Model on CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_sparse.py</span></code>)</p></td>
diff --git a/docs/how_to/extend_tvm/bring_your_own_datatypes.html b/docs/how_to/extend_tvm/bring_your_own_datatypes.html
index c48d82861..82336efa9 100644
--- a/docs/how_to/extend_tvm/bring_your_own_datatypes.html
+++ b/docs/how_to/extend_tvm/bring_your_own_datatypes.html
@@ -612,7 +612,7 @@ In this alpha state of the Bring Your Own Datatypes framework, we have not imple
<span class="n">module</span><span class="p">,</span> <a href="https://docs.python.org/3/library/stdtypes.html#dict" title="builtins.dict" class="sphx-glr-backref-module-builtins sphx-glr-backref-type-py-class sphx-glr-backref-instance"><span class="n">params</span></a> <span class="o">=</span> <span class="n">get_mobilenet</span><span class="p">()</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zip1efdd2e7-3622-4514-a4ad-5a86bdf60c5b from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zip93f8e8a8-e264-497e-8f53-3afeffa873ab from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
</pre></div>
</div>
<p>It’s easy to execute MobileNet with native TVM:</p>
diff --git a/docs/how_to/extend_tvm/sg_execution_times.html b/docs/how_to/extend_tvm/sg_execution_times.html
index 9060f255c..4737f8aae 100644
--- a/docs/how_to/extend_tvm/sg_execution_times.html
+++ b/docs/how_to/extend_tvm/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-extend-tvm-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:40.045</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
+<p><strong>00:40.364</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -336,19 +336,19 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="bring_your_own_datatypes.html#sphx-glr-how-to-extend-tvm-bring-your-own-datatypes-py"><span class="std std-ref">Bring Your Own Datatypes to TVM</span></a> (<code class="docutils literal notranslate"><span class="pre">bring_your_own_datatypes.py</span></code>)</p></td>
-<td><p>00:37.056</p></td>
+<td><p>00:37.285</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="use_pass_instrument.html#sphx-glr-how-to-extend-tvm-use-pass-instrument-py"><span class="std std-ref">How to Use TVM Pass Instrument</span></a> (<code class="docutils literal notranslate"><span class="pre">use_pass_instrument.py</span></code>)</p></td>
-<td><p>00:02.104</p></td>
+<td><p>00:02.140</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="use_pass_infra.html#sphx-glr-how-to-extend-tvm-use-pass-infra-py"><span class="std std-ref">How to Use TVM Pass Infra</span></a> (<code class="docutils literal notranslate"><span class="pre">use_pass_infra.py</span></code>)</p></td>
-<td><p>00:00.877</p></td>
+<td><p>00:00.932</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="low_level_custom_pass.html#sphx-glr-how-to-extend-tvm-low-level-custom-pass-py"><span class="std std-ref">Writing a Customized Pass</span></a> (<code class="docutils literal notranslate"><span class="pre">low_level_custom_pass.py</span></code>)</p></td>
-<td><p>00:00.008</p></td>
+<td><p>00:00.007</p></td>
<td><p>0.0 MB</p></td>
</tr>
</tbody>
diff --git a/docs/how_to/extend_tvm/use_pass_instrument.html b/docs/how_to/extend_tvm/use_pass_instrument.html
index d8156ec0e..a96c8d964 100644
--- a/docs/how_to/extend_tvm/use_pass_instrument.html
+++ b/docs/how_to/extend_tvm/use_pass_instrument.html
@@ -512,10 +512,10 @@ profile the execution time of each passes.</p>
</pre></div>
</div>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Printing results of timing profile...
-InferType: 6544us [6544us] (45.82%; 45.82%)
-FoldScaleAxis: 7737us [5us] (54.18%; 54.18%)
- FoldConstant: 7732us [1600us] (54.14%; 99.94%)
- InferType: 6132us [6132us] (42.94%; 79.31%)
+InferType: 6623us [6623us] (45.78%; 45.78%)
+FoldScaleAxis: 7842us [5us] (54.22%; 54.22%)
+ FoldConstant: 7837us [1623us] (54.18%; 99.93%)
+ InferType: 6213us [6213us] (42.96%; 79.29%)
</pre></div>
</div>
</div>
@@ -537,10 +537,10 @@ Refer to following sections and <a class="reference internal" href="../../refere
</pre></div>
</div>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Printing results of timing profile...
-InferType: 6253us [6253us] (44.97%; 44.97%)
-FoldScaleAxis: 7652us [4us] (55.03%; 55.03%)
- FoldConstant: 7647us [1572us] (55.00%; 99.94%)
- InferType: 6076us [6076us] (43.70%; 79.45%)
+InferType: 6269us [6269us] (44.39%; 44.39%)
+FoldScaleAxis: 7855us [4us] (55.61%; 55.61%)
+ FoldConstant: 7850us [1648us] (55.58%; 99.94%)
+ InferType: 6202us [6202us] (43.91%; 79.00%)
</pre></div>
</div>
<p>Register empty list to clear existing instruments.</p>
diff --git a/docs/how_to/optimize_operators/opt_conv_cuda.html b/docs/how_to/optimize_operators/opt_conv_cuda.html
index 06d757c7f..b4d5bd0d2 100644
--- a/docs/how_to/optimize_operators/opt_conv_cuda.html
+++ b/docs/how_to/optimize_operators/opt_conv_cuda.html
@@ -564,7 +564,7 @@ latency of convolution.</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Convolution: </span><span class="si">%f</span><span class="s2"> ms"</span> <span class="o">%</span> <span class="p">(</span><span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">w</span><span class="p">,</span> <span class="n">b</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span> <span class="o">*</span> <span cl [...]
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Convolution: 44.051947 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Convolution: 54.123335 ms
</pre></div>
</div>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-optimize-operators-opt-conv-cuda-py">
diff --git a/docs/how_to/optimize_operators/opt_conv_tensorcore.html b/docs/how_to/optimize_operators/opt_conv_tensorcore.html
index fbe1c7a10..d3021bbbb 100644
--- a/docs/how_to/optimize_operators/opt_conv_tensorcore.html
+++ b/docs/how_to/optimize_operators/opt_conv_tensorcore.html
@@ -906,7 +906,7 @@ be able to run on our build server</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"conv2d with tensor core: </span><span class="si">%f</span><span class="s2"> ms"</span> <span class="o">%</span> <span class="p">(</span><span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">w</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span> <span class="o">* [...]
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>conv2d with tensor core: 10.977927 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>conv2d with tensor core: 7.243840 ms
</pre></div>
</div>
</div>
diff --git a/docs/how_to/optimize_operators/opt_gemm.html b/docs/how_to/optimize_operators/opt_gemm.html
index 0d8056eb3..e22761ff3 100644
--- a/docs/how_to/optimize_operators/opt_gemm.html
+++ b/docs/how_to/optimize_operators/opt_gemm.html
@@ -461,8 +461,8 @@ Then we write a baseline implementation, the simplest way to write a matrix mult
<span class="nb">print</span><span class="p">(</span><span class="s2">"Baseline: </span><span class="si">%f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Numpy running time: 0.017506
-Baseline: 3.333864
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Numpy running time: 0.017938
+Baseline: 3.440580
</pre></div>
</div>
<p>In TVM, we can always inspect lower level IR to debug or optimize our schedule.
@@ -522,7 +522,7 @@ fill 32 * 32 * sizeof(float) which is 4KB in the cache whose total size is 32KB
<span class="nb">print</span><span class="p">(</span><span class="s2">"Opt1: </span><span class="si">%f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt1: 0.294581
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt1: 0.293736
</pre></div>
</div>
<p>Here is the generated IR after blocking.</p>
@@ -589,7 +589,7 @@ vastly.</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Opt2: </span><span class="si">%f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt2: 0.328113
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt2: 0.328085
</pre></div>
</div>
<p>Here is the generated IR after vectorization.</p>
@@ -650,7 +650,7 @@ the access pattern for A matrix is more cache friendly.</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Opt3: </span><span class="si">%f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt3: 0.111140
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt3: 0.113581
</pre></div>
</div>
<p>Here is the generated IR after loop permutation.</p>
@@ -733,7 +733,7 @@ flattening.</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Opt4: </span><span class="si">%f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt4: 0.107377
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt4: 0.109142
</pre></div>
</div>
<p>Here is the generated IR after array packing.</p>
@@ -819,7 +819,7 @@ write to C when all the block results are ready.</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Opt5: </span><span class="si">%f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt5: 0.108128
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt5: 0.111236
</pre></div>
</div>
<p>Here is the generated IR after blocking.</p>
@@ -909,7 +909,7 @@ write to C when all the block results are ready.</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Opt6: </span><span class="si">%f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">opt6_time</span><span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt6: 0.144638
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt6: 0.146516
</pre></div>
</div>
<p>Here is the generated IR after parallelization.</p>
diff --git a/docs/how_to/optimize_operators/sg_execution_times.html b/docs/how_to/optimize_operators/sg_execution_times.html
index 5d1846cc7..f8e97abc6 100644
--- a/docs/how_to/optimize_operators/sg_execution_times.html
+++ b/docs/how_to/optimize_operators/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-optimize-operators-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:33.822</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
+<p><strong>00:34.483</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -336,15 +336,15 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="opt_gemm.html#sphx-glr-how-to-optimize-operators-opt-gemm-py"><span class="std std-ref">How to optimize GEMM on CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">opt_gemm.py</span></code>)</p></td>
-<td><p>00:31.449</p></td>
+<td><p>00:32.049</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="opt_conv_tensorcore.html#sphx-glr-how-to-optimize-operators-opt-conv-tensorcore-py"><span class="std std-ref">How to optimize convolution using TensorCores</span></a> (<code class="docutils literal notranslate"><span class="pre">opt_conv_tensorcore.py</span></code>)</p></td>
-<td><p>00:01.288</p></td>
+<td><p>00:01.330</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="opt_conv_cuda.html#sphx-glr-how-to-optimize-operators-opt-conv-cuda-py"><span class="std std-ref">How to optimize convolution on GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">opt_conv_cuda.py</span></code>)</p></td>
-<td><p>00:01.086</p></td>
+<td><p>00:01.104</p></td>
<td><p>0.0 MB</p></td>
</tr>
</tbody>
diff --git a/docs/how_to/tune_with_autoscheduler/sg_execution_times.html b/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
index 87a16e1c0..981ad11f5 100644
--- a/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
+++ b/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-tune-with-autoscheduler-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>05:58.933</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
+<p><strong>06:10.469</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 85%" />
@@ -336,27 +336,27 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="tune_conv2d_layer_cuda.html#sphx-glr-how-to-tune-with-autoscheduler-tune-conv2d-layer-cuda-py"><span class="std std-ref">Auto-scheduling a Convolution Layer for GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_conv2d_layer_cuda.py</span></code>)</p></td>
-<td><p>03:17.057</p></td>
+<td><p>03:25.235</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="tune_network_x86.html#sphx-glr-how-to-tune-with-autoscheduler-tune-network-x86-py"><span class="std std-ref">Auto-scheduling a Neural Network for x86 CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_network_x86.py</span></code>)</p></td>
-<td><p>01:20.723</p></td>
+<td><p>01:22.299</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="tune_network_cuda.html#sphx-glr-how-to-tune-with-autoscheduler-tune-network-cuda-py"><span class="std std-ref">Auto-scheduling a Neural Network for NVIDIA GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_network_cuda.py</span></code>)</p></td>
-<td><p>00:45.656</p></td>
+<td><p>00:46.433</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="tune_sparse_x86.html#sphx-glr-how-to-tune-with-autoscheduler-tune-sparse-x86-py"><span class="std std-ref">Auto-scheduling Sparse Matrix Multiplication on CPU with Custom Sketch Rule</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_sparse_x86.py</span></code>)</p></td>
-<td><p>00:18.498</p></td>
+<td><p>00:19.276</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="tune_network_mali.html#sphx-glr-how-to-tune-with-autoscheduler-tune-network-mali-py"><span class="std std-ref">Auto-scheduling a Neural Network for mali GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_network_mali.py</span></code>)</p></td>
-<td><p>00:08.611</p></td>
+<td><p>00:08.687</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="tune_network_arm.html#sphx-glr-how-to-tune-with-autoscheduler-tune-network-arm-py"><span class="std std-ref">Auto-scheduling a Neural Network for ARM CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_network_arm.py</span></code>)</p></td>
-<td><p>00:08.388</p></td>
+<td><p>00:08.539</p></td>
<td><p>0.0 MB</p></td>
</tr>
</tbody>
diff --git a/docs/how_to/tune_with_autoscheduler/tune_conv2d_layer_cuda.html b/docs/how_to/tune_with_autoscheduler/tune_conv2d_layer_cuda.html
index 5f306977d..1ec289baf 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_conv2d_layer_cuda.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_conv2d_layer_cuda.html
@@ -493,53 +493,888 @@ cooperative fetching, unrolling and operator fusion.</p>
preflattened_buffer_map = {data_1: data_3: Buffer(data_2, float32, [1, 512, 7, 7], []), kernel_1: kernel_3: Buffer(kernel_2, float32, [512, 512, 3, 3], []), bias_1: bias_3: Buffer(bias_2, float32, [1, 512, 1, 1], []), compute_1: compute_3: Buffer(compute_2, float32, [1, 512, 7, 7], [])} {
attr [IterVar(blockIdx.x: int32, (nullptr), "ThreadIndex", "blockIdx.x")] "thread_extent" = 16;
allocate(conv2d_nchw: Pointer(local float32), float32, [8]), storage_scope = local;
- allocate(pad_temp.shared: Pointer(shared float32), float32, [1008]), storage_scope = shared;
- allocate(kernel.shared: Pointer(shared float32), float32, [1536]), storage_scope = shared;
+ allocate(pad_temp.shared: Pointer(shared float32), float32, [1568]), storage_scope = shared;
+ allocate(kernel.shared: Pointer(shared float32), float32, [1024]), storage_scope = shared;
attr [IterVar(threadIdx.x: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196 {
- for (ff.outer.inner.init: int32, 0, 2) {
- for (ff.inner.init: int32, 0, 2) {
- let cse_var_1: int32 = ((ff.outer.inner.init*2) + ff.inner.init)
- {
- conv2d_nchw_1: Buffer(conv2d_nchw, float32, [16], [], scope="local", align=16)[cse_var_1] = 0f32
- conv2d_nchw_1[(cse_var_1 + 4)] = 0f32
- }
- }
- }
- for (rc.outer.outer: int32, 0, 32) {
+ conv2d_nchw_1: Buffer(conv2d_nchw, float32, [4], [], scope="local", align=8)[0] = 0f32
+ conv2d_nchw_1[2] = 0f32
+ conv2d_nchw_1[4] = 0f32
+ conv2d_nchw_1[6] = 0f32
+ conv2d_nchw_1[1] = 0f32
+ conv2d_nchw_1[3] = 0f32
+ conv2d_nchw_1[5] = 0f32
+ conv2d_nchw_1[7] = 0f32
+ for (rc.outer.outer: int32, 0, 16) {
for (ry.outer.outer: int32, 0, 3) {
- for (ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer: int32, 0, 6) {
+ let cse_var_2: int32 = (rc.outer.outer*288)
+ let cse_var_1: int32 = (ry.outer.outer*3)
+ {
attr [IterVar(threadIdx.x_1: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
- if @tir.likely((((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer*7) + floordiv(threadIdx.x_1, 28)) < 36), dtype=bool) {
- pad_temp.shared_1: Buffer(pad_temp.shared, float32, [1008], [], scope="shared")[((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer*196) + threadIdx.x_1)] = @tir.if_then_else(((((1 <= (floordiv(floormod(((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer*7) + threadIdx.x_1), 63), 9) + ry.outer.outer)) && ((floordiv(floormod(((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer*7) + threadIdx.x_1), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod(((ax0.ax [...]
- }
- }
- for (ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1: int32, 0, 8) {
+ pad_temp.shared_1: Buffer(pad_temp.shared, float32, [1568], [], scope="shared")[threadIdx.x_1] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 196)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 188)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 392)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 384)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 588)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 580)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 784)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 776)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 980)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 972)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1176)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1168)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1372)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 7))), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1364)], 0f32, dtype=float32)
attr [IterVar(threadIdx.x_2: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
- if @tir.likely((((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1*49) + floordiv(threadIdx.x_2, 4)) < 384), dtype=bool) {
- kernel.shared_1: Buffer(kernel.shared, float32, [1536], [], scope="shared")[((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1*196) + threadIdx.x_2)] = kernel[((((((blockIdx.x*147456) + (floordiv(((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1*49) + floordiv(threadIdx.x_2, 4)), 12)*4608)) + (rc.outer.outer*144)) + (floordiv(floormod(((ax0.ax1.fused.ax2.fused.ax3.fused.outer.outer_1*4) + threadIdx.x_2), 48), 3)*9)) + (ry.outer.outer*3)) + floormod((threadIdx.x_2 + a [...]
+ kernel.shared_1: Buffer(kernel.shared, float32, [1024], [], scope="shared")[threadIdx.x_2] = kernel[(((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 32)*4608)) + cse_var_2) + (floormod(threadIdx.x_2, 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 196)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 196), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 4), 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 392)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 392), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 8), 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 588)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 588), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 12), 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 784)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 784), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 16), 32)*9)) + cse_var_1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ if @tir.likely((threadIdx.x_2 < 44), dtype=bool) {
+ kernel.shared_1[(threadIdx.x_2 + 980)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 980), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 20), 32)*9)) + cse_var_1)]
}
- }
- for (rc.outer.inner: int32, 0, 2) {
- for (ff.outer.inner: int32, 0, 2) {
- for (rc.inner: int32, 0, 8) {
- for (rx.inner: int32, 0, 3) {
- for (ff.inner: int32, 0, 2) {
- let cse_var_3: int32 = ((ff.outer.inner*2) + ff.inner)
- let cse_var_2: int32 = (cse_var_3 + 4)
- {
- conv2d_nchw_1[cse_var_3] = (conv2d_nchw_1[cse_var_3] + (pad_temp.shared_1[(((((rc.outer.inner*504) + (rc.inner*63)) + (floordiv(floormod(threadIdx.x, 49), 7)*9)) + rx.inner) + floormod(threadIdx.x, 7))]*kernel.shared_1[((((((floordiv(threadIdx.x, 49)*192) + (ff.outer.inner*96)) + (ff.inner*48)) + (rc.outer.inner*24)) + (rc.inner*3)) + rx.inner)]))
- conv2d_nchw_1[cse_var_2] = (conv2d_nchw_1[cse_var_2] + (pad_temp.shared_1[(((((rc.outer.inner*504) + (rc.inner*63)) + (floordiv(floormod(threadIdx.x, 49), 7)*9)) + rx.inner) + floormod(threadIdx.x, 7))]*kernel.shared_1[(((((((floordiv(threadIdx.x, 49)*192) + (ff.outer.inner*96)) + (ff.inner*48)) + (rc.outer.inner*24)) + (rc.inner*3)) + rx.inner) + 768)]))
- }
- }
- }
- }
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[(floordiv(threadIdx.x, 49)*64)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 256)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 512)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 768)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 32)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 288)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 544)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 800)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 1)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 257)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 513)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 769)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 33)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 289)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 545)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 801)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 2)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 258)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 514)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 770)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 34)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 290)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 546)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 802)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 3)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 259)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 515)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 771)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 35)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 291)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 547)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 803)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 4)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 260)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 516)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 772)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 36)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 292)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 548)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 804)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 5)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 261)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 517)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 773)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 37)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 293)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 549)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 805)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 6)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 262)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 518)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 774)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 38)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 294)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 550)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 806)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 7)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 263)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 519)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 775)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 39)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 295)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 551)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 807)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 8)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 264)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 520)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 776)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 40)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 296)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 552)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 808)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 9)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 265)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 521)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 777)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 41)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 297)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 553)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 809)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 10)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 266)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 522)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 778)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 42)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 298)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 554)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 810)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 11)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 267)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 523)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 779)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 43)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 299)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 555)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 811)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 12)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 268)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 524)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 780)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 44)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 300)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 556)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 812)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 13)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 269)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 525)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 781)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 45)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 301)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 557)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 813)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 14)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 270)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 526)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 782)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 46)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 302)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 558)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 814)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 15)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 271)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 527)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 783)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 47)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 303)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 559)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 815)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 16)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 272)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 528)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 784)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 48)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 304)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 560)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 816)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 17)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 273)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 529)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 785)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 49)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 305)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 561)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 817)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 18)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 274)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 530)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 786)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 50)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 306)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 562)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 818)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 19)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 275)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 531)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 787)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 51)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 307)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 563)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 819)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 20)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 276)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 532)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 788)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 52)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 308)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 564)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 820)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 21)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 277)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 533)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 789)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 53)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 309)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 565)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 821)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 22)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 278)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 534)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 790)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 54)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 310)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 566)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 822)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 23)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 279)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 535)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 791)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 55)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 311)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 567)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 823)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 24)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 280)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 536)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 792)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 56)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 312)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 568)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 824)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 25)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 281)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 537)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 793)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 57)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 313)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 569)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 825)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 26)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 282)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 538)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 794)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 58)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 314)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 570)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 826)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 27)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 283)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 539)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 795)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 59)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 315)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 571)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 827)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 28)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 284)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 540)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 796)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 60)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 316)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 572)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 828)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 29)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 285)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 541)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 797)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 61)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 317)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 573)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 829)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 30)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 286)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 542)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 798)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 62)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 318)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 574)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 830)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 31)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 287)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 543)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 799)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 63)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 319)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 575)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 831)]))
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[threadIdx.x_1] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) - 7)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 196)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 189)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 392)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 385)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 588)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 581)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 784)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 777)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 980)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 973)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1176)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1169)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1372)] = @tir.if_then_else(((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1365)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[threadIdx.x_2] = kernel[((((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 32)*4608)) + cse_var_2) + (floormod(threadIdx.x_2, 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 196)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 196), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 4), 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 392)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 392), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 8), 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 588)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 588), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 12), 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 784)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 784), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 16), 32)*9)) + cse_var_1) + 1)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ if @tir.likely((threadIdx.x_2 < 44), dtype=bool) {
+ kernel.shared_1[(threadIdx.x_2 + 980)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 980), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 20), 32)*9)) + cse_var_1) + 1)]
}
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[(floordiv(threadIdx.x, 49)*64)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 256)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 512)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 768)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 32)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 288)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 544)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 800)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 1)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 257)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 513)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 769)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 33)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 289)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 545)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 801)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 2)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 258)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 514)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 770)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 34)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 290)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 546)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 802)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 3)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 259)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 515)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 771)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 35)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 291)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 547)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 803)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 4)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 260)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 516)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 772)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 36)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 292)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 548)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 804)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 5)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 261)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 517)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 773)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 37)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 293)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 549)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 805)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 6)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 262)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 518)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 774)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 38)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 294)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 550)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 806)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 7)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 263)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 519)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 775)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 39)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 295)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 551)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 807)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 8)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 264)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 520)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 776)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 40)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 296)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 552)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 808)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 9)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 265)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 521)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 777)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 41)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 297)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 553)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 809)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 10)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 266)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 522)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 778)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 42)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 298)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 554)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 810)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 11)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 267)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 523)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 779)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 43)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 299)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 555)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 811)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 12)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 268)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 524)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 780)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 44)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 300)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 556)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 812)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 13)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 269)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 525)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 781)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 45)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 301)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 557)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 813)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 14)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 270)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 526)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 782)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 46)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 302)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 558)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 814)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 15)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 271)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 527)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 783)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 47)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 303)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 559)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 815)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 16)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 272)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 528)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 784)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 48)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 304)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 560)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 816)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 17)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 273)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 529)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 785)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 49)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 305)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 561)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 817)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 18)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 274)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 530)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 786)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 50)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 306)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 562)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 818)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 19)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 275)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 531)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 787)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 51)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 307)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 563)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 819)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 20)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 276)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 532)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 788)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 52)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 308)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 564)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 820)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 21)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 277)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 533)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 789)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 53)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 309)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 565)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 821)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 22)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 278)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 534)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 790)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 54)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 310)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 566)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 822)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 23)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 279)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 535)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 791)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 55)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 311)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 567)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 823)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 24)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 280)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 536)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 792)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 56)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 312)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 568)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 824)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 25)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 281)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 537)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 793)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 57)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 313)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 569)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 825)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 26)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 282)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 538)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 794)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 58)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 314)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 570)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 826)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 27)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 283)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 539)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 795)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 59)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 315)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 571)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 827)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 28)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 284)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 540)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 796)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 60)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 316)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 572)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 828)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 29)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 285)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 541)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 797)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 61)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 317)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 573)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 829)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 30)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 286)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 542)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 798)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 62)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 318)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 574)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 830)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 31)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 287)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 543)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 799)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 63)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 319)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 575)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 831)]))
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[threadIdx.x_1] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) - 6)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 196)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 190)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 392)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 386)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 588)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 582)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 784)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 778)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 980)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 974)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1176)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1170)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ pad_temp.shared_1[(threadIdx.x_1 + 1372)] = @tir.if_then_else((((1 <= (floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 49), 7) + ry.outer.outer) < 8)) && (floormod(threadIdx.x_1, 7) < 6)), data[((((rc.outer.outer*1568) + (ry.outer.outer*7)) + threadIdx.x_1) + 1366)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[threadIdx.x_2] = kernel[((((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 32)*4608)) + cse_var_2) + (floormod(threadIdx.x_2, 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 196)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 196), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 4), 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 392)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 392), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 8), 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 588)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 588), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 12), 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ kernel.shared_1[(threadIdx.x_2 + 784)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 784), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 16), 32)*9)) + cse_var_1) + 2)]
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 196;
+ if @tir.likely((threadIdx.x_2 < 44), dtype=bool) {
+ kernel.shared_1[(threadIdx.x_2 + 980)] = kernel[((((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 980), 32)*4608)) + cse_var_2) + (floormod((threadIdx.x_2 + 20), 32)*9)) + cse_var_1) + 2)]
+ }
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[(floordiv(threadIdx.x, 49)*64)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 256)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 512)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 768)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 32)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 288)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 544)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[floormod(threadIdx.x, 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 800)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 1)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 257)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 513)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 769)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 33)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 289)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 545)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 49)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 801)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 2)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 258)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 514)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 770)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 34)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 290)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 546)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 802)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 3)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 259)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 515)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 771)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 35)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 291)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 547)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 147)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 803)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 4)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 260)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 516)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 772)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 36)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 292)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 548)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 196)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 804)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 5)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 261)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 517)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 773)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 37)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 293)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 549)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 805)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 6)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 262)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 518)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 774)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 38)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 294)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 550)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 294)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 806)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 7)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 263)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 519)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 775)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 39)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 295)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 551)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 807)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 8)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 264)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 520)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 776)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 40)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 296)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 552)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 392)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 808)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 9)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 265)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 521)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 777)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 41)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 297)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 553)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 441)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 809)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 10)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 266)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 522)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 778)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 42)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 298)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 554)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 810)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 11)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 267)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 523)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 779)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 43)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 299)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 555)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 539)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 811)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 12)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 268)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 524)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 780)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 44)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 300)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 556)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 812)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 13)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 269)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 525)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 781)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 45)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 301)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 557)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 637)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 813)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 14)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 270)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 526)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 782)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 46)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 302)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 558)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 686)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 814)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 15)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 271)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 527)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 783)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 47)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 303)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 559)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 735)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 815)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 16)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 272)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 528)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 784)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 48)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 304)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 560)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 784)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 816)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 17)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 273)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 529)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 785)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 49)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 305)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 561)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 833)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 817)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 18)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 274)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 530)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 786)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 50)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 306)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 562)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 882)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 818)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 19)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 275)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 531)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 787)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 51)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 307)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 563)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 931)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 819)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 20)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 276)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 532)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 788)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 52)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 308)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 564)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 980)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 820)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 21)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 277)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 533)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 789)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 53)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 309)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 565)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1029)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 821)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 22)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 278)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 534)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 790)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 54)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 310)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 566)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1078)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 822)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 23)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 279)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 535)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 791)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 55)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 311)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 567)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1127)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 823)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 24)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 280)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 536)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 792)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 56)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 312)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 568)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1176)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 824)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 25)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 281)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 537)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 793)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 57)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 313)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 569)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1225)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 825)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 26)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 282)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 538)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 794)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 58)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 314)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 570)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1274)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 826)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 27)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 283)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 539)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 795)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 59)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 315)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 571)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1323)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 827)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 28)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 284)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 540)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 796)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 60)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 316)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 572)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1372)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 828)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 29)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 285)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 541)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 797)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 61)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 317)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 573)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1421)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 829)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 30)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 286)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 542)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 798)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 62)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 318)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 574)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1470)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 830)]))
+ conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 31)]))
+ conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 287)]))
+ conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 543)]))
+ conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 799)]))
+ conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 63)]))
+ conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 319)]))
+ conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 575)]))
+ conv2d_nchw_1[7] = (conv2d_nchw_1[7] + (pad_temp.shared_1[(floormod(threadIdx.x, 49) + 1519)]*kernel.shared_1[((floordiv(threadIdx.x, 49)*64) + 831)]))
}
}
}
- for (i1.inner: int32, 0, 4) {
- compute[((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*196)) + (i1.inner*49)) + floormod(threadIdx.x, 49))] = max((conv2d_nchw_1[i1.inner] + bias[(((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*4)) + i1.inner)]), 0f32)
- compute[(((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*196)) + (i1.inner*49)) + floormod(threadIdx.x, 49)) + 784)] = max((conv2d_nchw_1[(i1.inner + 4)] + bias[((((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*4)) + i1.inner) + 16)]), 0f32)
+ for (i1.inner: int32, 0, 2) {
+ compute[((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*98)) + (i1.inner*49)) + floormod(threadIdx.x, 49))] = max((conv2d_nchw_1[i1.inner] + bias[(((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*2)) + i1.inner)]), 0f32)
+ compute[(((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*98)) + (i1.inner*49)) + floormod(threadIdx.x, 49)) + 392)] = max((conv2d_nchw_1[(i1.inner + 2)] + bias[((((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*2)) + i1.inner) + 8)]), 0f32)
+ compute[(((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*98)) + (i1.inner*49)) + floormod(threadIdx.x, 49)) + 784)] = max((conv2d_nchw_1[(i1.inner + 4)] + bias[((((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*2)) + i1.inner) + 16)]), 0f32)
+ compute[(((((blockIdx.x*1568) + (floordiv(threadIdx.x, 49)*98)) + (i1.inner*49)) + floormod(threadIdx.x, 49)) + 1176)] = max((conv2d_nchw_1[(i1.inner + 6)] + bias[((((blockIdx.x*32) + (floordiv(threadIdx.x, 49)*2)) + i1.inner) + 24)]), 0f32)
}
}
}
@@ -576,7 +1411,7 @@ cooperative fetching, unrolling and operator fusion.</p>
<span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 0.366 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 0.243 ms
</pre></div>
</div>
</div>
@@ -605,10 +1440,10 @@ conv2d_nchw_nn_o_i, conv2d_nchw_nn_i = s[conv2d_nchw].split(conv2d_nchw_nn, fact
conv2d_nchw_nn_o_o_i, conv2d_nchw_nn_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_i, factor=1)
conv2d_nchw_nn_o_o_o_i, conv2d_nchw_nn_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_i, factor=1)
conv2d_nchw_nn_o_o_o_o, conv2d_nchw_nn_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_o_i, factor=1)
-conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=2)
+conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=1)
conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=2)
conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=4)
-conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=2)
+conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=4)
conv2d_nchw_yy_o_i, conv2d_nchw_yy_i = s[conv2d_nchw].split(conv2d_nchw_yy, factor=1)
conv2d_nchw_yy_o_o_i, conv2d_nchw_yy_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_i, factor=1)
conv2d_nchw_yy_o_o_o_i, conv2d_nchw_yy_o_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_o_i, factor=7)
@@ -617,19 +1452,19 @@ conv2d_nchw_xx_o_i, conv2d_nchw_xx_i = s[conv2d_nchw].split(conv2d_nchw_xx, fact
conv2d_nchw_xx_o_o_i, conv2d_nchw_xx_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_i, factor=1)
conv2d_nchw_xx_o_o_o_i, conv2d_nchw_xx_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_i, factor=7)
conv2d_nchw_xx_o_o_o_o, conv2d_nchw_xx_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_o_i, factor=1)
-conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=8)
-conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=2)
+conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=1)
+conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=32)
conv2d_nchw_ry_o_i, conv2d_nchw_ry_i = s[conv2d_nchw].split(conv2d_nchw_ry, factor=1)
conv2d_nchw_ry_o_o, conv2d_nchw_ry_o_i = s[conv2d_nchw].split(conv2d_nchw_ry_o_i, factor=1)
-conv2d_nchw_rx_o_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=3)
+conv2d_nchw_rx_o_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=1)
conv2d_nchw_rx_o_o, conv2d_nchw_rx_o_i = s[conv2d_nchw].split(conv2d_nchw_rx_o_i, factor=1)
s[conv2d_nchw].reorder(conv2d_nchw_nn_o_o_o_o, conv2d_nchw_ff_o_o_o_o, conv2d_nchw_yy_o_o_o_o, conv2d_nchw_xx_o_o_o_o, conv2d_nchw_nn_o_o_o_i, conv2d_nchw_ff_o_o_o_i, conv2d_nchw_yy_o_o_o_i, conv2d_nchw_xx_o_o_o_i, conv2d_nchw_nn_o_o_i, conv2d_nchw_ff_o_o_i, conv2d_nchw_yy_o_o_i, conv2d_nchw_xx_o_o_i, conv2d_nchw_rc_o_o, conv2d_nchw_ry_o_o, conv2d_nchw_rx_o_o, conv2d_nchw_rc_o_i, conv2d_nchw_ry_o_i, conv2d_nchw_rx_o_i, conv2d_nchw_nn_o_i, conv2d_nchw_ff_o_i, conv2d_nchw_yy_o_i, conv2d_nc [...]
compute_i0_o_i, compute_i0_i = s[compute].split(compute_i0, factor=1)
compute_i0_o_o_i, compute_i0_o_i = s[compute].split(compute_i0_o_i, factor=1)
compute_i0_o_o_o, compute_i0_o_o_i = s[compute].split(compute_i0_o_o_i, factor=1)
-compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=4)
+compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=2)
compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=4)
-compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=2)
+compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=4)
compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=7)
compute_i2_o_o_o, compute_i2_o_o_i = s[compute].split(compute_i2_o_o_i, factor=1)
@@ -661,7 +1496,7 @@ pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, pad_temp_shared_ax0_ax1_fus
s[pad_temp_shared].vectorize(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=196)
s[pad_temp_shared].bind(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
-s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 0)
+s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 1024)
s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
@@ -681,45 +1516,847 @@ CUDA source code:
#endif
extern "C" __global__ void __launch_bounds__(196) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
float conv2d_nchw[8];
- __shared__ float pad_temp_shared[1008];
- __shared__ float kernel_shared[1536];
- for (int ff_outer_inner_init = 0; ff_outer_inner_init < 2; ++ff_outer_inner_init) {
- for (int ff_inner_init = 0; ff_inner_init < 2; ++ff_inner_init) {
- conv2d_nchw[((ff_outer_inner_init * 2) + ff_inner_init)] = 0.000000e+00f;
- conv2d_nchw[(((ff_outer_inner_init * 2) + ff_inner_init) + 4)] = 0.000000e+00f;
- }
- }
- for (int rc_outer_outer = 0; rc_outer_outer < 32; ++rc_outer_outer) {
+ __shared__ float pad_temp_shared[1568];
+ __shared__ float kernel_shared[1024];
+ conv2d_nchw[0] = 0.000000e+00f;
+ conv2d_nchw[2] = 0.000000e+00f;
+ conv2d_nchw[4] = 0.000000e+00f;
+ conv2d_nchw[6] = 0.000000e+00f;
+ conv2d_nchw[1] = 0.000000e+00f;
+ conv2d_nchw[3] = 0.000000e+00f;
+ conv2d_nchw[5] = 0.000000e+00f;
+ conv2d_nchw[7] = 0.000000e+00f;
+ for (int rc_outer_outer = 0; rc_outer_outer < 16; ++rc_outer_outer) {
for (int ry_outer_outer = 0; ry_outer_outer < 3; ++ry_outer_outer) {
__syncthreads();
- for (int ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer = 0; ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer < 6; ++ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer) {
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) / 28)) < 36) {
- pad_temp_shared[((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 196) + ((int)threadIdx.x))] = (((((1 <= (((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + ((int)threadIdx.x)) % 63) / 9) + ry_outer_outer)) && ((((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + ((int)threadIdx.x)) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + ((int)threadIdx.x)) % 9))) && ((((ax0_ax1_fused_ax2_f [...]
- }
+ pad_temp_shared[((int)threadIdx.x)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 196)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 188)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 392)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 384)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 588)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 580)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 776)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 980)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 972)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1176)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1168)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1372)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 7))) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1364)] : 0.000000e+00f);
+ kernel_shared[((int)threadIdx.x)] = kernel[(((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) >> 5) * 4608)) + (rc_outer_outer * 288)) + ((((int)threadIdx.x) & 31) * 9)) + (ry_outer_outer * 3))];
+ kernel_shared[(((int)threadIdx.x) + 196)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 196) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 4) & 31) * 9)) + (ry_outer_outer * 3))];
+ kernel_shared[(((int)threadIdx.x) + 392)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 392) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 8) & 31) * 9)) + (ry_outer_outer * 3))];
+ kernel_shared[(((int)threadIdx.x) + 588)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 588) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 12) & 31) * 9)) + (ry_outer_outer * 3))];
+ kernel_shared[(((int)threadIdx.x) + 784)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 784) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 16) & 31) * 9)) + (ry_outer_outer * 3))];
+ if (((int)threadIdx.x) < 44) {
+ kernel_shared[(((int)threadIdx.x) + 980)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 980) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 20) & 31) * 9)) + (ry_outer_outer * 3))];
}
- for (int ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 = 0; ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 < 8; ++ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1) {
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 * 49) + (((int)threadIdx.x) >> 2)) < 384) {
- kernel_shared[((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 * 196) + ((int)threadIdx.x))] = kernel[((((((((int)blockIdx.x) * 147456) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 * 49) + (((int)threadIdx.x) >> 2)) / 12) * 4608)) + (rc_outer_outer * 144)) + (((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1 * 4) + ((int)threadIdx.x)) % 48) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer_1) % 3))];
- }
+ __syncthreads();
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[((((int)threadIdx.x) / 49) * 64)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 256)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 512)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 768)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 32)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 288)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 544)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 800)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 1)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 257)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 513)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 769)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 33)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 289)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 545)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 801)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 2)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 258)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 514)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 770)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 34)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 290)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 546)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 802)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 3)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 259)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 515)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 771)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 35)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 291)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 547)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 803)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 4)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 260)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 516)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 772)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 36)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 292)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 548)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 804)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 5)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 261)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 517)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 773)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 37)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 293)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 549)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 805)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 6)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 262)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 518)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 774)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 38)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 294)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 550)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 806)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 7)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 263)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 519)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 775)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 39)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 295)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 551)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 807)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 8)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 264)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 520)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 776)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 40)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 296)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 552)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 808)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 9)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 265)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 521)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 777)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 41)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 297)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 553)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 809)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 10)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 266)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 522)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 778)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 42)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 298)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 554)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 810)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 11)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 267)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 523)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 779)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 43)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 299)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 555)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 811)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 12)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 268)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 524)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 780)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 44)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 300)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 556)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 812)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 13)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 269)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 525)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 781)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 45)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 301)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 557)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 813)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 14)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 270)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 526)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 782)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 46)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 302)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 558)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 814)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 15)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 271)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 527)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 783)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 47)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 303)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 559)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 815)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 16)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 272)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 528)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 784)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 48)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 304)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 560)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 816)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 17)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 273)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 529)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 785)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 49)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 305)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 561)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 817)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 18)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 274)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 530)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 786)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 50)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 306)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 562)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 818)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 19)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 275)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 531)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 787)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 51)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 307)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 563)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 819)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 20)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 276)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 532)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 788)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 52)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 308)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 564)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 820)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 21)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 277)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 533)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 789)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 53)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 309)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 565)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 821)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 22)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 278)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 534)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 790)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 54)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 310)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 566)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 822)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 23)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 279)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 535)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 791)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 55)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 311)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 567)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 823)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 24)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 280)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 536)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 792)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 56)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 312)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 568)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 824)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 25)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 281)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 537)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 793)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 57)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 313)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 569)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 825)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 26)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 282)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 538)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 794)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 58)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 314)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 570)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 826)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 27)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 283)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 539)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 795)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 59)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 315)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 571)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 827)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 28)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 284)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 540)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 796)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 60)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 316)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 572)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 828)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 29)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 285)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 541)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 797)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 61)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 317)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 573)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 829)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 30)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 286)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 542)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 798)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 62)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 318)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 574)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 830)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 31)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 287)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 543)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 799)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 63)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 319)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 575)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 831)]));
+ __syncthreads();
+ pad_temp_shared[((int)threadIdx.x)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) - 7)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 196)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 189)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 392)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 385)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 588)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 581)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 777)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 980)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 973)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1176)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1169)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1372)] = (((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1365)] : 0.000000e+00f);
+ kernel_shared[((int)threadIdx.x)] = kernel[((((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) >> 5) * 4608)) + (rc_outer_outer * 288)) + ((((int)threadIdx.x) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ kernel_shared[(((int)threadIdx.x) + 196)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 196) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 4) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ kernel_shared[(((int)threadIdx.x) + 392)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 392) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 8) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ kernel_shared[(((int)threadIdx.x) + 588)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 588) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 12) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ kernel_shared[(((int)threadIdx.x) + 784)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 784) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 16) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
+ if (((int)threadIdx.x) < 44) {
+ kernel_shared[(((int)threadIdx.x) + 980)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 980) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 20) & 31) * 9)) + (ry_outer_outer * 3)) + 1)];
}
__syncthreads();
- for (int rc_outer_inner = 0; rc_outer_inner < 2; ++rc_outer_inner) {
- for (int ff_outer_inner = 0; ff_outer_inner < 2; ++ff_outer_inner) {
- for (int rc_inner = 0; rc_inner < 8; ++rc_inner) {
- for (int rx_inner = 0; rx_inner < 3; ++rx_inner) {
- for (int ff_inner = 0; ff_inner < 2; ++ff_inner) {
- conv2d_nchw[((ff_outer_inner * 2) + ff_inner)] = (conv2d_nchw[((ff_outer_inner * 2) + ff_inner)] + (pad_temp_shared[(((((rc_outer_inner * 504) + (rc_inner * 63)) + (((((int)threadIdx.x) % 49) / 7) * 9)) + rx_inner) + (((int)threadIdx.x) % 7))] * kernel_shared[(((((((((int)threadIdx.x) / 49) * 192) + (ff_outer_inner * 96)) + (ff_inner * 48)) + (rc_outer_inner * 24)) + (rc_inner * 3)) + rx_inner)]));
- conv2d_nchw[(((ff_outer_inner * 2) + ff_inner) + 4)] = (conv2d_nchw[(((ff_outer_inner * 2) + ff_inner) + 4)] + (pad_temp_shared[(((((rc_outer_inner * 504) + (rc_inner * 63)) + (((((int)threadIdx.x) % 49) / 7) * 9)) + rx_inner) + (((int)threadIdx.x) % 7))] * kernel_shared[((((((((((int)threadIdx.x) / 49) * 192) + (ff_outer_inner * 96)) + (ff_inner * 48)) + (rc_outer_inner * 24)) + (rc_inner * 3)) + rx_inner) + 768)]));
- }
- }
- }
- }
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[((((int)threadIdx.x) / 49) * 64)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 256)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 512)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 768)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 32)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 288)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 544)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 800)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 1)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 257)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 513)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 769)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 33)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 289)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 545)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 801)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 2)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 258)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 514)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 770)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 34)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 290)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 546)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 802)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 3)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 259)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 515)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 771)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 35)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 291)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 547)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 803)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 4)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 260)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 516)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 772)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 36)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 292)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 548)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 804)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 5)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 261)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 517)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 773)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 37)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 293)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 549)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 805)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 6)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 262)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 518)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 774)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 38)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 294)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 550)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 806)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 7)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 263)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 519)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 775)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 39)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 295)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 551)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 807)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 8)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 264)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 520)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 776)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 40)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 296)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 552)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 808)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 9)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 265)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 521)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 777)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 41)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 297)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 553)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 809)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 10)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 266)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 522)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 778)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 42)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 298)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 554)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 810)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 11)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 267)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 523)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 779)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 43)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 299)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 555)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 811)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 12)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 268)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 524)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 780)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 44)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 300)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 556)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 812)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 13)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 269)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 525)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 781)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 45)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 301)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 557)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 813)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 14)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 270)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 526)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 782)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 46)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 302)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 558)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 814)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 15)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 271)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 527)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 783)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 47)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 303)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 559)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 815)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 16)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 272)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 528)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 784)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 48)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 304)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 560)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 816)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 17)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 273)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 529)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 785)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 49)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 305)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 561)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 817)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 18)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 274)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 530)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 786)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 50)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 306)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 562)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 818)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 19)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 275)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 531)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 787)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 51)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 307)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 563)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 819)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 20)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 276)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 532)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 788)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 52)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 308)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 564)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 820)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 21)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 277)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 533)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 789)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 53)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 309)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 565)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 821)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 22)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 278)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 534)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 790)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 54)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 310)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 566)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 822)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 23)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 279)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 535)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 791)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 55)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 311)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 567)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 823)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 24)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 280)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 536)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 792)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 56)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 312)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 568)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 824)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 25)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 281)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 537)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 793)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 57)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 313)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 569)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 825)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 26)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 282)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 538)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 794)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 58)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 314)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 570)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 826)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 27)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 283)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 539)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 795)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 59)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 315)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 571)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 827)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 28)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 284)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 540)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 796)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 60)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 316)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 572)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 828)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 29)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 285)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 541)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 797)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 61)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 317)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 573)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 829)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 30)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 286)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 542)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 798)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 62)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 318)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 574)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 830)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 31)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 287)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 543)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 799)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 63)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 319)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 575)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 831)]));
+ __syncthreads();
+ pad_temp_shared[((int)threadIdx.x)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) - 6)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 196)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 190)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 392)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 386)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 588)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 582)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 778)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 980)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 974)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1176)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1170)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1372)] = ((((1 <= (((((int)threadIdx.x) % 49) / 7) + ry_outer_outer)) && ((((((int)threadIdx.x) % 49) / 7) + ry_outer_outer) < 8)) && ((((int)threadIdx.x) % 7) < 6)) ? data[((((rc_outer_outer * 1568) + (ry_outer_outer * 7)) + ((int)threadIdx.x)) + 1366)] : 0.000000e+00f);
+ kernel_shared[((int)threadIdx.x)] = kernel[((((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) >> 5) * 4608)) + (rc_outer_outer * 288)) + ((((int)threadIdx.x) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ kernel_shared[(((int)threadIdx.x) + 196)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 196) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 4) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ kernel_shared[(((int)threadIdx.x) + 392)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 392) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 8) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ kernel_shared[(((int)threadIdx.x) + 588)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 588) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 12) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ kernel_shared[(((int)threadIdx.x) + 784)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 784) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 16) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
+ if (((int)threadIdx.x) < 44) {
+ kernel_shared[(((int)threadIdx.x) + 980)] = kernel[((((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 980) >> 5) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 20) & 31) * 9)) + (ry_outer_outer * 3)) + 2)];
}
+ __syncthreads();
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[((((int)threadIdx.x) / 49) * 64)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 256)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 512)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 768)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 32)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 288)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 544)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((int)threadIdx.x) % 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 800)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 1)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 257)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 513)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 769)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 33)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 289)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 545)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 49)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 801)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 2)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 258)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 514)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 770)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 34)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 290)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 546)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 98)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 802)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 3)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 259)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 515)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 771)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 35)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 291)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 547)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 147)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 803)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 4)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 260)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 516)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 772)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 36)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 292)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 548)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 196)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 804)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 5)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 261)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 517)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 773)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 37)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 293)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 549)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 245)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 805)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 6)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 262)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 518)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 774)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 38)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 294)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 550)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 294)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 806)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 7)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 263)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 519)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 775)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 39)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 295)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 551)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 343)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 807)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 8)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 264)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 520)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 776)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 40)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 296)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 552)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 392)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 808)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 9)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 265)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 521)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 777)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 41)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 297)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 553)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 441)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 809)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 10)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 266)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 522)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 778)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 42)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 298)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 554)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 490)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 810)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 11)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 267)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 523)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 779)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 43)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 299)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 555)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 539)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 811)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 12)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 268)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 524)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 780)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 44)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 300)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 556)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 588)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 812)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 13)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 269)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 525)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 781)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 45)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 301)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 557)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 637)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 813)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 14)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 270)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 526)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 782)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 46)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 302)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 558)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 686)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 814)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 15)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 271)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 527)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 783)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 47)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 303)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 559)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 735)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 815)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 16)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 272)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 528)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 784)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 48)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 304)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 560)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 784)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 816)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 17)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 273)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 529)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 785)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 49)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 305)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 561)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 833)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 817)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 18)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 274)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 530)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 786)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 50)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 306)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 562)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 882)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 818)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 19)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 275)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 531)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 787)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 51)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 307)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 563)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 931)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 819)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 20)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 276)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 532)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 788)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 52)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 308)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 564)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 980)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 820)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 21)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 277)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 533)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 789)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 53)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 309)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 565)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1029)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 821)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 22)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 278)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 534)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 790)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 54)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 310)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 566)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1078)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 822)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 23)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 279)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 535)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 791)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 55)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 311)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 567)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1127)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 823)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 24)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 280)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 536)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 792)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 56)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 312)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 568)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1176)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 824)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 25)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 281)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 537)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 793)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 57)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 313)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 569)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1225)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 825)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 26)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 282)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 538)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 794)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 58)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 314)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 570)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1274)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 826)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 27)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 283)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 539)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 795)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 59)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 315)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 571)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1323)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 827)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 28)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 284)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 540)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 796)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 60)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 316)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 572)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1372)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 828)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 29)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 285)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 541)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 797)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 61)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 317)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 573)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1421)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 829)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 30)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 286)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 542)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 798)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 62)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 318)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 574)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1470)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 830)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 31)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 287)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 543)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 799)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 63)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 319)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 575)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 49) + 1519)] * kernel_shared[(((((int)threadIdx.x) / 49) * 64) + 831)]));
}
}
- for (int i1_inner = 0; i1_inner < 4; ++i1_inner) {
- compute[((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 196)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49))] = max((conv2d_nchw[i1_inner] + bias[(((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 4)) + i1_inner)]), 0.000000e+00f);
- compute[(((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 196)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49)) + 784)] = max((conv2d_nchw[(i1_inner + 4)] + bias[((((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 4)) + i1_inner) + 16)]), 0.000000e+00f);
+ for (int i1_inner = 0; i1_inner < 2; ++i1_inner) {
+ compute[((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 98)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49))] = max((conv2d_nchw[i1_inner] + bias[(((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 2)) + i1_inner)]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 98)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49)) + 392)] = max((conv2d_nchw[(i1_inner + 2)] + bias[((((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 2)) + i1_inner) + 8)]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 98)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49)) + 784)] = max((conv2d_nchw[(i1_inner + 4)] + bias[((((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 2)) + i1_inner) + 16)]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) * 1568) + ((((int)threadIdx.x) / 49) * 98)) + (i1_inner * 49)) + (((int)threadIdx.x) % 49)) + 1176)] = max((conv2d_nchw[(i1_inner + 6)] + bias[((((((int)blockIdx.x) * 32) + ((((int)threadIdx.x) / 49) * 2)) + i1_inner) + 24)]), 0.000000e+00f);
}
}
</pre></div>
@@ -756,7 +2393,7 @@ In the example below we resume the status and do more 5 trials.</p>
Get devices for measurement successfully!
</pre></div>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes 17.057 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes 25.235 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-tune-with-autoscheduler-tune-conv2d-layer-cuda-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/e3e540f3b477c0c52d8eb73e674e8ffd/tune_conv2d_layer_cuda.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">tune_conv2d_layer_cuda.py</span></code></a></p>
diff --git a/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html b/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
index 05b3aa684..7e330ec82 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
@@ -906,7 +906,7 @@ so we can read the log file and load the best schedules.</p>
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 9.8031 9.8079 9.8367 9.7646 0.0296
+ 9.9558 9.9761 10.0060 9.8852 0.0513
</pre></div>
</div>
</div>
diff --git a/docs/how_to/tune_with_autoscheduler/tune_network_x86.html b/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
index b6a971d30..60967fcc9 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
@@ -925,7 +925,7 @@ so we can read the log file and load the best schedules.</p>
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 741.7649 742.4460 742.8278 740.0208 1.2431
+ 754.8454 754.7219 755.2403 754.5740 0.2857
</pre></div>
</div>
</div>
@@ -947,7 +947,7 @@ to learn how to use the RPC Tracker and RPC Server.
To use the RPC Tracker in auto-scheduler, replace the runner in <code class="code docutils literal notranslate"><span class="pre">TuningOptions</span></code>
with <a class="reference internal" href="../../reference/api/python/auto_scheduler.html#tvm.auto_scheduler.RPCRunner" title="tvm.auto_scheduler.RPCRunner"><code class="xref any py py-class docutils literal notranslate"><span class="pre">auto_scheduler.RPCRunner</span></code></a>.</p></li>
</ol>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 20.723 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 22.299 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-tune-with-autoscheduler-tune-network-x86-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/e416b94ca1090b0897c0f6e0df95b911/tune_network_x86.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">tune_network_x86.py</span></code></a></p>
diff --git a/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html b/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
index 1868d98fb..2d992bfa7 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
@@ -625,77 +625,103 @@ layout transformation, parallelization, vectorization, unrolling, and operator f
placeholder_4: Buffer(placeholder_14: Pointer(float32), float32, [65536], []),
compute: Buffer(compute_2: Pointer(float32), float32, [65536], [])}
buffer_map = {placeholder_5: placeholder, placeholder_6: placeholder_1, placeholder_7: placeholder_2, placeholder_8: placeholder_3, placeholder_9: placeholder_4, compute_1: compute}
- preflattened_buffer_map = {placeholder_6: placeholder_15: Buffer(placeholder_11, float32, [4916, 16, 1], []), placeholder_9: placeholder_16: Buffer(placeholder_14, float32, [128, 512], []), placeholder_7: placeholder_17: Buffer(placeholder_12, int32, [4916], []), placeholder_5: placeholder_18: Buffer(placeholder_10, float32, [128, 256], []), placeholder_8: placeholder_19: Buffer(placeholder_13, int32, [33], []), compute_1: compute_3: Buffer(compute_2, float32, [128, 512], [])} {
- for (i0.outer.i1.outer.fused: int32, 0, 128) "parallel" {
- allocate(compute_4: Pointer(global float32), float32, [512]), storage_scope = global {
- for (nb_j.inner: int32, 0, 2) {
- for (i.inner.init: int32, 0, 16) {
- let cse_var_1: int32 = ((i.inner.init*32) + (nb_j.inner*16))
- {
- compute_5: Buffer(compute_4, float32, [512], [])[cse_var_1] = 0f32
- compute_5[(cse_var_1 + 1)] = 0f32
- compute_5[(cse_var_1 + 2)] = 0f32
- compute_5[(cse_var_1 + 3)] = 0f32
- compute_5[(cse_var_1 + 4)] = 0f32
- compute_5[(cse_var_1 + 5)] = 0f32
- compute_5[(cse_var_1 + 6)] = 0f32
- compute_5[(cse_var_1 + 7)] = 0f32
- compute_5[(cse_var_1 + 8)] = 0f32
- compute_5[(cse_var_1 + 9)] = 0f32
- compute_5[(cse_var_1 + 10)] = 0f32
- compute_5[(cse_var_1 + 11)] = 0f32
- compute_5[(cse_var_1 + 12)] = 0f32
- compute_5[(cse_var_1 + 13)] = 0f32
- compute_5[(cse_var_1 + 14)] = 0f32
- compute_5[(cse_var_1 + 15)] = 0f32
- }
+ preflattened_buffer_map = {placeholder_5: placeholder_15: Buffer(placeholder_10, float32, [128, 256], []), placeholder_6: placeholder_16: Buffer(placeholder_11, float32, [4916, 16, 1], []), placeholder_7: placeholder_17: Buffer(placeholder_12, int32, [4916], []), placeholder_8: placeholder_18: Buffer(placeholder_13, int32, [33], []), placeholder_9: placeholder_19: Buffer(placeholder_14, float32, [128, 512], []), compute_1: compute_3: Buffer(compute_2, float32, [128, 512], [])} {
+ allocate(compute_4: Pointer(global float32), float32, [2048]), storage_scope = global;
+ for (i1.outer: int32, 0, 32) {
+ for (i.outer.inner: int32, 0, 4) {
+ for (i.inner.init: int32, 0, 32) {
+ let cse_var_1: int32 = ((i.outer.inner*512) + (i.inner.init*16))
+ {
+ compute_5: Buffer(compute_4, float32, [2048], [])[cse_var_1] = 0f32
+ compute_5[(cse_var_1 + 1)] = 0f32
+ compute_5[(cse_var_1 + 2)] = 0f32
+ compute_5[(cse_var_1 + 3)] = 0f32
+ compute_5[(cse_var_1 + 4)] = 0f32
+ compute_5[(cse_var_1 + 5)] = 0f32
+ compute_5[(cse_var_1 + 6)] = 0f32
+ compute_5[(cse_var_1 + 7)] = 0f32
+ compute_5[(cse_var_1 + 8)] = 0f32
+ compute_5[(cse_var_1 + 9)] = 0f32
+ compute_5[(cse_var_1 + 10)] = 0f32
+ compute_5[(cse_var_1 + 11)] = 0f32
+ compute_5[(cse_var_1 + 12)] = 0f32
+ compute_5[(cse_var_1 + 13)] = 0f32
+ compute_5[(cse_var_1 + 14)] = 0f32
+ compute_5[(cse_var_1 + 15)] = 0f32
}
- for (elem_idx: int32, 0, let cse_var_2: int32 = ((floormod(i0.outer.i1.outer.fused, 16)*2) + nb_j.inner) in (placeholder_3[(cse_var_2 + 1)] - placeholder_3[cse_var_2])) {
- for (i.inner: int32, 0, 16) {
- let cse_var_21: int32 = (elem_idx*16)
- let cse_var_20: int32 = ((i.inner*32) + (nb_j.inner*16))
- let cse_var_19: int32 = ((floormod(i0.outer.i1.outer.fused, 16)*2) + nb_j.inner)
- let cse_var_18: int32 = ((floordiv(i0.outer.i1.outer.fused, 16)*4096) + (i.inner*256))
- let cse_var_17: int32 = (cse_var_20 + 9)
- let cse_var_16: int32 = (cse_var_20 + 8)
- let cse_var_15: int32 = (cse_var_20 + 7)
- let cse_var_14: int32 = (cse_var_20 + 6)
- let cse_var_13: int32 = (cse_var_20 + 5)
- let cse_var_12: int32 = (cse_var_20 + 4)
- let cse_var_11: int32 = (cse_var_20 + 3)
- let cse_var_10: int32 = (cse_var_20 + 2)
- let cse_var_9: int32 = (cse_var_20 + 15)
- let cse_var_8: int32 = (cse_var_20 + 14)
- let cse_var_7: int32 = (cse_var_20 + 13)
- let cse_var_6: int32 = (cse_var_20 + 12)
- let cse_var_5: int32 = (cse_var_20 + 11)
- let cse_var_4: int32 = (cse_var_20 + 10)
- let cse_var_3: int32 = (cse_var_20 + 1)
- {
- compute_5[cse_var_20] = (compute_5[cse_var_20] + (placeholder_1[((placeholder_3[cse_var_19]*16) + cse_var_21)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_3] = (compute_5[cse_var_3] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 1)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_10] = (compute_5[cse_var_10] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 2)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_11] = (compute_5[cse_var_11] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 3)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_12] = (compute_5[cse_var_12] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 4)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_13] = (compute_5[cse_var_13] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 5)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_14] = (compute_5[cse_var_14] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 6)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_15] = (compute_5[cse_var_15] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 7)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_16] = (compute_5[cse_var_16] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 8)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_17] = (compute_5[cse_var_17] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 9)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_4] = (compute_5[cse_var_4] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 10)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_5] = (compute_5[cse_var_5] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 11)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_6] = (compute_5[cse_var_6] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 12)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_7] = (compute_5[cse_var_7] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 13)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_8] = (compute_5[cse_var_8] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 14)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- compute_5[cse_var_9] = (compute_5[cse_var_9] + (placeholder_1[(((placeholder_3[cse_var_19]*16) + cse_var_21) + 15)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_19] + elem_idx)])], 0f32)))
- }
+ }
+ for (elem_idx: int32, 0, (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])) {
+ for (i.inner: int32, 0, 32) {
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_2: int32 = ((i.outer.inner*512) + (i.inner*16))
+ compute_5[cse_var_2] = (compute_5[cse_var_2] + (placeholder_1[((placeholder_3[i1.outer]*16) + (elem_idx*16))]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_3: int32 = (((i.outer.inner*512) + (i.inner*16)) + 1)
+ compute_5[cse_var_3] = (compute_5[cse_var_3] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 1)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_4: int32 = (((i.outer.inner*512) + (i.inner*16)) + 2)
+ compute_5[cse_var_4] = (compute_5[cse_var_4] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 2)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_5: int32 = (((i.outer.inner*512) + (i.inner*16)) + 3)
+ compute_5[cse_var_5] = (compute_5[cse_var_5] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 3)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_6: int32 = (((i.outer.inner*512) + (i.inner*16)) + 4)
+ compute_5[cse_var_6] = (compute_5[cse_var_6] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 4)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_7: int32 = (((i.outer.inner*512) + (i.inner*16)) + 5)
+ compute_5[cse_var_7] = (compute_5[cse_var_7] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 5)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_8: int32 = (((i.outer.inner*512) + (i.inner*16)) + 6)
+ compute_5[cse_var_8] = (compute_5[cse_var_8] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 6)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_9: int32 = (((i.outer.inner*512) + (i.inner*16)) + 7)
+ compute_5[cse_var_9] = (compute_5[cse_var_9] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 7)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_10: int32 = (((i.outer.inner*512) + (i.inner*16)) + 8)
+ compute_5[cse_var_10] = (compute_5[cse_var_10] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 8)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_11: int32 = (((i.outer.inner*512) + (i.inner*16)) + 9)
+ compute_5[cse_var_11] = (compute_5[cse_var_11] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 9)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_12: int32 = (((i.outer.inner*512) + (i.inner*16)) + 10)
+ compute_5[cse_var_12] = (compute_5[cse_var_12] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 10)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_13: int32 = (((i.outer.inner*512) + (i.inner*16)) + 11)
+ compute_5[cse_var_13] = (compute_5[cse_var_13] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 11)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_14: int32 = (((i.outer.inner*512) + (i.inner*16)) + 12)
+ compute_5[cse_var_14] = (compute_5[cse_var_14] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 12)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_15: int32 = (((i.outer.inner*512) + (i.inner*16)) + 13)
+ compute_5[cse_var_15] = (compute_5[cse_var_15] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 13)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_16: int32 = (((i.outer.inner*512) + (i.inner*16)) + 14)
+ compute_5[cse_var_16] = (compute_5[cse_var_16] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 14)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
+ }
+ if @tir.likely((elem_idx < (placeholder_3[(i1.outer + 1)] - placeholder_3[i1.outer])), dtype=bool) {
+ let cse_var_17: int32 = (((i.outer.inner*512) + (i.inner*16)) + 15)
+ compute_5[cse_var_17] = (compute_5[cse_var_17] + (placeholder_1[(((placeholder_3[i1.outer]*16) + (elem_idx*16)) + 15)]*max(placeholder[(((i.outer.inner*8192) + (i.inner*256)) + placeholder_2[(placeholder_3[i1.outer] + elem_idx)])], 0f32)))
}
}
}
- for (i0.inner: int32, 0, 16) {
- let cse_var_22: int32 = (((floordiv(i0.outer.i1.outer.fused, 16)*8192) + (i0.inner*512)) + (floormod(i0.outer.i1.outer.fused, 16)*32))
- compute[ramp(cse_var_22, 1, 32)] = max((compute_5[ramp((i0.inner*32), 1, 32)] + placeholder_4[ramp(cse_var_22, 1, 32)]), broadcast(0f32, 32))
- }
+ }
+ for (i0.inner: int32, 0, 128) {
+ let cse_var_18: int32 = ((i0.inner*512) + (i1.outer*16))
+ compute[ramp(cse_var_18, 1, 16)] = max((compute_5[ramp((i0.inner*16), 1, 16)] + placeholder_4[ramp(cse_var_18, 1, 16)]), broadcast(0f32, 16))
}
}
}
@@ -732,7 +758,7 @@ layout transformation, parallelization, vectorization, unrolling, and operator f
<span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 1.651 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 1.721 ms
</pre></div>
</div>
<div class="admonition note">
diff --git a/docs/how_to/tune_with_autotvm/sg_execution_times.html b/docs/how_to/tune_with_autotvm/sg_execution_times.html
index da9b6d8fb..b15ba69d3 100644
--- a/docs/how_to/tune_with_autotvm/sg_execution_times.html
+++ b/docs/how_to/tune_with_autotvm/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-tune-with-autotvm-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:44.846</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
+<p><strong>00:46.348</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -336,22 +336,22 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="tune_conv2d_cuda.html#sphx-glr-how-to-tune-with-autotvm-tune-conv2d-cuda-py"><span class="std std-ref">Tuning High Performance Convolution on NVIDIA GPUs</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_conv2d_cuda.py</span></code>)</p></td>
-<td><p>00:44.812</p></td>
+<td><p>00:46.314</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="tune_relay_x86.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-x86-py"><span class="std std-ref">Auto-tuning a Convolutional Network for x86 CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_x86.py</span></code>)</p></td>
-<td><p>00:00.019</p></td>
+<td><p>00:00.020</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="tune_relay_cuda.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-cuda-py"><span class="std std-ref">Auto-tuning a Convolutional Network for NVIDIA GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_cuda.py</span></code>)</p></td>
<td><p>00:00.005</p></td>
<td><p>0.0 MB</p></td>
</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="tune_relay_mobile_gpu.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-mobile-gpu-py"><span class="std std-ref">Auto-tuning a Convolutional Network for Mobile GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_mobile_gpu.py</span></code>)</p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="tune_relay_arm.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-arm-py"><span class="std std-ref">Auto-tuning a Convolutional Network for ARM CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_arm.py</span></code>)</p></td>
<td><p>00:00.005</p></td>
<td><p>0.0 MB</p></td>
</tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="tune_relay_arm.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-arm-py"><span class="std std-ref">Auto-tuning a Convolutional Network for ARM CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_arm.py</span></code>)</p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="tune_relay_mobile_gpu.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-mobile-gpu-py"><span class="std std-ref">Auto-tuning a Convolutional Network for Mobile GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_mobile_gpu.py</span></code>)</p></td>
<td><p>00:00.005</p></td>
<td><p>0.0 MB</p></td>
</tr>
diff --git a/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html b/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
index def5bec34..316713aeb 100644
--- a/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
+++ b/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
@@ -1436,8 +1436,8 @@ No: 8 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
TimeoutError
[('tile_f', [-1, 2, 1, 64]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4909501
-No: 9 GFLOPS: 190.31/190.31 result: MeasureResult(costs=(0.0012164683333333334,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.792853593826294, timestamp=1662973591.8327475) [('tile_f', [-1, 1, 4, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5072689
-No: 10 GFLOPS: 0.00/190.31 result: Traceback (most recent call last):
+No: 9 GFLOPS: 219.45/219.45 result: MeasureResult(costs=(0.0010549379172413794,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.907240867614746, timestamp=1663013704.4100664) [('tile_f', [-1, 1, 4, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5072689
+No: 10 GFLOPS: 0.00/219.45 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1560,8 +1560,8 @@ Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 4, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 64, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5092711
-No: 11 GFLOPS: 260.87/260.87 result: MeasureResult(costs=(0.0008874141546961326,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7072300910949707, timestamp=1662973592.7237387) [('tile_f', [-1, 8, 2, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4264713
-No: 12 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+No: 11 GFLOPS: 259.35/259.35 result: MeasureResult(costs=(0.0008926184475138122,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6609346866607666, timestamp=1663013705.284125) [('tile_f', [-1, 8, 2, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4264713
+No: 12 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1684,7 +1684,7 @@ Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 128, 1, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 256]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,183542
-No: 13 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+No: 13 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1807,7 +1807,7 @@ Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 8, 8]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 64]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2482196
-No: 14 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+No: 14 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -1930,9 +1930,9 @@ Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 64, 1, 4]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10306226
-No: 15 GFLOPS: 5.46/260.87 result: MeasureResult(costs=(0.042424332,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7882673740386963, timestamp=1662973597.2073998) [('tile_f', [-1, 2, 2, 8]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5330964
-No: 16 GFLOPS: 3.35/260.87 result: MeasureResult(costs=(0.06920571675,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.458357572555542, timestamp=1662973598.4397767) [('tile_f', [-1, 8, 4, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2140058
-No: 17 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+No: 15 GFLOPS: 5.27/259.35 result: MeasureResult(costs=(0.0439689325,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.8793528079986572, timestamp=1663013709.848698) [('tile_f', [-1, 2, 2, 8]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5330964
+No: 16 GFLOPS: 3.34/259.35 result: MeasureResult(costs=(0.06932482525,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.540909290313721, timestamp=1663013711.0937493) [('tile_f', [-1, 8, 4, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2140058
+No: 17 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 142, in build
res = future.result()
File "/usr/lib/python3.7/concurrent/futures/_base.py", line 435, in result
@@ -1950,8 +1950,8 @@ No: 17 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
TimeoutError
[('tile_f', [-1, 2, 2, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 16]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10195251
-No: 18 GFLOPS: 26.14/260.87 result: MeasureResult(costs=(0.008855842416666667,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.1080987453460693, timestamp=1662973609.323898) [('tile_f', [-1, 4, 8, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6068603
-No: 19 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+No: 18 GFLOPS: 28.28/259.35 result: MeasureResult(costs=(0.008186122,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.3027231693267822, timestamp=1663013722.1563756) [('tile_f', [-1, 4, 8, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6068603
+No: 19 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -2074,7 +2074,7 @@ Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 871, in verify_pass
raise InstantiationError("Skipped because of invalid gpu kernel")
tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 16, 4, 8]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6956993
-No: 20 GFLOPS: 0.00/260.87 result: Traceback (most recent call last):
+No: 20 GFLOPS: 0.00/259.35 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 588, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 540, in _build_func_common
@@ -2237,7 +2237,7 @@ and measure running time.</p>
Best config:
[('tile_f', [-1, 8, 2, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4264713
Finish loading 20 records
-Time cost of this operator: 0.001305
+Time cost of this operator: 0.001259
</pre></div>
</div>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-tune-with-autotvm-tune-conv2d-cuda-py">
diff --git a/docs/how_to/work_with_microtvm/micro_autotune.html b/docs/how_to/work_with_microtvm/micro_autotune.html
index 31a029af0..ac845f9c6 100644
--- a/docs/how_to/work_with_microtvm/micro_autotune.html
+++ b/docs/how_to/work_with_microtvm/micro_autotune.html
@@ -584,10 +584,10 @@ the tuned operator.</p>
########## Build without Autotuning ##########
Node Name Ops Time(us) Time(%) Shape Inputs Outputs Measurements(us)
--------- --- -------- ------- ----- ------ ------- ----------------
-tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 305.2 98.66 (1, 2, 10, 10, 3) 2 1 [305.2]
-tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.215 1.039 (1, 6, 10, 10) 1 1 [3.215]
-tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.932 0.301 (1, 1, 10, 10, 3) 1 1 [0.932]
-Total_time - 309.347 - - - - -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 311.4 98.634 (1, 2, 10, 10, 3) 2 1 [311.4]
+tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.171 1.005 (1, 6, 10, 10) 1 1 [3.171]
+tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 1.142 0.362 (1, 1, 10, 10, 3) 1 1 [1.142]
+Total_time - 315.713 - - - - -
</pre></div>
</div>
</div>
@@ -640,10 +640,10 @@ Total_time -
########## Build with Autotuning ##########
Node Name Ops Time(us) Time(%) Shape Inputs Outputs Measurements(us)
--------- --- -------- ------- ----- ------ ------- ----------------
-tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 128.7 97.884 (1, 6, 10, 10, 1) 2 1 [128.7]
-tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.806 1.373 (1, 6, 10, 10) 1 1 [1.806]
-tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.976 0.743 (1, 1, 10, 10, 3) 1 1 [0.976]
-Total_time - 131.482 - - - - -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 218.0 98.617 (1, 1, 10, 10, 6) 2 1 [218.0]
+tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 2.208 0.999 (1, 6, 10, 10) 1 1 [2.208]
+tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.849 0.384 (1, 3, 10, 10, 1) 1 1 [0.849]
+Total_time - 221.057 - - - - -
</pre></div>
</div>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-work-with-microtvm-micro-autotune-py">
diff --git a/docs/how_to/work_with_microtvm/micro_train.html b/docs/how_to/work_with_microtvm/micro_train.html
index f76a726b5..f2f03291e 100644
--- a/docs/how_to/work_with_microtvm/micro_train.html
+++ b/docs/how_to/work_with_microtvm/micro_train.html
@@ -516,7 +516,7 @@ take about <strong>2 minutes</strong> to download the Stanford Cars, while COCO
<a href="https://docs.python.org/3/library/shutil.html#shutil.move" title="shutil.move" class="sphx-glr-backref-module-shutil sphx-glr-backref-type-py-function"><span class="n">shutil</span><span class="o">.</span><span class="n">move</span></a><span class="p">(</span><span class="sa">f</span><span class="s2">"</span><span class="si">{</span><a href="https://docs.python.org/3/library/stdtypes.html#str" title="builtins.str" class="sphx-glr-backref-module-builtins sphx-glr-backref-typ [...]
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>'/tmp/tmpkzbcrnpu/images/random'
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>'/tmp/tmp75inl8ex/images/random'
</pre></div>
</div>
</div>
@@ -576,8 +576,8 @@ objects to other stuff? We can display some examples from our datasets using <co
<span class="n">plt</span><span class="o">.</span><span class="n">axis</span><span class="p">(</span><span class="s2">"off"</span><span class="p">)</span>
</pre></div>
</div>
-<img src="../../_images/sphx_glr_micro_train_001.png" srcset="../../_images/sphx_glr_micro_train_001.png" alt="[1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0]" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>/tmp/tmpkzbcrnpu/images/target contains 8144 images
-/tmp/tmpkzbcrnpu/images/random contains 5000 images
+<img src="../../_images/sphx_glr_micro_train_001.png" srcset="../../_images/sphx_glr_micro_train_001.png" alt="[1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0]" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>/tmp/tmp75inl8ex/images/target contains 8144 images
+/tmp/tmp75inl8ex/images/random contains 5000 images
</pre></div>
</div>
</div>
@@ -689,13 +689,13 @@ the time on our validation set).</p>
</pre></div>
</div>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Epoch 1/3
-328/328 - 46s - loss: 0.2312 - accuracy: 0.9217 - val_loss: 0.1355 - val_accuracy: 0.9562 - 46s/epoch - 139ms/step
+328/328 - 47s - loss: 0.2088 - accuracy: 0.9299 - val_loss: 0.1318 - val_accuracy: 0.9558 - 47s/epoch - 142ms/step
Epoch 2/3
-328/328 - 42s - loss: 0.0931 - accuracy: 0.9651 - val_loss: 0.1191 - val_accuracy: 0.9622 - 42s/epoch - 129ms/step
+328/328 - 43s - loss: 0.1011 - accuracy: 0.9610 - val_loss: 0.1258 - val_accuracy: 0.9630 - 43s/epoch - 132ms/step
Epoch 3/3
-328/328 - 42s - loss: 0.0649 - accuracy: 0.9750 - val_loss: 0.2244 - val_accuracy: 0.9316 - 42s/epoch - 129ms/step
+328/328 - 43s - loss: 0.0675 - accuracy: 0.9744 - val_loss: 0.1085 - val_accuracy: 0.9630 - 43s/epoch - 131ms/step
-<keras.callbacks.History object at 0x7f6154ba6390>
+<keras.callbacks.History object at 0x7fda749133d0>
</pre></div>
</div>
</div>
@@ -961,7 +961,7 @@ as intended.</p>
<p>From here, we could modify the model to read live images from the camera - we have another
Arduino tutorial for how to do that <a class="reference external" href="https://github.com/guberti/tvm-arduino-demos/tree/master/examples/person_detection">on GitHub</a>. Alternatively, we could also
<a class="reference external" href="https://tvm.apache.org/docs/how_to/work_with_microtvm/micro_autotune.html">use TVM’s autotuning capabilities</a> to dramatically improve the model’s performance.</p>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 4 minutes 36.292 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 4 minutes 48.525 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-work-with-microtvm-micro-train-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/b52cec46baf4f78d6bcd94cbe269c8a6/micro_train.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">micro_train.py</span></code></a></p>
diff --git a/docs/how_to/work_with_microtvm/sg_execution_times.html b/docs/how_to/work_with_microtvm/sg_execution_times.html
index ef965d641..cb652253e 100644
--- a/docs/how_to/work_with_microtvm/sg_execution_times.html
+++ b/docs/how_to/work_with_microtvm/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-work-with-microtvm-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>05:27.627</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
+<p><strong>05:41.583</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -336,19 +336,19 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="micro_train.html#sphx-glr-how-to-work-with-microtvm-micro-train-py"><span class="std std-ref">Training Vision Models for microTVM on Arduino</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_train.py</span></code>)</p></td>
-<td><p>04:36.292</p></td>
+<td><p>04:48.525</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="micro_autotune.html#sphx-glr-how-to-work-with-microtvm-micro-autotune-py"><span class="std std-ref">Autotuning with microTVM</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_autotune.py</span></code>)</p></td>
-<td><p>00:40.411</p></td>
+<td><p>00:41.646</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="micro_aot.html#sphx-glr-how-to-work-with-microtvm-micro-aot-py"><span class="std std-ref">microTVM Host-Driven AoT</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_aot.py</span></code>)</p></td>
-<td><p>00:07.799</p></td>
+<td><p>00:08.116</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="micro_tflite.html#sphx-glr-how-to-work-with-microtvm-micro-tflite-py"><span class="std std-ref">microTVM with TFLite Models</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_tflite.py</span></code>)</p></td>
-<td><p>00:03.124</p></td>
+<td><p>00:03.295</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="micro_ethosu.html#sphx-glr-how-to-work-with-microtvm-micro-ethosu-py"><span class="std std-ref">Running TVM on bare metal Arm(R) Cortex(R)-M55 CPU and Ethos(TM)-U55 NPU with CMSIS-NN</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_ethosu.py</span></code>)</p></td>
diff --git a/docs/how_to/work_with_relay/sg_execution_times.html b/docs/how_to/work_with_relay/sg_execution_times.html
index 4eb10e00f..615062327 100644
--- a/docs/how_to/work_with_relay/sg_execution_times.html
+++ b/docs/how_to/work_with_relay/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-work-with-relay-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:38.996</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
+<p><strong>00:42.559</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -336,15 +336,15 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="using_pipeline_executor.html#sphx-glr-how-to-work-with-relay-using-pipeline-executor-py"><span class="std std-ref">Using Pipeline Executor in Relay</span></a> (<code class="docutils literal notranslate"><span class="pre">using_pipeline_executor.py</span></code>)</p></td>
-<td><p>00:30.095</p></td>
+<td><p>00:31.307</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="using_external_lib.html#sphx-glr-how-to-work-with-relay-using-external-lib-py"><span class="std std-ref">Using External Libraries in Relay</span></a> (<code class="docutils literal notranslate"><span class="pre">using_external_lib.py</span></code>)</p></td>
-<td><p>00:07.682</p></td>
+<td><p>00:09.811</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="build_gcn.html#sphx-glr-how-to-work-with-relay-build-gcn-py"><span class="std std-ref">Building a Graph Convolutional Network</span></a> (<code class="docutils literal notranslate"><span class="pre">build_gcn.py</span></code>)</p></td>
-<td><p>00:01.212</p></td>
+<td><p>00:01.435</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="using_relay_viz.html#sphx-glr-how-to-work-with-relay-using-relay-viz-py"><span class="std std-ref">Use Relay Visualizer to Visualize Relay</span></a> (<code class="docutils literal notranslate"><span class="pre">using_relay_viz.py</span></code>)</p></td>
diff --git a/docs/how_to/work_with_schedules/intrin_math.html b/docs/how_to/work_with_schedules/intrin_math.html
index ed53c29ca..561083f31 100644
--- a/docs/how_to/work_with_schedules/intrin_math.html
+++ b/docs/how_to/work_with_schedules/intrin_math.html
@@ -522,7 +522,7 @@ The following example customizes CUDA lowering rule for <code class="code docuti
<a href="../../reference/api/python/ir.html#tvm.ir.register_intrin_lowering" title="tvm.ir.register_intrin_lowering" class="sphx-glr-backref-module-tvm-ir sphx-glr-backref-type-py-function"><span class="n">register_intrin_lowering</span></a><span class="p">(</span><span class="s2">"tir.exp"</span><span class="p">,</span> <span class="n">target</span><span class="o">=</span><span class="s2">"cuda"</span><span class="p">,</span> <span class="n">f</span><span class="o">= [...]
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span><function my_cuda_math_rule at 0x7f60f3ae8950>
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span><function my_cuda_math_rule at 0x7fd9f52db710>
</pre></div>
</div>
<p>Register the rule to TVM with override option to override existing rule.
diff --git a/docs/how_to/work_with_schedules/sg_execution_times.html b/docs/how_to/work_with_schedules/sg_execution_times.html
index c6c0c1aa9..466b31194 100644
--- a/docs/how_to/work_with_schedules/sg_execution_times.html
+++ b/docs/how_to/work_with_schedules/sg_execution_times.html
@@ -327,7 +327,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-work-with-schedules-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:04.082</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
+<p><strong>00:07.923</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -336,27 +336,27 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="intrin_math.html#sphx-glr-how-to-work-with-schedules-intrin-math-py"><span class="std std-ref">Intrinsics and Math Functions</span></a> (<code class="docutils literal notranslate"><span class="pre">intrin_math.py</span></code>)</p></td>
-<td><p>00:02.023</p></td>
+<td><p>00:05.711</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="tensorize.html#sphx-glr-how-to-work-with-schedules-tensorize-py"><span class="std std-ref">Use Tensorize to Leverage Hardware Intrinsics</span></a> (<code class="docutils literal notranslate"><span class="pre">tensorize.py</span></code>)</p></td>
-<td><p>00:00.897</p></td>
+<td><p>00:00.963</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="reduction.html#sphx-glr-how-to-work-with-schedules-reduction-py"><span class="std std-ref">Reduction</span></a> (<code class="docutils literal notranslate"><span class="pre">reduction.py</span></code>)</p></td>
-<td><p>00:00.496</p></td>
+<td><p>00:00.542</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="scan.html#sphx-glr-how-to-work-with-schedules-scan-py"><span class="std std-ref">Scan and Recurrent Kernel</span></a> (<code class="docutils literal notranslate"><span class="pre">scan.py</span></code>)</p></td>
-<td><p>00:00.489</p></td>
+<td><p>00:00.527</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="extern_op.html#sphx-glr-how-to-work-with-schedules-extern-op-py"><span class="std std-ref">External Tensor Functions</span></a> (<code class="docutils literal notranslate"><span class="pre">extern_op.py</span></code>)</p></td>
-<td><p>00:00.096</p></td>
+<td><p>00:00.097</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="schedule_primitives.html#sphx-glr-how-to-work-with-schedules-schedule-primitives-py"><span class="std std-ref">Schedule Primitives in TVM</span></a> (<code class="docutils literal notranslate"><span class="pre">schedule_primitives.py</span></code>)</p></td>
-<td><p>00:00.041</p></td>
+<td><p>00:00.042</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="tedd.html#sphx-glr-how-to-work-with-schedules-tedd-py"><span class="std std-ref">Use Tensor Expression Debug Display (TEDD) for Visualization</span></a> (<code class="docutils literal notranslate"><span class="pre">tedd.py</span></code>)</p></td>
diff --git a/docs/how_to/work_with_schedules/tensorize.html b/docs/how_to/work_with_schedules/tensorize.html
index f8257d7cf..3618fc3e6 100644
--- a/docs/how_to/work_with_schedules/tensorize.html
+++ b/docs/how_to/work_with_schedules/tensorize.html
@@ -577,7 +577,7 @@ The importing needs to happen before the tensorized GEMV being executed.</p>
C: Buffer(C_2: Pointer(float32), float32, [524288], [])}
buffer_map = {A_1: A, B_1: B, C_1: C}
preflattened_buffer_map = {A_1: A_3: Buffer(A_2, float32, [1024, 64], []), B_1: B_3: Buffer(B_2, float32, [512, 64], []), C_1: C_3: Buffer(C_2, float32, [1024, 512], [])} {
- attr [IterVar(i: int32, (nullptr), "DataPar", "")] "pragma_import_llvm" = "; ModuleID = '/tmp/tmphvuuz9sb/input0.cc'\nsource_filename = \"/tmp/tmphvuuz9sb/input0.cc\"\ntarget datalayout = \"e-m:e-i64:64-f80:128-n8:16:32:64-S128\"\ntarget triple = \"x86_64-pc-linux-gnu\"\n\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local i32 @gemv_update(float*, float*, float*, i32, i32, i32) #0 {\n %7 = allo [...]
+ attr [IterVar(i: int32, (nullptr), "DataPar", "")] "pragma_import_llvm" = "; ModuleID = '/tmp/tmpozmw_jum/input0.cc'\nsource_filename = \"/tmp/tmpozmw_jum/input0.cc\"\ntarget datalayout = \"e-m:e-i64:64-f80:128-n8:16:32:64-S128\"\ntarget triple = \"x86_64-pc-linux-gnu\"\n\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local i32 @gemv_update(float*, float*, float*, i32, i32, i32) #0 {\n %7 = allo [...]
for (i, 0, 1024) {
for (j.outer: int32, 0, 32) {
@tir.call_extern("gemv_update", @tir.tvm_access_ptr(@tir.type_annotation(, dtype=float32), C_2, ((i*512) + (j.outer*16)), 16, 2, dtype=handle), @tir.tvm_access_ptr(@tir.type_annotation(, dtype=float32), A_2, (i*64), 64, 1, dtype=handle), @tir.tvm_access_ptr(@tir.type_annotation(, dtype=float32), B_2, (j.outer*1024), 1024, 1, dtype=handle), 16, 64, 64, dtype=int32)
diff --git a/docs/install/nnpack.html b/docs/install/nnpack.html
index 3153785d7..aa2238b85 100644
--- a/docs/install/nnpack.html
+++ b/docs/install/nnpack.html
@@ -224,7 +224,17 @@
<p class="caption" role="heading"><span class="caption-text">Getting Started</span></p>
<ul class="current">
<li class="toctree-l1 current"><a class="reference internal" href="index.html">Installing TVM</a><ul class="current">
-<li class="toctree-l2"><a class="reference internal" href="from_source.html">Install from Source</a></li>
+<li class="toctree-l2 current"><a class="reference internal" href="from_source.html">Install from Source</a><ul class="current">
+<li class="toctree-l3"><a class="reference internal" href="from_source.html#developers-get-source-from-github">Developers: Get Source from Github</a></li>
+<li class="toctree-l3"><a class="reference internal" href="from_source.html#build-the-shared-library">Build the Shared Library</a></li>
+<li class="toctree-l3"><a class="reference internal" href="from_source.html#python-package-installation">Python Package Installation</a></li>
+<li class="toctree-l3 current"><a class="reference internal" href="from_source.html#install-contrib-libraries">Install Contrib Libraries</a><ul class="current">
+<li class="toctree-l4 current"><a class="current reference internal" href="#">NNPACK Contrib Installation</a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="from_source.html#enable-c-tests">Enable C++ Tests</a></li>
+</ul>
+</li>
<li class="toctree-l2"><a class="reference internal" href="docker.html">Docker Images</a></li>
<li class="toctree-l2 current"><a class="current reference internal" href="#">NNPACK Contrib Installation</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#conditions">Conditions</a></li>
diff --git a/docs/reference/api/doxygen/annotated.html b/docs/reference/api/doxygen/annotated.html
index ad7cdfe03..343a733d5 100644
--- a/docs/reference/api/doxygen/annotated.html
+++ b/docs/reference/api/doxygen/annotated.html
@@ -558,125 +558,127 @@ $(function() {
<tr id="row_1_6_244_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1relay_1_1WildcardPatternNode.html" target="_self">WildcardPatternNode</a></td><td class="desc">Wildcard <a class="el" href="classtvm_1_1relay_1_1Pattern.html" title="Pattern is the base type for an ADT match pattern in Relay. ">Pattern</a> </td></tr>
<tr id="row_1_6_245_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1relay_1_1YoloReorgAttrs.html" target="_self">YoloReorgAttrs</a></td><td class="desc">Attributes used in yolo reorg operators </td></tr>
<tr id="row_1_7_" class="even" style="display:none;"><td class="entry"><span style="width:16px;display:inline-block;"> </span><span id="arr_1_7_" class="arrow" onclick="toggleFolder('1_7_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime.html" target="_self">runtime</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_0_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_0_" class="arrow" onclick="toggleFolder('1_7_0_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1metadata.html" target="_self">metadata</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_0_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayAccessor.html" target="_self">ArrayAccessor</a></td><td class="desc">A span-like class which permits access to <a class="el" href="classtvm_1_1runtime_1_1Array.html" title="Array, container representing a contiguous sequence of ObjectRefs. ">A [...]
-<tr id="row_1_7_0_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayAccessor_3_01const_01char_01_5_00_01_1_1tvm_1_1runtime_1_1String_01_4.html" target="_self">ArrayAccessor< const char *, ::tvm::runtime::String ></a></td><td class="desc">A specialization of <a class="el" href="classtvm_1_1runtime_1_1meta [...]
-<tr id="row_1_7_0_2_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayIterator.html" target="_self">ArrayIterator</a></td><td class="desc">An iterator implementation that lazily instantiates the C++ wrapping <a class="el" href="classtvm_1_1runtime_1_1metadata_1_1Metadata.html">Metadata</a> class </td></tr>
-<tr id="row_1_7_0_3_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ConstantInfoMetadata.html" target="_self">ConstantInfoMetadata</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_0_4_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ConstantInfoMetadataNode.html" target="_self">ConstantInfoMetadataNode</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_0_5_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1Metadata.html" target="_self">Metadata</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_0_6_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArray.html" target="_self">MetadataArray</a></td><td class="desc">Reference class for <a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArray.html" title="Reference class for MetadataArray. ">MetadataArray</a> </td></tr>
-<tr id="row_1_7_0_7_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArrayNode.html" target="_self">MetadataArrayNode</a></td><td class="desc">Container for arrays in the metadata </td></tr>
-<tr id="row_1_7_0_8_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBase.html" target="_self">MetadataBase</a></td><td class="desc">Reference class for the common <a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBaseNode.html" title="Common base class for all Metadata. ">MetadataBaseNode</a> c [...]
-<tr id="row_1_7_0_9_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBaseNode.html" target="_self">MetadataBaseNode</a></td><td class="desc">Common base class for all <a class="el" href="classtvm_1_1runtime_1_1metadata_1_1Metadata.html">Metadata</a> </td></tr>
-<tr id="row_1_7_0_10_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataNode.html" target="_self">MetadataNode</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_0_11_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1TensorInfo.html" target="_self">TensorInfo</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_0_12_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1TensorInfoNode.html" target="_self">TensorInfoNode</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_1_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_1_" class="arrow" onclick="toggleFolder('1_7_1_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1micro__rpc.html" target="_self">micro_rpc</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_1_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1FrameBuffer.html" target="_self">FrameBuffer</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_1_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1Framer.html" target="_self">Framer</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_1_2_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1PacketFieldSizeBytes.html" target="_self">PacketFieldSizeBytes</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_1_3_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1Session.html" target="_self">Session</a></td><td class="desc">CRT communication session management class. Assumes the following properties provided by the underlying transport: </td></tr>
-<tr id="row_1_7_1_4_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1micro__rpc_1_1SessionHeader.html" target="_self">SessionHeader</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_1_5_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1Unframer.html" target="_self">Unframer</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_1_6_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1WriteStream.html" target="_self">WriteStream</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_2_" class="arrow" onclick="toggleFolder('1_7_2_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1profiling.html" target="_self">profiling</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1profiling_1_1CallFrame.html" target="_self">CallFrame</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1CountNode.html" target="_self">CountNode</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_2_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1DeviceWrapper.html" target="_self">DeviceWrapper</a></td><td class="desc">Wrapper for <code>Device</code> </td></tr>
-<tr id="row_1_7_2_3_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1profiling_1_1DeviceWrapperNode.html" target="_self">DeviceWrapperNode</a></td><td class="desc">Wrapper for <code>Device</code> because <code>Device</code> is not passable across the <a class="el" href="classtvm_1_1runtime_1_1PackedFunc.html" title="Packed fun [...]
-<tr id="row_1_7_2_4_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1DurationNode.html" target="_self">DurationNode</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_5_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollector.html" target="_self">MetricCollector</a></td><td class="desc">Wrapper for <code><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollectorNode.html" title="Interface for user defined profiling metric collection. ">Metr [...]
-<tr id="row_1_7_2_6_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollectorNode.html" target="_self">MetricCollectorNode</a></td><td class="desc">Interface for user defined profiling metric collection </td></tr>
-<tr id="row_1_7_2_7_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1PercentNode.html" target="_self">PercentNode</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_8_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1Profiler.html" target="_self">Profiler</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_9_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1RatioNode.html" target="_self">RatioNode</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_10_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1Report.html" target="_self">Report</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_2_11_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1ReportNode.html" target="_self">ReportNode</a></td><td class="desc">Data collected from a profiling run. Includes per-call metrics and per-device metrics </td></tr>
-<tr id="row_1_7_3_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_3_" class="arrow" onclick="toggleFolder('1_7_3_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1threading.html" target="_self">threading</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_3_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1threading_1_1ThreadGroup.html" target="_self">ThreadGroup</a></td><td class="desc">A platform-agnostic abstraction for managing a collection of thread pool threads </td></tr>
-<tr id="row_1_7_4_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_4_" class="arrow" onclick="toggleFolder('1_7_4_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1vm.html" target="_self">vm</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_4_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Allocator.html" target="_self">Allocator</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_4_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1vm_1_1Buffer.html" target="_self">Buffer</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_4_2_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Executable.html" target="_self">Executable</a></td><td class="desc">The executable emitted by the VM compiler </td></tr>
-<tr id="row_1_7_4_3_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1vm_1_1Instruction.html" target="_self">Instruction</a></td><td class="desc">A single virtual machine instruction </td></tr>
-<tr id="row_1_7_4_4_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1MemoryManager.html" target="_self">MemoryManager</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_4_5_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Storage.html" target="_self">Storage</a></td><td class="desc">Reference to storage </td></tr>
-<tr id="row_1_7_4_6_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1StorageObj.html" target="_self">StorageObj</a></td><td class="desc">An object representing a storage allocation </td></tr>
-<tr id="row_1_7_4_7_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1VirtualMachine.html" target="_self">VirtualMachine</a></td><td class="desc">The virtual machine </td></tr>
-<tr id="row_1_7_4_8_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1VMClosure.html" target="_self">VMClosure</a></td><td class="desc">Reference to closure </td></tr>
-<tr id="row_1_7_4_9_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1VMClosureObj.html" target="_self">VMClosureObj</a></td><td class="desc">An object representing a vm closure </td></tr>
-<tr id="row_1_7_4_10_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1vm_1_1VMFrame.html" target="_self">VMFrame</a></td><td class="desc">A representation of a stack frame </td></tr>
-<tr id="row_1_7_4_11_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1vm_1_1VMFunction.html" target="_self">VMFunction</a></td><td class="desc">A representation of a Relay function in the VM </td></tr>
-<tr id="row_1_7_5_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ADT.html" target="_self">ADT</a></td><td class="desc">Reference to algebraic data type objects </td></tr>
-<tr id="row_1_7_6_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ADTObj.html" target="_self">ADTObj</a></td><td class="desc">An object representing a structure or enumeration </td></tr>
-<tr id="row_1_7_7_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_7_" class="arrow" onclick="toggleFolder('1_7_7_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Array.html" target="_self">Array</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Array.html" title="Array, container representing a contiguous sequence of Obje [...]
-<tr id="row_1_7_7_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1Array_1_1ValueConverter.html" target="_self">ValueConverter</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_8_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ArrayNode.html" target="_self">ArrayNode</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Array.html" title="Array, container representing a contiguous sequence of ObjectRefs. ">Array</a> node content in array </td></tr>
-<tr id="row_1_7_9_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Closure.html" target="_self">Closure</a></td><td class="desc">Reference to closure </td></tr>
-<tr id="row_1_7_10_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ClosureObj.html" target="_self">ClosureObj</a></td><td class="desc">An object representing a closure. This object is used by both the Relay VM and interpreter </td></tr>
-<tr id="row_1_7_11_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1DataType.html" target="_self">DataType</a></td><td class="desc">Runtime primitive data type </td></tr>
-<tr id="row_1_7_12_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1DenseMapNode.html" target="_self">DenseMapNode</a></td><td class="desc">A specialization of hash map that implements the idea of array-based hash map. Another reference implementation can be found [1] </td></tr>
-<tr id="row_1_7_13_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1DeviceAPI.html" target="_self">DeviceAPI</a></td><td class="desc">TVM Runtime Device API, abstracts the device specific interface for memory management </td></tr>
-<tr id="row_1_7_14_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1InplaceArrayBase.html" target="_self">InplaceArrayBase</a></td><td class="desc">Base template for classes with array like memory layout </td></tr>
-<tr id="row_1_7_15_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1IterAdapter.html" target="_self">IterAdapter</a></td><td class="desc">Iterator adapter that adapts TIter to return another type </td></tr>
-<tr id="row_1_7_16_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_16_" class="arrow" onclick="toggleFolder('1_7_16_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Map.html" target="_self">Map</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Map.html" title="Map container of NodeRef->NodeRef in DSL graph. Map impleme [...]
-<tr id="row_1_7_16_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Map_1_1iterator.html" target="_self">iterator</a></td><td class="desc">Iterator of the hash map </td></tr>
-<tr id="row_1_7_17_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_17_" class="arrow" onclick="toggleFolder('1_7_17_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1MapNode.html" target="_self">MapNode</a></td><td class="desc">Shared content of all specializations of hash map </td></tr>
-<tr id="row_1_7_17_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1MapNode_1_1iterator.html" target="_self">iterator</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_18_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Module.html" target="_self">Module</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Module.html" title="Module container of TVM. ">Module</a> container of TVM </td></tr>
-<tr id="row_1_7_19_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ModuleNode.html" target="_self">ModuleNode</a></td><td class="desc">Base container of module </td></tr>
-<tr id="row_1_7_20_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_20_" class="arrow" onclick="toggleFolder('1_7_20_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1NDArray.html" target="_self">NDArray</a></td><td class="desc">Managed <a class="el" href="classtvm_1_1runtime_1_1NDArray.html" title="Managed NDArray. The array is backed by re [...]
-<tr id="row_1_7_20_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1NDArray_1_1Container.html" target="_self">Container</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Object.html" title="base class of all object containers. ">Object</a> container class that backs <a class="el" href="classtvm_1_1runtime_1_ [...]
-<tr id="row_1_7_20_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1NDArray_1_1ContainerBase.html" target="_self">ContainerBase</a></td><td class="desc">The container base structure contains all the fields except for the <a class="el" href="classtvm_1_1runtime_1_1Object.html" title="base class of all object containers. ">Obje [...]
-<tr id="row_1_7_21_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1NullOptType.html" target="_self">NullOptType</a></td><td class="desc">Helper to represent nullptr for optional </td></tr>
-<tr id="row_1_7_22_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ObjAllocatorBase.html" target="_self">ObjAllocatorBase</a></td><td class="desc">Base class of object allocators that implements make. Use curiously recurring template pattern </td></tr>
-<tr id="row_1_7_23_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Object.html" target="_self">Object</a></td><td class="desc">Base class of all object containers </td></tr>
-<tr id="row_1_7_24_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectEqual.html" target="_self">ObjectEqual</a></td><td class="desc">String-aware <a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" title="Base class of all object reference. ">ObjectRef</a> hash functor </td></tr>
-<tr id="row_1_7_25_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectHash.html" target="_self">ObjectHash</a></td><td class="desc">String-aware <a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" title="Base class of all object reference. ">ObjectRef</a> equal functor </td></tr>
-<tr id="row_1_7_26_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ObjectPtr.html" target="_self">ObjectPtr</a></td><td class="desc">A custom smart pointer for <a class="el" href="classtvm_1_1runtime_1_1Object.html" title="base class of all object containers. ">Object</a> </td></tr>
-<tr id="row_1_7_27_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectPtrEqual.html" target="_self">ObjectPtrEqual</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" title="Base class of all object reference. ">ObjectRef</a> equal functor </td></tr>
-<tr id="row_1_7_28_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectPtrHash.html" target="_self">ObjectPtrHash</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" title="Base class of all object reference. ">ObjectRef</a> hash functor </td></tr>
-<tr id="row_1_7_29_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" target="_self">ObjectRef</a></td><td class="desc">Base class of all object reference </td></tr>
-<tr id="row_1_7_30_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker.html" target="_self">ObjectTypeChecker</a></td><td class="desc"><a class="el" href="classtvm_1_1Type.html" title="Managed reference to TypeNode. ">Type</a> traits for runtime type check during FFI conversion </td></tr>
-<tr id="row_1_7_31_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker_3_01Array_3_01T_01_4_01_4.html" target="_self">ObjectTypeChecker< Array< T > ></a></td><td class="desc"></td></tr>
-<tr id="row_1_7_32_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker_3_01Map_3_01K_00_01V_01_4_01_4.html" target="_self">ObjectTypeChecker< Map< K, V > ></a></td><td class="desc"></td></tr>
-<tr id="row_1_7_33_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Optional.html" target="_self">Optional</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Optional.html" title="Optional container that to represent to a Nullable variant of T. ">Optional</a> container that to represent to a Nullable variant of [...]
-<tr id="row_1_7_34_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1PackedFunc.html" target="_self">PackedFunc</a></td><td class="desc">Packed function is a type-erased function. The arguments are passed by packed format </td></tr>
-<tr id="row_1_7_35_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_35_" class="arrow" onclick="toggleFolder('1_7_35_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1PackedFuncObj.html" target="_self">PackedFuncObj</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Object.html" title="base class of all object containers. "> [...]
-<tr id="row_1_7_35_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncObj_1_1Extractor.html" target="_self">Extractor</a></td><td class="desc">Internal struct for extracting the callable method from callable type </td></tr>
-<tr id="row_1_7_36_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1PackedFuncSubObj.html" target="_self">PackedFuncSubObj</a></td><td class="desc">Derived object class for constructing <a class="el" href="classtvm_1_1runtime_1_1PackedFuncObj.html" title="Object container class that backs PackedFunc. ">PackedFuncObj</a> </td></tr>
-<tr id="row_1_7_37_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter.html" target="_self">PackedFuncValueConverter</a></td><td class="desc"><a class="el" href="classtvm_1_1Type.html" title="Managed reference to TypeNode. ">Type</a> trait to specify special value conversion rules from <a class="el" href= [...]
-<tr id="row_1_7_38_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01Optional_3_01T_01_4_01_4.html" target="_self">PackedFuncValueConverter< Optional< T > ></a></td><td class="desc"></td></tr>
-<tr id="row_1_7_39_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01PrimExpr_01_4.html" target="_self">PackedFuncValueConverter< PrimExpr ></a></td><td class="desc"></td></tr>
-<tr id="row_1_7_40_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01tvm_1_1Bool_01_4.html" target="_self">PackedFuncValueConverter< tvm::Bool ></a></td><td class="desc"></td></tr>
-<tr id="row_1_7_41_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01tvm_1_1Integer_01_4.html" target="_self">PackedFuncValueConverter< tvm::Integer ></a></td><td class="desc"></td></tr>
-<tr id="row_1_7_42_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_1_1tvm_1_1runtime_1_1String_01_4.html" target="_self">PackedFuncValueConverter<::tvm::runtime::String ></a></td><td class="desc"></td></tr>
-<tr id="row_1_7_43_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Registry.html" target="_self">Registry</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Registry.html" title="Registry for global function. ">Registry</a> for global function </td></tr>
-<tr id="row_1_7_44_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ReverseIterAdapter.html" target="_self">ReverseIterAdapter</a></td><td class="desc">Iterator adapter that adapts TIter to return another type </td></tr>
-<tr id="row_1_7_45_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ShapeTuple.html" target="_self">ShapeTuple</a></td><td class="desc">Reference to shape tuple objects </td></tr>
-<tr id="row_1_7_46_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_46_" class="arrow" onclick="toggleFolder('1_7_46_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ShapeTupleObj.html" target="_self">ShapeTupleObj</a></td><td class="desc">An object representing a shape tuple </td></tr>
-<tr id="row_1_7_46_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ShapeTupleObj_1_1FromStd.html" target="_self">FromStd</a></td><td class="desc">An object representing shape tuple moved from std::vector </td></tr>
-<tr id="row_1_7_47_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1SignaturePrinter.html" target="_self">SignaturePrinter</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_48_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_48_" class="arrow" onclick="toggleFolder('1_7_48_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator.html" target="_self">SimpleObjAllocator</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_48_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator_1_1ArrayHandler.html" target="_self">ArrayHandler</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_48_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator_1_1Handler.html" target="_self">Handler</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_49_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1SmallMapNode.html" target="_self">SmallMapNode</a></td><td class="desc">A specialization of small-sized hash map </td></tr>
-<tr id="row_1_7_50_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1String.html" target="_self">String</a></td><td class="desc">Reference to string objects </td></tr>
-<tr id="row_1_7_51_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_51_" class="arrow" onclick="toggleFolder('1_7_51_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1StringObj.html" target="_self">StringObj</a></td><td class="desc">An object representing string. It's POD type </td></tr>
-<tr id="row_1_7_51_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1StringObj_1_1FromStd.html" target="_self">FromStd</a></td><td class="desc">An object representing string moved from std::string </td></tr>
-<tr id="row_1_7_52_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Timer.html" target="_self">Timer</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Timer.html" title="Timer for a specific device. ">Timer</a> for a specific device </td></tr>
-<tr id="row_1_7_53_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TimerNode.html" target="_self">TimerNode</a></td><td class="desc">Base class for all implementations </td></tr>
-<tr id="row_1_7_54_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMArgs.html" target="_self">TVMArgs</a></td><td class="desc">Arguments into TVM functions </td></tr>
-<tr id="row_1_7_55_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMArgsSetter.html" target="_self">TVMArgsSetter</a></td><td class="desc"></td></tr>
-<tr id="row_1_7_56_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMArgValue.html" target="_self">TVMArgValue</a></td><td class="desc">A single argument value to <a class="el" href="classtvm_1_1runtime_1_1PackedFunc.html" title="Packed function is a type-erased function. The arguments are passed by packed format. ">PackedFun [...]
-<tr id="row_1_7_57_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMMovableArgValue__.html" target="_self">TVMMovableArgValue_</a></td><td class="desc">Internal auxiliary struct for <a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc.html" title="Please refer to TypedPackedFunc<R(Args..)>. ">TypedPackedFunc</a> [...]
-<tr id="row_1_7_58_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMMovableArgValueWithContext__.html" target="_self">TVMMovableArgValueWithContext_</a></td><td class="desc">Internal auxiliary struct for <a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc.html" title="Please refer to TypedPackedFunc<R(Args..)>. [...]
-<tr id="row_1_7_59_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMPODValue__.html" target="_self">TVMPODValue_</a></td><td class="desc">Internal base class to handle conversion to POD values </td></tr>
-<tr id="row_1_7_60_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMRetValue.html" target="_self">TVMRetValue</a></td><td class="desc">Return Value container, Unlike <a class="el" href="classtvm_1_1runtime_1_1TVMArgValue.html" title="A single argument value to PackedFunc. Containing both type_code and TVMValue. ">TVMArgValue [...]
-<tr id="row_1_7_61_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc.html" target="_self">TypedPackedFunc</a></td><td class="desc">Please refer to <a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc_3_01R_07Args_8_8_8_08_4.html#TypedPackedFuncAnchor">TypedPackedFunc<R(Args..)></a> </td></tr>
-<tr id="row_1_7_62_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc_3_01R_07Args_8_8_8_08_4.html" target="_self">TypedPackedFunc< R(Args...)></a></td><td class="desc">A <a class="el" href="classtvm_1_1runtime_1_1PackedFunc.html" title="Packed function is a type-erased function. The arguments are passed by [...]
-<tr id="row_1_7_63_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1TypeIndex.html" target="_self">TypeIndex</a></td><td class="desc">Namespace for the list of type index </td></tr>
+<tr id="row_1_7_0_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_0_" class="arrow" onclick="toggleFolder('1_7_0_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1hexagon.html" target="_self">hexagon</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_0_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1hexagon_1_1SDLTensor.html" target="_self">SDLTensor</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_1_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_1_" class="arrow" onclick="toggleFolder('1_7_1_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1metadata.html" target="_self">metadata</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_1_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayAccessor.html" target="_self">ArrayAccessor</a></td><td class="desc">A span-like class which permits access to <a class="el" href="classtvm_1_1runtime_1_1Array.html" title="Array, container representing a contiguous sequence of ObjectRefs. ">A [...]
+<tr id="row_1_7_1_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayAccessor_3_01const_01char_01_5_00_01_1_1tvm_1_1runtime_1_1String_01_4.html" target="_self">ArrayAccessor< const char *, ::tvm::runtime::String ></a></td><td class="desc">A specialization of <a class="el" href="classtvm_1_1runtime_1_1meta [...]
+<tr id="row_1_7_1_2_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayIterator.html" target="_self">ArrayIterator</a></td><td class="desc">An iterator implementation that lazily instantiates the C++ wrapping <a class="el" href="classtvm_1_1runtime_1_1metadata_1_1Metadata.html">Metadata</a> class </td></tr>
+<tr id="row_1_7_1_3_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ConstantInfoMetadata.html" target="_self">ConstantInfoMetadata</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_1_4_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ConstantInfoMetadataNode.html" target="_self">ConstantInfoMetadataNode</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_1_5_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1Metadata.html" target="_self">Metadata</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_1_6_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArray.html" target="_self">MetadataArray</a></td><td class="desc">Reference class for <a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArray.html" title="Reference class for MetadataArray. ">MetadataArray</a> </td></tr>
+<tr id="row_1_7_1_7_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArrayNode.html" target="_self">MetadataArrayNode</a></td><td class="desc">Container for arrays in the metadata </td></tr>
+<tr id="row_1_7_1_8_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBase.html" target="_self">MetadataBase</a></td><td class="desc">Reference class for the common <a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBaseNode.html" title="Common base class for all Metadata. ">MetadataBaseNode</a> c [...]
+<tr id="row_1_7_1_9_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBaseNode.html" target="_self">MetadataBaseNode</a></td><td class="desc">Common base class for all <a class="el" href="classtvm_1_1runtime_1_1metadata_1_1Metadata.html">Metadata</a> </td></tr>
+<tr id="row_1_7_1_10_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataNode.html" target="_self">MetadataNode</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_1_11_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1TensorInfo.html" target="_self">TensorInfo</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_1_12_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1TensorInfoNode.html" target="_self">TensorInfoNode</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_2_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_2_" class="arrow" onclick="toggleFolder('1_7_2_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1micro__rpc.html" target="_self">micro_rpc</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_2_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1FrameBuffer.html" target="_self">FrameBuffer</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_2_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1Framer.html" target="_self">Framer</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_2_2_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1PacketFieldSizeBytes.html" target="_self">PacketFieldSizeBytes</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_2_3_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1Session.html" target="_self">Session</a></td><td class="desc">CRT communication session management class. Assumes the following properties provided by the underlying transport: </td></tr>
+<tr id="row_1_7_2_4_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1micro__rpc_1_1SessionHeader.html" target="_self">SessionHeader</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_2_5_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1Unframer.html" target="_self">Unframer</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_2_6_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1WriteStream.html" target="_self">WriteStream</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_3_" class="arrow" onclick="toggleFolder('1_7_3_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1profiling.html" target="_self">profiling</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1profiling_1_1CallFrame.html" target="_self">CallFrame</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1CountNode.html" target="_self">CountNode</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_2_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1DeviceWrapper.html" target="_self">DeviceWrapper</a></td><td class="desc">Wrapper for <code>Device</code> </td></tr>
+<tr id="row_1_7_3_3_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1profiling_1_1DeviceWrapperNode.html" target="_self">DeviceWrapperNode</a></td><td class="desc">Wrapper for <code>Device</code> because <code>Device</code> is not passable across the <a class="el" href="classtvm_1_1runtime_1_1PackedFunc.html" title="Packed fun [...]
+<tr id="row_1_7_3_4_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1DurationNode.html" target="_self">DurationNode</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_5_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollector.html" target="_self">MetricCollector</a></td><td class="desc">Wrapper for <code><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollectorNode.html" title="Interface for user defined profiling metric collection. ">Metr [...]
+<tr id="row_1_7_3_6_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollectorNode.html" target="_self">MetricCollectorNode</a></td><td class="desc">Interface for user defined profiling metric collection </td></tr>
+<tr id="row_1_7_3_7_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1PercentNode.html" target="_self">PercentNode</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_8_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1Profiler.html" target="_self">Profiler</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_9_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1RatioNode.html" target="_self">RatioNode</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_10_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1Report.html" target="_self">Report</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_3_11_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1ReportNode.html" target="_self">ReportNode</a></td><td class="desc">Data collected from a profiling run. Includes per-call metrics and per-device metrics </td></tr>
+<tr id="row_1_7_4_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_4_" class="arrow" onclick="toggleFolder('1_7_4_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1threading.html" target="_self">threading</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_4_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1threading_1_1ThreadGroup.html" target="_self">ThreadGroup</a></td><td class="desc">A platform-agnostic abstraction for managing a collection of thread pool threads </td></tr>
+<tr id="row_1_7_5_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_5_" class="arrow" onclick="toggleFolder('1_7_5_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1runtime_1_1vm.html" target="_self">vm</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_5_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Allocator.html" target="_self">Allocator</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_5_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1vm_1_1Buffer.html" target="_self">Buffer</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_5_2_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Executable.html" target="_self">Executable</a></td><td class="desc">The executable emitted by the VM compiler </td></tr>
+<tr id="row_1_7_5_3_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1vm_1_1Instruction.html" target="_self">Instruction</a></td><td class="desc">A single virtual machine instruction </td></tr>
+<tr id="row_1_7_5_4_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1MemoryManager.html" target="_self">MemoryManager</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_5_5_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Storage.html" target="_self">Storage</a></td><td class="desc">Reference to storage </td></tr>
+<tr id="row_1_7_5_6_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1StorageObj.html" target="_self">StorageObj</a></td><td class="desc">An object representing a storage allocation </td></tr>
+<tr id="row_1_7_5_7_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1VirtualMachine.html" target="_self">VirtualMachine</a></td><td class="desc">The virtual machine </td></tr>
+<tr id="row_1_7_5_8_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1VMClosure.html" target="_self">VMClosure</a></td><td class="desc">Reference to closure </td></tr>
+<tr id="row_1_7_5_9_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1vm_1_1VMClosureObj.html" target="_self">VMClosureObj</a></td><td class="desc">An object representing a vm closure </td></tr>
+<tr id="row_1_7_5_10_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1vm_1_1VMFrame.html" target="_self">VMFrame</a></td><td class="desc">A representation of a stack frame </td></tr>
+<tr id="row_1_7_5_11_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1vm_1_1VMFunction.html" target="_self">VMFunction</a></td><td class="desc">A representation of a Relay function in the VM </td></tr>
+<tr id="row_1_7_6_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ADT.html" target="_self">ADT</a></td><td class="desc">Reference to algebraic data type objects </td></tr>
+<tr id="row_1_7_7_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ADTObj.html" target="_self">ADTObj</a></td><td class="desc">An object representing a structure or enumeration </td></tr>
+<tr id="row_1_7_8_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_8_" class="arrow" onclick="toggleFolder('1_7_8_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Array.html" target="_self">Array</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Array.html" title="Array, container representing a contiguous sequence of Obje [...]
+<tr id="row_1_7_8_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1Array_1_1ValueConverter.html" target="_self">ValueConverter</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_9_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ArrayNode.html" target="_self">ArrayNode</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Array.html" title="Array, container representing a contiguous sequence of ObjectRefs. ">Array</a> node content in array </td></tr>
+<tr id="row_1_7_10_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Closure.html" target="_self">Closure</a></td><td class="desc">Reference to closure </td></tr>
+<tr id="row_1_7_11_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ClosureObj.html" target="_self">ClosureObj</a></td><td class="desc">An object representing a closure. This object is used by both the Relay VM and interpreter </td></tr>
+<tr id="row_1_7_12_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1DataType.html" target="_self">DataType</a></td><td class="desc">Runtime primitive data type </td></tr>
+<tr id="row_1_7_13_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1DenseMapNode.html" target="_self">DenseMapNode</a></td><td class="desc">A specialization of hash map that implements the idea of array-based hash map. Another reference implementation can be found [1] </td></tr>
+<tr id="row_1_7_14_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1DeviceAPI.html" target="_self">DeviceAPI</a></td><td class="desc">TVM Runtime Device API, abstracts the device specific interface for memory management </td></tr>
+<tr id="row_1_7_15_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1InplaceArrayBase.html" target="_self">InplaceArrayBase</a></td><td class="desc">Base template for classes with array like memory layout </td></tr>
+<tr id="row_1_7_16_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1IterAdapter.html" target="_self">IterAdapter</a></td><td class="desc">Iterator adapter that adapts TIter to return another type </td></tr>
+<tr id="row_1_7_17_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_17_" class="arrow" onclick="toggleFolder('1_7_17_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Map.html" target="_self">Map</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Map.html" title="Map container of NodeRef->NodeRef in DSL graph. Map impleme [...]
+<tr id="row_1_7_17_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Map_1_1iterator.html" target="_self">iterator</a></td><td class="desc">Iterator of the hash map </td></tr>
+<tr id="row_1_7_18_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_18_" class="arrow" onclick="toggleFolder('1_7_18_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1MapNode.html" target="_self">MapNode</a></td><td class="desc">Shared content of all specializations of hash map </td></tr>
+<tr id="row_1_7_18_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1MapNode_1_1iterator.html" target="_self">iterator</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_19_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Module.html" target="_self">Module</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Module.html" title="Module container of TVM. ">Module</a> container of TVM </td></tr>
+<tr id="row_1_7_20_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ModuleNode.html" target="_self">ModuleNode</a></td><td class="desc">Base container of module </td></tr>
+<tr id="row_1_7_21_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_21_" class="arrow" onclick="toggleFolder('1_7_21_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1NDArray.html" target="_self">NDArray</a></td><td class="desc">Managed <a class="el" href="classtvm_1_1runtime_1_1NDArray.html" title="Managed NDArray. The array is backed by re [...]
+<tr id="row_1_7_21_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1NDArray_1_1Container.html" target="_self">Container</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Object.html" title="base class of all object containers. ">Object</a> container class that backs <a class="el" href="classtvm_1_1runtime_1_ [...]
+<tr id="row_1_7_21_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1NDArray_1_1ContainerBase.html" target="_self">ContainerBase</a></td><td class="desc">The container base structure contains all the fields except for the <a class="el" href="classtvm_1_1runtime_1_1Object.html" title="base class of all object containers. ">Obje [...]
+<tr id="row_1_7_22_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1NullOptType.html" target="_self">NullOptType</a></td><td class="desc">Helper to represent nullptr for optional </td></tr>
+<tr id="row_1_7_23_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ObjAllocatorBase.html" target="_self">ObjAllocatorBase</a></td><td class="desc">Base class of object allocators that implements make. Use curiously recurring template pattern </td></tr>
+<tr id="row_1_7_24_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Object.html" target="_self">Object</a></td><td class="desc">Base class of all object containers </td></tr>
+<tr id="row_1_7_25_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectEqual.html" target="_self">ObjectEqual</a></td><td class="desc">String-aware <a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" title="Base class of all object reference. ">ObjectRef</a> hash functor </td></tr>
+<tr id="row_1_7_26_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectHash.html" target="_self">ObjectHash</a></td><td class="desc">String-aware <a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" title="Base class of all object reference. ">ObjectRef</a> equal functor </td></tr>
+<tr id="row_1_7_27_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ObjectPtr.html" target="_self">ObjectPtr</a></td><td class="desc">A custom smart pointer for <a class="el" href="classtvm_1_1runtime_1_1Object.html" title="base class of all object containers. ">Object</a> </td></tr>
+<tr id="row_1_7_28_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectPtrEqual.html" target="_self">ObjectPtrEqual</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" title="Base class of all object reference. ">ObjectRef</a> equal functor </td></tr>
+<tr id="row_1_7_29_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectPtrHash.html" target="_self">ObjectPtrHash</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" title="Base class of all object reference. ">ObjectRef</a> hash functor </td></tr>
+<tr id="row_1_7_30_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html" target="_self">ObjectRef</a></td><td class="desc">Base class of all object reference </td></tr>
+<tr id="row_1_7_31_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker.html" target="_self">ObjectTypeChecker</a></td><td class="desc"><a class="el" href="classtvm_1_1Type.html" title="Managed reference to TypeNode. ">Type</a> traits for runtime type check during FFI conversion </td></tr>
+<tr id="row_1_7_32_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker_3_01Array_3_01T_01_4_01_4.html" target="_self">ObjectTypeChecker< Array< T > ></a></td><td class="desc"></td></tr>
+<tr id="row_1_7_33_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker_3_01Map_3_01K_00_01V_01_4_01_4.html" target="_self">ObjectTypeChecker< Map< K, V > ></a></td><td class="desc"></td></tr>
+<tr id="row_1_7_34_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Optional.html" target="_self">Optional</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Optional.html" title="Optional container that to represent to a Nullable variant of T. ">Optional</a> container that to represent to a Nullable variant of [...]
+<tr id="row_1_7_35_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1PackedFunc.html" target="_self">PackedFunc</a></td><td class="desc">Packed function is a type-erased function. The arguments are passed by packed format </td></tr>
+<tr id="row_1_7_36_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_36_" class="arrow" onclick="toggleFolder('1_7_36_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1PackedFuncObj.html" target="_self">PackedFuncObj</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Object.html" title="base class of all object containers. "> [...]
+<tr id="row_1_7_36_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncObj_1_1Extractor.html" target="_self">Extractor</a></td><td class="desc">Internal struct for extracting the callable method from callable type </td></tr>
+<tr id="row_1_7_37_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1PackedFuncSubObj.html" target="_self">PackedFuncSubObj</a></td><td class="desc">Derived object class for constructing <a class="el" href="classtvm_1_1runtime_1_1PackedFuncObj.html" title="Object container class that backs PackedFunc. ">PackedFuncObj</a> </td></tr>
+<tr id="row_1_7_38_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter.html" target="_self">PackedFuncValueConverter</a></td><td class="desc"><a class="el" href="classtvm_1_1Type.html" title="Managed reference to TypeNode. ">Type</a> trait to specify special value conversion rules from <a class="el" href= [...]
+<tr id="row_1_7_39_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01Optional_3_01T_01_4_01_4.html" target="_self">PackedFuncValueConverter< Optional< T > ></a></td><td class="desc"></td></tr>
+<tr id="row_1_7_40_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01PrimExpr_01_4.html" target="_self">PackedFuncValueConverter< PrimExpr ></a></td><td class="desc"></td></tr>
+<tr id="row_1_7_41_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01tvm_1_1Bool_01_4.html" target="_self">PackedFuncValueConverter< tvm::Bool ></a></td><td class="desc"></td></tr>
+<tr id="row_1_7_42_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01tvm_1_1Integer_01_4.html" target="_self">PackedFuncValueConverter< tvm::Integer ></a></td><td class="desc"></td></tr>
+<tr id="row_1_7_43_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_1_1tvm_1_1runtime_1_1String_01_4.html" target="_self">PackedFuncValueConverter<::tvm::runtime::String ></a></td><td class="desc"></td></tr>
+<tr id="row_1_7_44_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Registry.html" target="_self">Registry</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Registry.html" title="Registry for global function. ">Registry</a> for global function </td></tr>
+<tr id="row_1_7_45_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ReverseIterAdapter.html" target="_self">ReverseIterAdapter</a></td><td class="desc">Iterator adapter that adapts TIter to return another type </td></tr>
+<tr id="row_1_7_46_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ShapeTuple.html" target="_self">ShapeTuple</a></td><td class="desc">Reference to shape tuple objects </td></tr>
+<tr id="row_1_7_47_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_47_" class="arrow" onclick="toggleFolder('1_7_47_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ShapeTupleObj.html" target="_self">ShapeTupleObj</a></td><td class="desc">An object representing a shape tuple </td></tr>
+<tr id="row_1_7_47_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1ShapeTupleObj_1_1FromStd.html" target="_self">FromStd</a></td><td class="desc">An object representing shape tuple moved from std::vector </td></tr>
+<tr id="row_1_7_48_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1SignaturePrinter.html" target="_self">SignaturePrinter</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_49_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_49_" class="arrow" onclick="toggleFolder('1_7_49_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator.html" target="_self">SimpleObjAllocator</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_49_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator_1_1ArrayHandler.html" target="_self">ArrayHandler</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_49_1_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator_1_1Handler.html" target="_self">Handler</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_50_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1SmallMapNode.html" target="_self">SmallMapNode</a></td><td class="desc">A specialization of small-sized hash map </td></tr>
+<tr id="row_1_7_51_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1String.html" target="_self">String</a></td><td class="desc">Reference to string objects </td></tr>
+<tr id="row_1_7_52_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_7_52_" class="arrow" onclick="toggleFolder('1_7_52_')">►</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1StringObj.html" target="_self">StringObj</a></td><td class="desc">An object representing string. It's POD type </td></tr>
+<tr id="row_1_7_52_0_" class="even" style="display:none;"><td class="entry"><span style="width:64px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1StringObj_1_1FromStd.html" target="_self">FromStd</a></td><td class="desc">An object representing string moved from std::string </td></tr>
+<tr id="row_1_7_53_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1Timer.html" target="_self">Timer</a></td><td class="desc"><a class="el" href="classtvm_1_1runtime_1_1Timer.html" title="Timer for a specific device. ">Timer</a> for a specific device </td></tr>
+<tr id="row_1_7_54_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TimerNode.html" target="_self">TimerNode</a></td><td class="desc">Base class for all implementations </td></tr>
+<tr id="row_1_7_55_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMArgs.html" target="_self">TVMArgs</a></td><td class="desc">Arguments into TVM functions </td></tr>
+<tr id="row_1_7_56_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMArgsSetter.html" target="_self">TVMArgsSetter</a></td><td class="desc"></td></tr>
+<tr id="row_1_7_57_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMArgValue.html" target="_self">TVMArgValue</a></td><td class="desc">A single argument value to <a class="el" href="classtvm_1_1runtime_1_1PackedFunc.html" title="Packed function is a type-erased function. The arguments are passed by packed format. ">PackedFun [...]
+<tr id="row_1_7_58_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMMovableArgValue__.html" target="_self">TVMMovableArgValue_</a></td><td class="desc">Internal auxiliary struct for <a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc.html" title="Please refer to TypedPackedFunc<R(Args..)>. ">TypedPackedFunc</a> [...]
+<tr id="row_1_7_59_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMMovableArgValueWithContext__.html" target="_self">TVMMovableArgValueWithContext_</a></td><td class="desc">Internal auxiliary struct for <a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc.html" title="Please refer to TypedPackedFunc<R(Args..)>. [...]
+<tr id="row_1_7_60_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMPODValue__.html" target="_self">TVMPODValue_</a></td><td class="desc">Internal base class to handle conversion to POD values </td></tr>
+<tr id="row_1_7_61_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TVMRetValue.html" target="_self">TVMRetValue</a></td><td class="desc">Return Value container, Unlike <a class="el" href="classtvm_1_1runtime_1_1TVMArgValue.html" title="A single argument value to PackedFunc. Containing both type_code and TVMValue. ">TVMArgValue [...]
+<tr id="row_1_7_62_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc.html" target="_self">TypedPackedFunc</a></td><td class="desc">Please refer to <a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc_3_01R_07Args_8_8_8_08_4.html#TypedPackedFuncAnchor">TypedPackedFunc<R(Args..)></a> </td></tr>
+<tr id="row_1_7_63_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtvm_1_1runtime_1_1TypedPackedFunc_3_01R_07Args_8_8_8_08_4.html" target="_self">TypedPackedFunc< R(Args...)></a></td><td class="desc">A <a class="el" href="classtvm_1_1runtime_1_1PackedFunc.html" title="Packed function is a type-erased function. The arguments are passed by [...]
+<tr id="row_1_7_64_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span class="icona"><span class="icon">C</span></span><a class="el" href="structtvm_1_1runtime_1_1TypeIndex.html" target="_self">TypeIndex</a></td><td class="desc">Namespace for the list of type index </td></tr>
<tr id="row_1_8_" class="even" style="display:none;"><td class="entry"><span style="width:16px;display:inline-block;"> </span><span id="arr_1_8_" class="arrow" onclick="toggleFolder('1_8_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1script.html" target="_self">script</a></td><td class="desc"></td></tr>
<tr id="row_1_8_0_" class="even" style="display:none;"><td class="entry"><span style="width:32px;display:inline-block;"> </span><span id="arr_1_8_0_" class="arrow" onclick="toggleFolder('1_8_0_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1script_1_1ir__builder.html" target="_self">ir_builder</a></td><td class="desc"></td></tr>
<tr id="row_1_8_0_0_" class="even" style="display:none;"><td class="entry"><span style="width:48px;display:inline-block;"> </span><span id="arr_1_8_0_0_" class="arrow" onclick="toggleFolder('1_8_0_0_')">►</span><span class="icona"><span class="icon">N</span></span><a class="el" href="namespacetvm_1_1script_1_1ir__builder_1_1details.html" target="_self">details</a></td><td class="desc"></td></tr>
diff --git a/docs/reference/api/doxygen/c__runtime__api_8h.html b/docs/reference/api/doxygen/c__runtime__api_8h.html
index 75f968bcc..fdc6baa06 100644
--- a/docs/reference/api/doxygen/c__runtime__api_8h.html
+++ b/docs/reference/api/doxygen/c__runtime__api_8h.html
@@ -84,7 +84,7 @@ Include dependency graph for c_runtime_api.h:</div>
</div><div class="textblock"><div class="dynheader">
This graph shows which files directly or indirectly include this file:</div>
<div class="dyncontent">
-<div class="center"><iframe scrolling="no" frameborder="0" src="c__runtime__api_8h__dep__incl.svg" width="3588" height="1124"><p><b>This browser is not able to show SVG: try Firefox, Chrome, Safari, or Opera instead.</b></p></iframe>
+<div class="center"><iframe scrolling="no" frameborder="0" src="c__runtime__api_8h__dep__incl.svg" width="4416" height="1124"><p><b>This browser is not able to show SVG: try Firefox, Chrome, Safari, or Opera instead.</b></p></iframe>
</div>
</div>
</div>
diff --git a/docs/reference/api/doxygen/c__runtime__api_8h__dep__incl.svg b/docs/reference/api/doxygen/c__runtime__api_8h__dep__incl.svg
index fc64aa387..7104fb5ce 100644
--- a/docs/reference/api/doxygen/c__runtime__api_8h__dep__incl.svg
+++ b/docs/reference/api/doxygen/c__runtime__api_8h__dep__incl.svg
@@ -4,1224 +4,1230 @@
<!-- Generated by graphviz version 2.40.1 (20161225.0304)
-->
<!-- Title: include/tvm/runtime/c_runtime_api.h Pages: 1 -->
-<svg width="2691pt" height="843pt"
- viewBox="0.00 0.00 2691.00 843.00" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
+<svg width="3312pt" height="843pt"
+ viewBox="0.00 0.00 3312.00 843.00" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
<g id="graph0" class="graph" transform="scale(1 1) rotate(0) translate(4 839)">
<title>include/tvm/runtime/c_runtime_api.h</title>
-<polygon fill="#ffffff" stroke="transparent" points="-4,4 -4,-839 2687,-839 2687,4 -4,4"/>
+<polygon fill="#ffffff" stroke="transparent" points="-4,4 -4,-839 3308,-839 3308,4 -4,4"/>
<!-- Node4 -->
<g id="node1" class="node">
<title>Node4</title>
-<polygon fill="#bfbfbf" stroke="#000000" points="1500,-804.5 1500,-834.5 1616,-834.5 1616,-804.5 1500,-804.5"/>
-<text text-anchor="start" x="1508" y="-822.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1558" y="-811.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/c_runtime_api.h</text>
+<polygon fill="#bfbfbf" stroke="#000000" points="2102,-804.5 2102,-834.5 2218,-834.5 2218,-804.5 2102,-804.5"/>
+<text text-anchor="start" x="2110" y="-822.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2160" y="-811.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/c_runtime_api.h</text>
</g>
<!-- Node5 -->
<g id="node2" class="node">
<title>Node5</title>
<g id="a_node2"><a xlink:href="compute__dag_8h.html" target="_top" xlink:title="The auto-scheduler's computational graph and related program analyses. ">
-<polygon fill="#ffffff" stroke="#000000" points="1770,-469.5 1770,-499.5 1922,-499.5 1922,-469.5 1770,-469.5"/>
-<text text-anchor="start" x="1778" y="-487.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
-<text text-anchor="middle" x="1846" y="-476.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/compute_dag.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="38,-469.5 38,-499.5 190,-499.5 190,-469.5 38,-469.5"/>
+<text text-anchor="start" x="46" y="-487.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
+<text text-anchor="middle" x="114" y="-476.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/compute_dag.h</text>
</a>
</g>
</g>
<!-- Node4->Node5 -->
<g id="edge1" class="edge">
<title>Node4->Node5</title>
-<path fill="none" stroke="#191970" d="M1572.5071,-795.6108C1577.5776,-787.0009 1583.1988,-777.1514 1588,-768 1610.3535,-725.3927 1605.0574,-708.4368 1634,-670 1690.3362,-595.1835 1781.0179,-528.4985 1823.1367,-499.6449"/>
-<polygon fill="#191970" stroke="#191970" points="1569.397,-793.9929 1567.286,-804.3753 1575.4108,-797.5754 1569.397,-793.9929"/>
+<path fill="none" stroke="#191970" d="M2091.5627,-818.544C1750.2112,-813.6718 246.3625,-790.9374 204,-768 164.8091,-746.7798 147,-730.0671 147,-685.5 147,-685.5 147,-685.5 147,-618.5 147,-574.4155 129.9521,-524.4779 120.3627,-499.8751"/>
+<polygon fill="#191970" stroke="#191970" points="2091.7459,-822.0469 2101.7948,-818.6897 2091.8457,-815.0476 2091.7459,-822.0469"/>
</g>
<!-- Node13 -->
<g id="node7" class="node">
<title>Node13</title>
<g id="a_node7"><a xlink:href="node_8h.html" target="_top" xlink:title="Definitions and helper macros for IR/AST nodes. ">
-<polygon fill="#ffffff" stroke="#000000" points="858.5,-475 858.5,-494 995.5,-494 995.5,-475 858.5,-475"/>
-<text text-anchor="middle" x="927" y="-482" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/node/node.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="937.5,-475 937.5,-494 1074.5,-494 1074.5,-475 937.5,-475"/>
+<text text-anchor="middle" x="1006" y="-482" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/node/node.h</text>
</a>
</g>
</g>
<!-- Node4->Node13 -->
<g id="edge6" class="edge">
<title>Node4->Node13</title>
-<path fill="none" stroke="#191970" d="M1489.7942,-818.327C1256.9775,-814.0592 505.6994,-798.0116 464,-768 427.4253,-741.6768 412.4032,-707.0867 438,-670 491.1883,-592.9365 776.0801,-519.6405 884.8007,-494.0629"/>
-<polygon fill="#191970" stroke="#191970" points="1489.7677,-821.827 1499.8297,-818.5095 1489.895,-814.8282 1489.7677,-821.827"/>
+<path fill="none" stroke="#191970" d="M2091.9783,-817.8079C1773.5981,-809.4502 446,-769.56 446,-685.5 446,-685.5 446,-685.5 446,-618.5 446,-518.5332 787.7969,-492.9774 937.333,-486.5803"/>
+<polygon fill="#191970" stroke="#191970" points="2091.9021,-821.307 2101.99,-818.0691 2092.0847,-814.3094 2091.9021,-821.307"/>
</g>
<!-- Node176 -->
<g id="node20" class="node">
<title>Node176</title>
<g id="a_node20"><a xlink:href="tir_2expr_8h.html" target="_top" xlink:title="TIR expressions. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="157.5,-140 157.5,-159 278.5,-159 278.5,-140 157.5,-140"/>
-<text text-anchor="middle" x="218" y="-147" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/tir/expr.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="909.5,-140 909.5,-159 1030.5,-159 1030.5,-140 909.5,-140"/>
+<text text-anchor="middle" x="970" y="-147" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/tir/expr.h</text>
</a>
</g>
</g>
<!-- Node4->Node176 -->
-<g id="edge121" class="edge">
+<g id="edge122" class="edge">
<title>Node4->Node176</title>
-<path fill="none" stroke="#191970" d="M1489.7755,-818.3853C1231.2208,-813.961 319.8704,-796.4816 193,-768 101.9794,-747.5665 0,-778.786 0,-685.5 0,-685.5 0,-685.5 0,-278 0,-204.632 92.7704,-172.3028 157.2748,-158.6916"/>
-<polygon fill="#191970" stroke="#191970" points="1489.727,-821.8849 1499.7852,-818.5556 1489.8462,-814.8859 1489.727,-821.8849"/>
+<path fill="none" stroke="#191970" d="M2091.634,-818.4163C1836.3978,-814.1752 939.3832,-797.4683 655,-768 458.3997,-747.6279 218,-883.153 218,-685.5 218,-685.5 218,-685.5 218,-484.5 218,-437.9897 231.747,-423.5184 265,-391 282.5893,-373.7993 641.5364,-208.3962 665,-201 747.4588,-175.0073 846.8344,-161.3391 909.2343,-154.7742"/>
+<polygon fill="#191970" stroke="#191970" points="2091.8181,-821.9198 2101.8746,-818.5855 2091.9338,-814.9207 2091.8181,-821.9198"/>
</g>
<!-- Node190 -->
<g id="node29" class="node">
<title>Node190</title>
<g id="a_node29"><a xlink:href="reflection_8h.html" target="_top" xlink:title="Reflection and serialization of compiler IR/AST nodes. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="861.5,-542 861.5,-561 1020.5,-561 1020.5,-542 861.5,-542"/>
-<text text-anchor="middle" x="941" y="-549" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/node/reflection.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1205.5,-542 1205.5,-561 1364.5,-561 1364.5,-542 1205.5,-542"/>
+<text text-anchor="middle" x="1285" y="-549" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/node/reflection.h</text>
</a>
</g>
</g>
<!-- Node4->Node190 -->
<g id="edge57" class="edge">
<title>Node4->Node190</title>
-<path fill="none" stroke="#191970" d="M1489.6438,-818.9697C1271.756,-816.8874 602.6937,-807.3259 514,-768 475.5024,-750.9305 463.1156,-739.9065 447,-701 441.7275,-688.271 438.265,-680.6549 447,-670 498.7235,-606.9074 547.5117,-652.374 627,-634 729.693,-610.2622 849.9035,-577.1831 907.2441,-561.0759"/>
-<polygon fill="#191970" stroke="#191970" points="1489.642,-822.4698 1499.6742,-819.0633 1489.7073,-815.4701 1489.642,-822.4698"/>
+<path fill="none" stroke="#191970" d="M2091.5384,-817.5661C1909.5971,-812.082 1423.3151,-795.0524 1356,-768 1270.2499,-733.539 1235.9665,-717.7989 1197,-634 1191.1907,-621.5068 1190.1463,-614.9521 1197,-603 1208.9758,-582.1156 1233.0632,-568.8863 1253.0138,-561.0725"/>
+<polygon fill="#191970" stroke="#191970" points="2091.6847,-821.072 2101.7849,-817.8725 2091.894,-814.0751 2091.6847,-821.072"/>
</g>
<!-- Node196 -->
-<g id="node31" class="node">
+<g id="node30" class="node">
<title>Node196</title>
-<g id="a_node31"><a xlink:href="serialization_8h.html" target="_top" xlink:title="include/tvm/node/serialization.h">
-<polygon fill="#ffffff" stroke="#000000" points="488.5,-676 488.5,-695 661.5,-695 661.5,-676 488.5,-676"/>
-<text text-anchor="middle" x="575" y="-683" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/node/serialization.h</text>
+<g id="a_node30"><a xlink:href="serialization_8h.html" target="_top" xlink:title="include/tvm/node/serialization.h">
+<polygon fill="#ffffff" stroke="#000000" points="1920.5,-676 1920.5,-695 2093.5,-695 2093.5,-676 1920.5,-676"/>
+<text text-anchor="middle" x="2007" y="-683" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/node/serialization.h</text>
</a>
</g>
</g>
<!-- Node4->Node196 -->
-<g id="edge61" class="edge">
+<g id="edge60" class="edge">
<title>Node4->Node196</title>
-<path fill="none" stroke="#191970" d="M1489.5909,-816.4183C1336.914,-809.2314 972.2842,-790.2689 851,-768 756.3107,-750.6141 647.9656,-712.7267 600.0422,-695.0007"/>
-<polygon fill="#191970" stroke="#191970" points="1489.6212,-819.9234 1499.7739,-816.8949 1489.9485,-812.9311 1489.6212,-819.9234"/>
+<path fill="none" stroke="#191970" d="M2135.1335,-797.7215C2101.3294,-768.1153 2042.3003,-716.4166 2017.9659,-695.1042"/>
+<polygon fill="#191970" stroke="#191970" points="2132.9176,-800.4334 2142.7464,-804.389 2137.5296,-795.1675 2132.9176,-800.4334"/>
</g>
<!-- Node24 -->
-<g id="node32" class="node">
+<g id="node31" class="node">
<title>Node24</title>
-<g id="a_node32"><a xlink:href="relay_2qnn_2transform_8h.html" target="_top" xlink:title="include/tvm/relay/qnn\l/transform.h">
-<polygon fill="#ffffff" stroke="#000000" points="1815.5,-737.5 1815.5,-767.5 1938.5,-767.5 1938.5,-737.5 1815.5,-737.5"/>
-<text text-anchor="start" x="1823.5" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/relay/qnn</text>
-<text text-anchor="middle" x="1877" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/transform.h</text>
+<g id="a_node31"><a xlink:href="relay_2qnn_2transform_8h.html" target="_top" xlink:title="include/tvm/relay/qnn\l/transform.h">
+<polygon fill="#ffffff" stroke="#000000" points="2436.5,-737.5 2436.5,-767.5 2559.5,-767.5 2559.5,-737.5 2436.5,-737.5"/>
+<text text-anchor="start" x="2444.5" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/relay/qnn</text>
+<text text-anchor="middle" x="2498" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/transform.h</text>
</a>
</g>
</g>
<!-- Node4->Node24 -->
-<g id="edge62" class="edge">
+<g id="edge61" class="edge">
<title>Node4->Node24</title>
-<path fill="none" stroke="#191970" d="M1626.1779,-805.5368C1676.0203,-795.2814 1745.2416,-780.9403 1806,-768 1809.085,-767.343 1812.2445,-766.6664 1815.4347,-765.9805"/>
-<polygon fill="#191970" stroke="#191970" points="1625.2516,-802.154 1616.1615,-807.5964 1626.6615,-809.0105 1625.2516,-802.154"/>
+<path fill="none" stroke="#191970" d="M2227.9806,-806.2747C2280.884,-795.9493 2356.1322,-781.1903 2422,-768 2426.6214,-767.0746 2431.3997,-766.1124 2436.2056,-765.141"/>
+<polygon fill="#191970" stroke="#191970" points="2227.1835,-802.8641 2218.0388,-808.2142 2228.5239,-809.7346 2227.1835,-802.8641"/>
</g>
<!-- Node197 -->
-<g id="node33" class="node">
+<g id="node32" class="node">
<title>Node197</title>
-<g id="a_node33"><a xlink:href="builtin__fp16_8h.html" target="_top" xlink:title="Functions for conversion between fp32 and fp16. ">
-<polygon fill="#ffffff" stroke="#000000" points="1957,-737.5 1957,-767.5 2073,-767.5 2073,-737.5 1957,-737.5"/>
-<text text-anchor="start" x="1965" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="2015" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/builtin_fp16.h</text>
+<g id="a_node32"><a xlink:href="builtin__fp16_8h.html" target="_top" xlink:title="Functions for conversion between fp32 and fp16. ">
+<polygon fill="#ffffff" stroke="#000000" points="2578,-737.5 2578,-767.5 2694,-767.5 2694,-737.5 2578,-737.5"/>
+<text text-anchor="start" x="2586" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2636" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/builtin_fp16.h</text>
</a>
</g>
</g>
<!-- Node4->Node197 -->
-<g id="edge63" class="edge">
+<g id="edge62" class="edge">
<title>Node4->Node197</title>
-<path fill="none" stroke="#191970" d="M1626.4254,-812.3798C1704.3544,-803.8218 1835.3473,-788.0716 1947,-768 1950.2177,-767.4216 1953.513,-766.7927 1956.8339,-766.1305"/>
-<polygon fill="#191970" stroke="#191970" points="1625.8179,-808.9252 1616.2563,-813.4887 1626.5769,-815.8839 1625.8179,-808.9252"/>
+<path fill="none" stroke="#191970" d="M2228.1377,-812.974C2309.3967,-804.7133 2449.1665,-788.9908 2568,-768 2571.2194,-767.4313 2574.5161,-766.8102 2577.8381,-766.154"/>
+<polygon fill="#191970" stroke="#191970" points="2227.6369,-809.5067 2218.0388,-813.9928 2228.3396,-816.4714 2227.6369,-809.5067"/>
</g>
<!-- Node198 -->
-<g id="node34" class="node">
+<g id="node33" class="node">
<title>Node198</title>
-<g id="a_node34"><a xlink:href="c__backend__api_8h.html" target="_top" xlink:title="TVM runtime backend API. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="2091,-737.5 2091,-767.5 2207,-767.5 2207,-737.5 2091,-737.5"/>
-<text text-anchor="start" x="2099" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="2149" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/c_backend_api.h</text>
+<g id="a_node33"><a xlink:href="c__backend__api_8h.html" target="_top" xlink:title="TVM runtime backend API. ">
+<polygon fill="#ffffff" stroke="#ff0000" points="2712,-737.5 2712,-767.5 2828,-767.5 2828,-737.5 2712,-737.5"/>
+<text text-anchor="start" x="2720" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2770" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/c_backend_api.h</text>
</a>
</g>
</g>
<!-- Node4->Node198 -->
-<g id="edge64" class="edge">
+<g id="edge63" class="edge">
<title>Node4->Node198</title>
-<path fill="none" stroke="#191970" d="M1626.0861,-815.6303C1726.3954,-809.2922 1919.5734,-794.6551 2082,-768 2084.9454,-767.5166 2087.9545,-766.9799 2090.9867,-766.4047"/>
-<polygon fill="#191970" stroke="#191970" points="1625.8476,-812.1383 1616.0849,-816.2541 1626.2834,-819.1247 1625.8476,-812.1383"/>
+<path fill="none" stroke="#191970" d="M2228.4773,-815.9294C2331.818,-809.8823 2533.5629,-795.5092 2703,-768 2705.9462,-767.5217 2708.956,-766.9889 2711.9887,-766.417"/>
+<polygon fill="#191970" stroke="#191970" points="2227.9654,-812.453 2218.1836,-816.5231 2228.3685,-819.4414 2227.9654,-812.453"/>
</g>
<!-- Node202 -->
-<g id="node35" class="node">
+<g id="node34" class="node">
<title>Node202</title>
-<g id="a_node35"><a xlink:href="graph__executor_8h.html" target="_top" xlink:title="Tiny AoT executor. ">
-<polygon fill="#ffffff" stroke="#000000" points="2304.5,-603.5 2304.5,-633.5 2423.5,-633.5 2423.5,-603.5 2304.5,-603.5"/>
-<text text-anchor="start" x="2312.5" y="-621.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="2364" y="-610.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/crt/graph_executor.h</text>
+<g id="a_node34"><a xlink:href="graph__executor_8h.html" target="_top" xlink:title="Tiny AoT executor. ">
+<polygon fill="#ffffff" stroke="#000000" points="2839.5,-603.5 2839.5,-633.5 2958.5,-633.5 2958.5,-603.5 2839.5,-603.5"/>
+<text text-anchor="start" x="2847.5" y="-621.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2899" y="-610.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/crt/graph_executor.h</text>
</a>
</g>
</g>
<!-- Node4->Node202 -->
-<g id="edge65" class="edge">
+<g id="edge64" class="edge">
<title>Node4->Node202</title>
-<path fill="none" stroke="#191970" d="M1626.2715,-817.7835C1820.1927,-812.5781 2362.0458,-795.6145 2388,-768 2423.106,-730.6482 2390.805,-663.4573 2373.4184,-633.5574"/>
-<polygon fill="#191970" stroke="#191970" points="1626.1046,-814.2866 1616.2014,-818.0516 1626.291,-821.2841 1626.1046,-814.2866"/>
+<path fill="none" stroke="#191970" d="M2228.4661,-818.3835C2390.7678,-815.1625 2788.9219,-803.6773 2837,-768 2874.5145,-740.1615 2856.6792,-712.9729 2875,-670 2880.2612,-657.6594 2886.7058,-643.8961 2891.5916,-633.7044"/>
+<polygon fill="#191970" stroke="#191970" points="2228.0532,-814.8908 2218.1229,-818.5839 2228.1888,-821.8895 2228.0532,-814.8908"/>
</g>
<!-- Node201 -->
-<g id="node36" class="node">
+<g id="node35" class="node">
<title>Node201</title>
-<g id="a_node36"><a xlink:href="crt_2packed__func_8h.html" target="_top" xlink:title="Type-erased function used across TVM API. ">
-<polygon fill="#ffffff" stroke="#000000" points="2263,-670.5 2263,-700.5 2379,-700.5 2379,-670.5 2263,-670.5"/>
-<text text-anchor="start" x="2271" y="-688.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="2321" y="-677.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/crt/packed_func.h</text>
+<g id="a_node35"><a xlink:href="crt_2packed__func_8h.html" target="_top" xlink:title="Type-erased function used across TVM API. ">
+<polygon fill="#ffffff" stroke="#000000" points="2884,-670.5 2884,-700.5 3000,-700.5 3000,-670.5 2884,-670.5"/>
+<text text-anchor="start" x="2892" y="-688.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2942" y="-677.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/crt/packed_func.h</text>
</a>
</g>
</g>
<!-- Node4->Node201 -->
-<g id="edge66" class="edge">
+<g id="edge65" class="edge">
<title>Node4->Node201</title>
-<path fill="none" stroke="#191970" d="M1626.163,-816.4728C1783.4186,-809.1594 2162.2712,-789.5349 2216,-768 2236.2314,-759.8911 2236.9134,-750.5318 2254,-737 2269.6417,-724.6125 2287.5733,-710.852 2300.9472,-700.6735"/>
-<polygon fill="#191970" stroke="#191970" points="1625.9588,-812.9785 1616.1311,-816.9367 1626.2821,-819.971 1625.9588,-812.9785"/>
+<path fill="none" stroke="#191970" d="M2228.4466,-817.878C2425.5927,-812.8753 2982.3723,-796.2919 3009,-768 3029.9921,-745.6959 2995.5885,-717.6286 2968.8039,-700.6504"/>
+<polygon fill="#191970" stroke="#191970" points="2228.1235,-814.3849 2218.2147,-818.1354 2228.2996,-821.3827 2228.1235,-814.3849"/>
</g>
<!-- Node203 -->
-<g id="node37" class="node">
+<g id="node36" class="node">
<title>Node203</title>
-<g id="a_node37"><a xlink:href="page__allocator_8h.html" target="_top" xlink:title="An implementation of a dynamic memory allocator for microcontrollers. ">
-<polygon fill="#ffffff" stroke="#000000" points="2435,-737.5 2435,-767.5 2551,-767.5 2551,-737.5 2435,-737.5"/>
-<text text-anchor="start" x="2443" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="2493" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/crt/page_allocator.h</text>
+<g id="a_node36"><a xlink:href="page__allocator_8h.html" target="_top" xlink:title="An implementation of a dynamic memory allocator for microcontrollers. ">
+<polygon fill="#ffffff" stroke="#000000" points="3056,-737.5 3056,-767.5 3172,-767.5 3172,-737.5 3056,-737.5"/>
+<text text-anchor="start" x="3064" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="3114" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/crt/page_allocator.h</text>
</a>
</g>
</g>
<!-- Node4->Node203 -->
-<g id="edge68" class="edge">
+<g id="edge67" class="edge">
<title>Node4->Node203</title>
-<path fill="none" stroke="#191970" d="M1626.3199,-816.7715C1802.2609,-809.5286 2267.8125,-788.9353 2421,-768 2425.5607,-767.3767 2430.2625,-766.6222 2434.9756,-765.7857"/>
-<polygon fill="#191970" stroke="#191970" points="1626.0032,-813.2815 1616.155,-817.1884 1626.29,-820.2756 1626.0032,-813.2815"/>
+<path fill="none" stroke="#191970" d="M2228.2783,-816.8739C2406.8958,-809.7861 2884.9873,-789.3556 3042,-768 3046.5611,-767.3796 3051.2632,-766.6272 3055.9765,-765.7922"/>
+<polygon fill="#191970" stroke="#191970" points="2228.073,-813.3792 2218.2191,-817.2716 2228.3495,-820.3738 2228.073,-813.3792"/>
</g>
<!-- Node204 -->
-<g id="node38" class="node">
+<g id="node37" class="node">
<title>Node204</title>
-<g id="a_node38"><a xlink:href="platform_8h.html" target="_top" xlink:title="The virtual memory manager for micro-controllers. ">
-<polygon fill="#ffffff" stroke="#000000" points="2263,-737.5 2263,-767.5 2379,-767.5 2379,-737.5 2263,-737.5"/>
-<text text-anchor="start" x="2271" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="2321" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/crt/platform.h</text>
+<g id="a_node37"><a xlink:href="platform_8h.html" target="_top" xlink:title="The virtual memory manager for micro-controllers. ">
+<polygon fill="#ffffff" stroke="#000000" points="2884,-737.5 2884,-767.5 3000,-767.5 3000,-737.5 2884,-737.5"/>
+<text text-anchor="start" x="2892" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2942" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/crt/platform.h</text>
</a>
</g>
</g>
<!-- Node4->Node204 -->
-<g id="edge69" class="edge">
+<g id="edge68" class="edge">
<title>Node4->Node204</title>
-<path fill="none" stroke="#191970" d="M1626.2139,-817.5278C1750.5859,-813.242 2022.3006,-800.691 2249,-768 2253.556,-767.343 2258.2544,-766.5643 2262.9651,-765.711"/>
-<polygon fill="#191970" stroke="#191970" points="1625.9545,-814.0345 1616.0783,-817.8699 1626.1907,-821.0305 1625.9545,-814.0345"/>
+<path fill="none" stroke="#191970" d="M2228.6275,-817.7049C2355.7417,-813.6751 2636.1755,-801.4818 2870,-768 2874.5566,-767.3475 2879.2555,-766.572 2883.9666,-765.7211"/>
+<polygon fill="#191970" stroke="#191970" points="2228.1617,-814.2176 2218.2749,-818.0257 2228.3786,-821.2142 2228.1617,-814.2176"/>
</g>
<!-- Node205 -->
-<g id="node39" class="node">
+<g id="node38" class="node">
<title>Node205</title>
-<g id="a_node39"><a xlink:href="data__type_8h.html" target="_top" xlink:title="include/tvm/runtime\l/data_type.h">
-<polygon fill="#ffffff" stroke="#ff0000" points="202,-737.5 202,-767.5 318,-767.5 318,-737.5 202,-737.5"/>
-<text text-anchor="start" x="210" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="260" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/data_type.h</text>
+<g id="a_node38"><a xlink:href="data__type_8h.html" target="_top" xlink:title="include/tvm/runtime\l/data_type.h">
+<polygon fill="#ffffff" stroke="#ff0000" points="996,-737.5 996,-767.5 1112,-767.5 1112,-737.5 996,-737.5"/>
+<text text-anchor="start" x="1004" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1054" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/data_type.h</text>
</a>
</g>
</g>
<!-- Node4->Node205 -->
-<g id="edge71" class="edge">
+<g id="edge70" class="edge">
<title>Node4->Node205</title>
-<path fill="none" stroke="#191970" d="M1489.5664,-818.1584C1266.0139,-813.535 558.6086,-796.9377 332,-768 327.434,-767.4169 322.7283,-766.6913 318.0126,-765.8748"/>
-<polygon fill="#191970" stroke="#191970" points="1489.7723,-821.6633 1499.8421,-818.3696 1489.9162,-814.6648 1489.7723,-821.6633"/>
+<path fill="none" stroke="#191970" d="M2091.646,-817.5453C1892.2648,-811.6136 1313.6329,-792.7031 1126,-768 1121.4363,-767.3992 1116.7323,-766.6608 1112.0177,-765.8354"/>
+<polygon fill="#191970" stroke="#191970" points="2091.5987,-821.0453 2101.6979,-817.8429 2091.8059,-814.0484 2091.5987,-821.0453"/>
</g>
<!-- Node208 -->
-<g id="node40" class="node">
+<g id="node39" class="node">
<title>Node208</title>
-<g id="a_node40"><a xlink:href="ndarray_8h.html" target="_top" xlink:title="A device-independent managed NDArray abstraction. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="1014,-670.5 1014,-700.5 1130,-700.5 1130,-670.5 1014,-670.5"/>
-<text text-anchor="start" x="1022" y="-688.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1072" y="-677.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/ndarray.h</text>
+<g id="a_node39"><a xlink:href="ndarray_8h.html" target="_top" xlink:title="A device-independent managed NDArray abstraction. ">
+<polygon fill="#ffffff" stroke="#ff0000" points="1330,-670.5 1330,-700.5 1446,-700.5 1446,-670.5 1330,-670.5"/>
+<text text-anchor="start" x="1338" y="-688.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1388" y="-677.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/ndarray.h</text>
</a>
</g>
</g>
<!-- Node4->Node208 -->
-<g id="edge100" class="edge">
+<g id="edge101" class="edge">
<title>Node4->Node208</title>
-<path fill="none" stroke="#191970" d="M1489.6894,-813.4807C1426.2466,-806.6398 1329.6992,-793.0212 1249,-768 1193.1404,-750.6804 1132.1945,-719.243 1098.3796,-700.5776"/>
-<polygon fill="#191970" stroke="#191970" points="1489.6577,-816.9966 1499.9687,-814.561 1490.3893,-810.0349 1489.6577,-816.9966"/>
+<path fill="none" stroke="#191970" d="M2091.9188,-817.425C1927.0082,-811.9838 1515.5884,-795.7733 1460,-768 1430.2395,-753.1309 1407.5216,-720.0384 1396.1248,-700.5856"/>
+<polygon fill="#191970" stroke="#191970" points="2091.8384,-820.9242 2101.9474,-817.7527 2092.067,-813.9279 2091.8384,-820.9242"/>
</g>
<!-- Node210 -->
-<g id="node41" class="node">
+<g id="node40" class="node">
<title>Node210</title>
-<g id="a_node41"><a xlink:href="device__api_8h.html" target="_top" xlink:title="Abstract device memory management API. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="1327,-536.5 1327,-566.5 1443,-566.5 1443,-536.5 1327,-536.5"/>
-<text text-anchor="start" x="1335" y="-554.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1385" y="-543.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/device_api.h</text>
+<g id="a_node40"><a xlink:href="device__api_8h.html" target="_top" xlink:title="Abstract device memory management API. ">
+<polygon fill="#ffffff" stroke="#ff0000" points="1858,-536.5 1858,-566.5 1974,-566.5 1974,-536.5 1858,-536.5"/>
+<text text-anchor="start" x="1866" y="-554.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1916" y="-543.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/device_api.h</text>
</a>
</g>
</g>
<!-- Node4->Node210 -->
<g id="edge95" class="edge">
<title>Node4->Node210</title>
-<path fill="none" stroke="#191970" d="M1550.5323,-794.9129C1536.7167,-752.3471 1503.9449,-663.6728 1454,-603 1442.0298,-588.4587 1425.1276,-575.7557 1411.0481,-566.604"/>
-<polygon fill="#191970" stroke="#191970" points="1547.2011,-795.9871 1553.5621,-804.46 1553.8732,-793.8697 1547.2011,-795.9871"/>
+<path fill="none" stroke="#191970" d="M2161.3915,-794.0966C2162.3833,-747.5781 2156.9456,-649.252 2098,-603 2078.6889,-587.8474 2019.8637,-572.8291 1974.2614,-562.9559"/>
+<polygon fill="#191970" stroke="#191970" points="2157.8888,-794.1209 2161.0532,-804.2321 2164.8849,-794.3545 2157.8888,-794.1209"/>
</g>
<!-- Node211 -->
-<g id="node42" class="node">
+<g id="node41" class="node">
<title>Node211</title>
-<g id="a_node42"><a xlink:href="profiling_8h.html" target="_top" xlink:title="Runtime profiling including timers. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="1318,-469.5 1318,-499.5 1434,-499.5 1434,-469.5 1318,-469.5"/>
-<text text-anchor="start" x="1326" y="-487.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1376" y="-476.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/profiling.h</text>
+<g id="a_node41"><a xlink:href="conv2d_8h.html" target="_top" xlink:title="include/tvm/runtime\l/hexagon/ops/conv2d.h">
+<polygon fill="#ffffff" stroke="#000000" points="2171.5,-469.5 2171.5,-499.5 2300.5,-499.5 2300.5,-469.5 2171.5,-469.5"/>
+<text text-anchor="start" x="2179.5" y="-487.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2236" y="-476.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/hexagon/ops/conv2d.h</text>
</a>
</g>
</g>
<!-- Node4->Node211 -->
-<g id="edge117" class="edge">
+<g id="edge96" class="edge">
<title>Node4->Node211</title>
-<path fill="none" stroke="#191970" d="M1563.2316,-794.0743C1569.9357,-763.7705 1582.8931,-711.9611 1601,-670 1626.9259,-609.919 1703.9705,-586.2031 1662,-536 1633.5642,-501.9865 1509.3762,-490.4024 1434.3104,-486.4822"/>
-<polygon fill="#191970" stroke="#191970" points="1559.7331,-793.6947 1561.0466,-804.2078 1566.5758,-795.1702 1559.7331,-793.6947"/>
+<path fill="none" stroke="#191970" d="M2187.2194,-797.543C2196.1906,-789.1203 2205.5185,-778.895 2212,-768 2231.5237,-735.1816 2236,-723.6867 2236,-685.5 2236,-685.5 2236,-685.5 2236,-618.5 2236,-575.4618 2236,-524.6482 2236,-499.7729"/>
+<polygon fill="#191970" stroke="#191970" points="2184.6203,-795.1723 2179.5081,-804.4522 2189.2915,-800.3858 2184.6203,-795.1723"/>
</g>
-<!-- Node213 -->
-<g id="node43" class="node">
-<title>Node213</title>
-<g id="a_node43"><a xlink:href="packed__func_8h.html" target="_top" xlink:title="Type-erased function used across TVM API. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="1329,-603.5 1329,-633.5 1445,-633.5 1445,-603.5 1329,-603.5"/>
-<text text-anchor="start" x="1337" y="-621.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1387" y="-610.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/packed_func.h</text>
+<!-- Node212 -->
+<g id="node42" class="node">
+<title>Node212</title>
+<g id="a_node42"><a xlink:href="profiling_8h.html" target="_top" xlink:title="Runtime profiling including timers. ">
+<polygon fill="#ffffff" stroke="#ff0000" points="1815,-469.5 1815,-499.5 1931,-499.5 1931,-469.5 1815,-469.5"/>
+<text text-anchor="start" x="1823" y="-487.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1873" y="-476.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/profiling.h</text>
</a>
</g>
</g>
-<!-- Node4->Node213 -->
-<g id="edge116" class="edge">
-<title>Node4->Node213</title>
-<path fill="none" stroke="#191970" d="M1512.6297,-799.8571C1497.169,-791.6181 1480.6202,-780.9167 1468,-768 1428.1844,-727.2492 1402.3649,-662.971 1392.0682,-633.802"/>
-<polygon fill="#191970" stroke="#191970" points="1511.0713,-802.9912 1521.5692,-804.4217 1514.2547,-796.7569 1511.0713,-802.9912"/>
+<!-- Node4->Node212 -->
+<g id="edge118" class="edge">
+<title>Node4->Node212</title>
+<path fill="none" stroke="#191970" d="M2168.7197,-794.8459C2171.2775,-786.4406 2173.7283,-776.9138 2175,-768 2185.8044,-692.2681 2181.8586,-655.2673 2126,-603 2069.2454,-549.8943 1984.4077,-516.8406 1928.8847,-499.544"/>
+<polygon fill="#191970" stroke="#191970" points="2165.3685,-793.8339 2165.6187,-804.4257 2172.0283,-795.9897 2165.3685,-793.8339"/>
</g>
<!-- Node214 -->
-<g id="node44" class="node">
+<g id="node43" class="node">
<title>Node214</title>
-<g id="a_node44"><a xlink:href="runtime_2module_8h.html" target="_top" xlink:title="Runtime container of the functions generated by TVM, This is used to support dynamically link...">
-<polygon fill="#ffffff" stroke="#ff0000" points="1537,-536.5 1537,-566.5 1653,-566.5 1653,-536.5 1537,-536.5"/>
-<text text-anchor="start" x="1545" y="-554.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1595" y="-543.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/module.h</text>
+<g id="a_node43"><a xlink:href="packed__func_8h.html" target="_top" xlink:title="Type-erased function used across TVM API. ">
+<polygon fill="#ffffff" stroke="#ff0000" points="1648,-603.5 1648,-633.5 1764,-633.5 1764,-603.5 1648,-603.5"/>
+<text text-anchor="start" x="1656" y="-621.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1706" y="-610.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/packed_func.h</text>
</a>
</g>
</g>
<!-- Node4->Node214 -->
-<g id="edge99" class="edge">
+<g id="edge117" class="edge">
<title>Node4->Node214</title>
-<path fill="none" stroke="#191970" d="M1557.0717,-793.7247C1556.4068,-763.7999 1556.5568,-713.0424 1563,-670 1568.6787,-632.065 1582.192,-588.8915 1589.7151,-566.6006"/>
-<polygon fill="#191970" stroke="#191970" points="1553.5836,-794.2133 1557.355,-804.1141 1560.581,-794.0224 1553.5836,-794.2133"/>
+<path fill="none" stroke="#191970" d="M2091.6252,-817.0784C1966.226,-811.9294 1709.792,-797.8403 1683,-768 1648.9649,-730.0924 1679.7227,-663.7837 1696.6173,-633.8889"/>
+<polygon fill="#191970" stroke="#191970" points="2091.7495,-820.5862 2101.8823,-817.4916 2092.0313,-813.5918 2091.7495,-820.5862"/>
</g>
-<!-- Node219 -->
-<g id="node45" class="node">
-<title>Node219</title>
-<g id="a_node45"><a xlink:href="serializer_8h.html" target="_top" xlink:title="Serializer extension to support TVM data types Include this file to enable serialization of DLDataTyp...">
-<polygon fill="#ffffff" stroke="#000000" points="1119,-603.5 1119,-633.5 1235,-633.5 1235,-603.5 1119,-603.5"/>
-<text text-anchor="start" x="1127" y="-621.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1177" y="-610.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/serializer.h</text>
+<!-- Node215 -->
+<g id="node44" class="node">
+<title>Node215</title>
+<g id="a_node44"><a xlink:href="runtime_2module_8h.html" target="_top" xlink:title="Runtime container of the functions generated by TVM, This is used to support dynamically link...">
+<polygon fill="#ffffff" stroke="#ff0000" points="1686,-536.5 1686,-566.5 1802,-566.5 1802,-536.5 1686,-536.5"/>
+<text text-anchor="start" x="1694" y="-554.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1744" y="-543.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/module.h</text>
</a>
</g>
</g>
-<!-- Node4->Node219 -->
-<g id="edge118" class="edge">
-<title>Node4->Node219</title>
-<path fill="none" stroke="#191970" d="M1511.4471,-800.5202C1489.5206,-791.2419 1463.1559,-779.5941 1440,-768 1350.0655,-722.97 1247.5135,-661.6528 1201.5459,-633.6131"/>
-<polygon fill="#191970" stroke="#191970" points="1510.2898,-803.8303 1520.865,-804.4757 1513.0005,-797.3764 1510.2898,-803.8303"/>
+<!-- Node4->Node215 -->
+<g id="edge100" class="edge">
+<title>Node4->Node215</title>
+<path fill="none" stroke="#191970" d="M2156.8771,-794.2636C2151.478,-761.2336 2137.3469,-703.6078 2102,-670 2058.2902,-628.4407 1892.0524,-585.3203 1802.3018,-564.4139"/>
+<polygon fill="#191970" stroke="#191970" points="2153.4592,-795.0752 2158.3914,-804.452 2160.3831,-794.046 2153.4592,-795.0752"/>
</g>
<!-- Node220 -->
-<g id="node46" class="node">
+<g id="node45" class="node">
<title>Node220</title>
-<g id="a_node46"><a xlink:href="memory__manager_8h.html" target="_top" xlink:title="Abstract device memory management API. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="849.5,-603.5 849.5,-633.5 986.5,-633.5 986.5,-603.5 849.5,-603.5"/>
-<text text-anchor="start" x="857.5" y="-621.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="918" y="-610.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/vm/memory_manager.h</text>
+<g id="a_node45"><a xlink:href="serializer_8h.html" target="_top" xlink:title="Serializer extension to support TVM data types Include this file to enable serialization of DLDataTyp...">
+<polygon fill="#ffffff" stroke="#000000" points="1206,-603.5 1206,-633.5 1322,-633.5 1322,-603.5 1206,-603.5"/>
+<text text-anchor="start" x="1214" y="-621.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1264" y="-610.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/serializer.h</text>
</a>
</g>
</g>
<!-- Node4->Node220 -->
<g id="edge119" class="edge">
<title>Node4->Node220</title>
-<path fill="none" stroke="#191970" d="M1489.5912,-817.0723C1317.8091,-810.6865 871.5715,-792.2295 725,-768 613.2014,-749.5187 547.7975,-791.0411 479,-701 414.679,-616.8177 351.8214,-687.7094 797,-634 813.926,-631.9579 832.1718,-629.6675 849.2335,-627.4868"/>
-<polygon fill="#191970" stroke="#191970" points="1489.648,-820.5767 1499.7705,-817.4485 1489.9066,-813.5814 1489.648,-820.5767"/>
+<path fill="none" stroke="#191970" d="M2091.7895,-816.7774C1920.6899,-809.6611 1482.4854,-789.5581 1420,-768 1369.7759,-750.6722 1358.7486,-738.3866 1321,-701 1300.2198,-680.419 1282.1657,-651.3208 1272.2091,-633.7477"/>
+<polygon fill="#191970" stroke="#191970" points="2091.7993,-820.2808 2101.9355,-817.1972 2092.0888,-813.2867 2091.7993,-820.2808"/>
+</g>
+<!-- Node221 -->
+<g id="node46" class="node">
+<title>Node221</title>
+<g id="a_node46"><a xlink:href="memory__manager_8h.html" target="_top" xlink:title="Abstract device memory management API. ">
+<polygon fill="#ffffff" stroke="#ff0000" points="1416.5,-603.5 1416.5,-633.5 1553.5,-633.5 1553.5,-603.5 1416.5,-603.5"/>
+<text text-anchor="start" x="1424.5" y="-621.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1485" y="-610.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/vm/memory_manager.h</text>
+</a>
+</g>
+</g>
+<!-- Node4->Node221 -->
+<g id="edge120" class="edge">
+<title>Node4->Node221</title>
+<path fill="none" stroke="#191970" d="M2091.7713,-815.436C1975.3472,-807.9956 1743.3042,-790.8271 1665,-768 1601.0757,-749.3649 1579.13,-747.0118 1531,-701 1510.8487,-681.7356 1497.1758,-651.676 1490.2635,-633.6537"/>
+<polygon fill="#191970" stroke="#191970" points="2091.7881,-818.944 2101.9893,-816.083 2092.2305,-811.958 2091.7881,-818.944"/>
</g>
<!-- Node168 -->
<g id="node47" class="node">
<title>Node168</title>
<g id="a_node47"><a xlink:href="metadata_8h.html" target="_top" xlink:title="Defines types which can be used in Metadata. ">
-<polygon fill="#ffffff" stroke="#000000" points="1643,-670.5 1643,-700.5 1759,-700.5 1759,-670.5 1643,-670.5"/>
-<text text-anchor="start" x="1651" y="-688.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1701" y="-677.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/metadata.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="2264,-670.5 2264,-700.5 2380,-700.5 2380,-670.5 2264,-670.5"/>
+<text text-anchor="start" x="2272" y="-688.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2322" y="-677.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/metadata.h</text>
</a>
</g>
</g>
<!-- Node4->Node168 -->
-<g id="edge96" class="edge">
+<g id="edge97" class="edge">
<title>Node4->Node168</title>
-<path fill="none" stroke="#191970" d="M1590.134,-798.5005C1602.7463,-789.678 1617.0353,-778.9336 1629,-768 1652.436,-746.5838 1675.9351,-718.001 1689.4417,-700.7155"/>
-<polygon fill="#191970" stroke="#191970" points="1587.9808,-795.7333 1581.7195,-804.2801 1591.944,-801.5034 1587.9808,-795.7333"/>
+<path fill="none" stroke="#191970" d="M2228.4661,-813.4225C2288.4744,-806.5095 2369.1458,-792.8151 2389,-768 2407.9045,-744.3719 2375.0908,-717.1593 2349.0644,-700.6518"/>
+<polygon fill="#191970" stroke="#191970" points="2227.8269,-809.9719 2218.2747,-814.555 2228.6001,-816.9291 2227.8269,-809.9719"/>
</g>
-<!-- Node222 -->
+<!-- Node223 -->
<g id="node48" class="node">
-<title>Node222</title>
+<title>Node223</title>
<g id="a_node48"><a xlink:href="metadata__types_8h.html" target="_top" xlink:title="Defines types which can be used in metadata here which are also shared between C and C++ code bases...">
-<polygon fill="#ffffff" stroke="#ff0000" points="1681,-737.5 1681,-767.5 1797,-767.5 1797,-737.5 1681,-737.5"/>
-<text text-anchor="start" x="1689" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="1739" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/metadata_types.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="2264,-737.5 2264,-767.5 2380,-767.5 2380,-737.5 2264,-737.5"/>
+<text text-anchor="start" x="2272" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="2322" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/metadata_types.h</text>
</a>
</g>
</g>
-<!-- Node4->Node222 -->
-<g id="edge97" class="edge">
-<title>Node4->Node222</title>
-<path fill="none" stroke="#191970" d="M1608.3423,-800.865C1636.6747,-790.3773 1671.5938,-777.4515 1698.2029,-767.6017"/>
-<polygon fill="#191970" stroke="#191970" points="1606.783,-797.71 1598.6199,-804.4639 1609.2131,-804.2747 1606.783,-797.71"/>
+<!-- Node4->Node223 -->
+<g id="edge98" class="edge">
+<title>Node4->Node223</title>
+<path fill="none" stroke="#191970" d="M2205.7235,-800.5897C2230.9614,-790.1518 2261.8776,-777.3654 2285.4855,-767.6017"/>
+<polygon fill="#191970" stroke="#191970" points="2204.2592,-797.4077 2196.356,-804.4639 2206.9345,-803.8763 2204.2592,-797.4077"/>
</g>
-<!-- Node224 -->
+<!-- Node225 -->
<g id="node49" class="node">
-<title>Node224</title>
+<title>Node225</title>
<g id="a_node49"><a xlink:href="object_8h.html" target="_top" xlink:title="A managed object in the TVM runtime. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="860,-737.5 860,-767.5 976,-767.5 976,-737.5 860,-737.5"/>
-<text text-anchor="start" x="868" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
-<text text-anchor="middle" x="918" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/object.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1692,-737.5 1692,-767.5 1808,-767.5 1808,-737.5 1692,-737.5"/>
+<text text-anchor="start" x="1700" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/runtime</text>
+<text text-anchor="middle" x="1750" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/object.h</text>
</a>
</g>
</g>
-<!-- Node4->Node224 -->
-<g id="edge101" class="edge">
-<title>Node4->Node224</title>
-<path fill="none" stroke="#191970" d="M1489.8393,-812.3644C1363.4117,-799.129 1094.6799,-770.9962 976.1898,-758.5917"/>
-<polygon fill="#191970" stroke="#191970" points="1489.4818,-815.846 1499.7919,-813.4063 1490.2107,-808.8841 1489.4818,-815.846"/>
+<!-- Node4->Node225 -->
+<g id="edge102" class="edge">
+<title>Node4->Node225</title>
+<path fill="none" stroke="#191970" d="M2091.7972,-808.3547C2012.4631,-795.3903 1882.849,-774.2095 1808.1065,-761.9954"/>
+<polygon fill="#191970" stroke="#191970" points="2091.2588,-811.813 2101.6924,-809.9717 2092.3878,-804.9047 2091.2588,-811.813"/>
</g>
-<!-- Node238 -->
+<!-- Node239 -->
<g id="node50" class="node">
-<title>Node238</title>
+<title>Node239</title>
<g id="a_node50"><a xlink:href="parallel__for_8h.html" target="_top" xlink:title="An implementation to run loop in parallel. ">
-<polygon fill="#ffffff" stroke="#000000" points="2569,-737.5 2569,-767.5 2683,-767.5 2683,-737.5 2569,-737.5"/>
-<text text-anchor="start" x="2577" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/support</text>
-<text text-anchor="middle" x="2626" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/parallel_for.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="3190,-737.5 3190,-767.5 3304,-767.5 3304,-737.5 3190,-737.5"/>
+<text text-anchor="start" x="3198" y="-755.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/support</text>
+<text text-anchor="middle" x="3247" y="-744.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/parallel_for.h</text>
</a>
</g>
</g>
-<!-- Node4->Node238 -->
-<g id="edge120" class="edge">
-<title>Node4->Node238</title>
-<path fill="none" stroke="#191970" d="M1626.2523,-817.5884C1821.3569,-811.8702 2379.0302,-793.6891 2560,-768 2562.9133,-767.5864 2565.887,-767.1058 2568.8816,-766.5752"/>
-<polygon fill="#191970" stroke="#191970" points="1626.0258,-814.0934 1616.132,-817.8833 1626.2297,-821.0904 1626.0258,-814.0934"/>
+<!-- Node4->Node239 -->
+<g id="edge121" class="edge">
+<title>Node4->Node239</title>
+<path fill="none" stroke="#191970" d="M2228.5027,-817.6566C2426.3533,-812.0765 2996.2714,-794.1328 3181,-768 3183.9135,-767.5878 3186.8874,-767.1084 3189.8821,-766.5786"/>
+<polygon fill="#191970" stroke="#191970" points="2228.1421,-814.1652 2218.2442,-817.9442 2228.3384,-821.1625 2228.1421,-814.1652"/>
</g>
<!-- Node6 -->
<g id="node3" class="node">
<title>Node6</title>
<g id="a_node3"><a xlink:href="auto__scheduler_2cost__model_8h.html" target="_top" xlink:title="Cost models that estimate the performance of programs. ">
-<polygon fill="#ffffff" stroke="#000000" points="1718,-397 1718,-427 1870,-427 1870,-397 1718,-397"/>
-<text text-anchor="start" x="1726" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
-<text text-anchor="middle" x="1794" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/cost_model.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="398,-397 398,-427 550,-427 550,-397 398,-397"/>
+<text text-anchor="start" x="406" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
+<text text-anchor="middle" x="474" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/cost_model.h</text>
</a>
</g>
</g>
<!-- Node5->Node6 -->
<g id="edge2" class="edge">
<title>Node5->Node6</title>
-<path fill="none" stroke="#191970" d="M1829.3918,-461.3443C1821.3269,-450.1 1811.8693,-436.9139 1804.7606,-427.0028"/>
-<polygon fill="#191970" stroke="#191970" points="1826.5532,-463.3919 1835.2256,-469.478 1832.2413,-459.3121 1826.5532,-463.3919"/>
+<path fill="none" stroke="#191970" d="M198.5727,-467.468C259.5205,-455.1938 341.0016,-438.7844 399.5033,-427.0028"/>
+<polygon fill="#191970" stroke="#191970" points="197.7044,-464.0725 188.5922,-469.478 199.0864,-470.9347 197.7044,-464.0725"/>
</g>
<!-- Node7 -->
<g id="node4" class="node">
<title>Node7</title>
<g id="a_node4"><a xlink:href="auto__scheduler_2feature_8h.html" target="_top" xlink:title="Feature extraction for the cost model. We extract one feature vector per BufferStoreNode statement in...">
-<polygon fill="#ffffff" stroke="#000000" points="1926,-397 1926,-427 2078,-427 2078,-397 1926,-397"/>
-<text text-anchor="start" x="1934" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
-<text text-anchor="middle" x="2002" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/feature.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="0,-397 0,-427 152,-427 152,-397 0,-397"/>
+<text text-anchor="start" x="8" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
+<text text-anchor="middle" x="76" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/feature.h</text>
</a>
</g>
</g>
<!-- Node5->Node7 -->
<g id="edge3" class="edge">
<title>Node5->Node7</title>
-<path fill="none" stroke="#191970" d="M1887.4772,-465.2237C1913.2457,-453.248 1945.8832,-438.0799 1969.7181,-427.0028"/>
-<polygon fill="#191970" stroke="#191970" points="1885.9167,-462.0894 1878.3233,-469.478 1888.8669,-468.4373 1885.9167,-462.0894"/>
+<path fill="none" stroke="#191970" d="M101.3406,-460.3473C95.5646,-449.3272 88.9055,-436.6224 83.8635,-427.0028"/>
+<polygon fill="#191970" stroke="#191970" points="98.384,-462.2457 106.1264,-469.478 104.584,-458.996 98.384,-462.2457"/>
</g>
<!-- Node8 -->
<g id="node5" class="node">
<title>Node8</title>
<g id="a_node5"><a xlink:href="search__task_8h.html" target="_top" xlink:title="Meta information and hardware parameters for a search task. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="1466,-134.5 1466,-164.5 1618,-164.5 1618,-134.5 1466,-134.5"/>
-<text text-anchor="start" x="1474" y="-152.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
-<text text-anchor="middle" x="1542" y="-141.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/search_task.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="222,-134.5 222,-164.5 374,-164.5 374,-134.5 222,-134.5"/>
+<text text-anchor="start" x="230" y="-152.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
+<text text-anchor="middle" x="298" y="-141.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/search_task.h</text>
</a>
</g>
</g>
<!-- Node5->Node8 -->
<g id="edge4" class="edge">
<title>Node5->Node8</title>
-<path fill="none" stroke="#191970" d="M1864.5768,-461.1698C1877.5896,-441.8047 1890.7311,-413.8391 1879,-391 1816.104,-268.5492 1659.099,-194.2726 1583.7003,-164.5326"/>
-<polygon fill="#191970" stroke="#191970" points="1861.6547,-459.2392 1858.6968,-469.4128 1867.3534,-463.3044 1861.6547,-459.2392"/>
+<path fill="none" stroke="#191970" d="M139.4697,-461.9485C147.4794,-453.6014 155.6264,-443.5757 161,-433 180.2088,-395.1953 180,-381.9049 180,-339.5 180,-339.5 180,-339.5 180,-278 180,-225.45 234.2002,-185.1068 269.134,-164.5619"/>
+<polygon fill="#191970" stroke="#191970" points="136.8802,-459.5871 132.2361,-469.1098 141.805,-464.5616 136.8802,-459.5871"/>
</g>
<!-- Node12 -->
<g id="node6" class="node">
<title>Node12</title>
<g id="a_node6"><a xlink:href="search__policy_8h.html" target="_top" xlink:title="The base class of search policies, including the abstract definition of search policy and other suppo...">
-<polygon fill="#ffffff" stroke="#ff0000" points="1406,-67.5 1406,-97.5 1558,-97.5 1558,-67.5 1406,-67.5"/>
-<text text-anchor="start" x="1414" y="-85.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
-<text text-anchor="middle" x="1482" y="-74.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/search_policy.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="278,-67.5 278,-97.5 430,-97.5 430,-67.5 278,-67.5"/>
+<text text-anchor="start" x="286" y="-85.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
+<text text-anchor="middle" x="354" y="-74.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/search_policy.h</text>
</a>
</g>
</g>
<!-- Node8->Node12 -->
<g id="edge5" class="edge">
<title>Node8->Node12</title>
-<path fill="none" stroke="#191970" d="M1521.527,-126.6385C1512.954,-117.0653 1503.2227,-106.1987 1495.561,-97.6432"/>
-<polygon fill="#191970" stroke="#191970" points="1519.1065,-129.1821 1528.3851,-134.2967 1524.3212,-124.5122 1519.1065,-129.1821"/>
+<path fill="none" stroke="#191970" d="M317.3485,-126.3509C325.2911,-116.8482 334.2636,-106.1132 341.343,-97.6432"/>
+<polygon fill="#191970" stroke="#191970" points="314.4349,-124.3792 310.7073,-134.2967 319.8059,-128.8684 314.4349,-124.3792"/>
</g>
<!-- Node13->Node6 -->
<g id="edge7" class="edge">
<title>Node13->Node6</title>
-<path fill="none" stroke="#191970" d="M978.8388,-472.9801C987.2172,-471.4317 995.8238,-470.0263 1004,-469 1314.8584,-429.9807 1398.0035,-478.3608 1708,-433 1718.0368,-431.5313 1728.6146,-429.4242 1738.7873,-427.1053"/>
-<polygon fill="#191970" stroke="#191970" points="977.9312,-469.5907 968.7766,-474.9239 979.2589,-476.4636 977.9312,-469.5907"/>
+<path fill="none" stroke="#191970" d="M948.7303,-473.1505C940.1265,-471.6355 931.3437,-470.1952 923,-469 764.2646,-446.2611 722.1887,-459.2737 564,-433 554.1867,-431.3701 543.8478,-429.2979 533.8135,-427.0933"/>
+<polygon fill="#191970" stroke="#191970" points="948.2482,-476.6199 958.7111,-474.953 949.4923,-469.7314 948.2482,-476.6199"/>
</g>
<!-- Node13->Node12 -->
<g id="edge8" class="edge">
<title>Node13->Node12</title>
-<path fill="none" stroke="#191970" d="M981.1941,-473.0013C988.8333,-471.5527 996.6017,-470.1707 1004,-469 1141.435,-447.2516 1188.3835,-490.703 1315,-433 1340.1916,-421.5194 1339.5924,-408.5537 1361,-391 1381.8967,-373.8652 1394.8387,-377.3697 1410,-355 1429.8372,-325.7313 1430,-313.3578 1430,-278 1430,-278 1430,-278 1430,-216.5 1430,-170.8339 1456.5721,-121.9818 1471.7265,-97.8342"/>
-<polygon fill="#191970" stroke="#191970" points="980.4159,-469.587 971.2681,-474.9319 981.7524,-476.4582 980.4159,-469.587"/>
+<path fill="none" stroke="#191970" d="M927.2212,-476.2597C811.1867,-463.8987 606.404,-441.1384 592,-433 571.3316,-421.3221 578.5377,-404.4845 559,-391 514.922,-360.5783 481.3548,-393.3792 444,-355 372.9304,-281.9813 421.697,-228.2611 383,-134 377.7782,-121.2803 370.0773,-107.7662 363.9234,-97.7744"/>
+<polygon fill="#191970" stroke="#191970" points="927.1632,-479.7732 937.4772,-477.3498 927.9032,-472.8124 927.1632,-479.7732"/>
</g>
<!-- Node14 -->
<g id="node8" class="node">
<title>Node14</title>
<g id="a_node8"><a xlink:href="transform__step_8h.html" target="_top" xlink:title="Transformation steps. These steps are used to manipulate LoopState. They are similar to the schedule ...">
-<polygon fill="#ffffff" stroke="#ff0000" points="622,-397 622,-427 774,-427 774,-397 622,-397"/>
-<text text-anchor="start" x="630" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
-<text text-anchor="middle" x="698" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/transform_step.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="948,-397 948,-427 1100,-427 1100,-397 948,-397"/>
+<text text-anchor="start" x="956" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/auto_scheduler</text>
+<text text-anchor="middle" x="1024" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/transform_step.h</text>
</a>
</g>
</g>
<!-- Node13->Node14 -->
<g id="edge9" class="edge">
<title>Node13->Node14</title>
-<path fill="none" stroke="#191970" d="M887.0556,-471.8538C847.6954,-459.3926 787.5397,-440.3477 745.6934,-427.0995"/>
-<polygon fill="#191970" stroke="#191970" points="886.0005,-475.191 896.5906,-474.8726 888.1134,-468.5174 886.0005,-475.191"/>
+<path fill="none" stroke="#191970" d="M1010.8707,-464.8817C1013.8157,-453.02 1017.5138,-438.1248 1020.2341,-427.168"/>
+<polygon fill="#191970" stroke="#191970" points="1007.4404,-464.1736 1008.4276,-474.7223 1014.2341,-465.8604 1007.4404,-464.1736"/>
</g>
<!-- Node16 -->
<g id="node9" class="node">
<title>Node16</title>
<g id="a_node9"><a xlink:href="ir_2adt_8h.html" target="_top" xlink:title="Algebraic data type definitions. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="415,-207 415,-226 527,-226 527,-207 415,-207"/>
-<text text-anchor="middle" x="471" y="-214" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/adt.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1531,-207 1531,-226 1643,-226 1643,-207 1531,-207"/>
+<text text-anchor="middle" x="1587" y="-214" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/adt.h</text>
</a>
</g>
</g>
<!-- Node13->Node16 -->
<g id="edge10" class="edge">
<title>Node13->Node16</title>
-<path fill="none" stroke="#191970" d="M848.4287,-478.4497C742.6931,-469.6356 565.1108,-452.2082 542,-433 525.1164,-418.9674 484.1223,-266.4615 473.531,-226.1997"/>
-<polygon fill="#191970" stroke="#191970" points="848.1752,-481.9405 858.4293,-479.2755 848.7513,-474.9643 848.1752,-481.9405"/>
+<path fill="none" stroke="#191970" d="M1084.969,-482.3672C1232.9254,-477.6539 1540.3281,-464.195 1577,-433 1639.1926,-380.0958 1603.1345,-261.3209 1590.7047,-226.3595"/>
+<polygon fill="#191970" stroke="#191970" points="1084.4832,-478.8806 1074.5976,-482.6916 1084.7021,-485.8772 1084.4832,-478.8806"/>
</g>
<!-- Node140 -->
<g id="node10" class="node">
<title>Node140</title>
<g id="a_node10"><a xlink:href="ir_2expr_8h.html" target="_top" xlink:title="Base expr nodes in TVM. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="592.5,-268.5 592.5,-287.5 709.5,-287.5 709.5,-268.5 592.5,-268.5"/>
-<text text-anchor="middle" x="651" y="-275.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/expr.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1267.5,-268.5 1267.5,-287.5 1384.5,-287.5 1384.5,-268.5 1267.5,-268.5"/>
+<text text-anchor="middle" x="1326" y="-275.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/expr.h</text>
</a>
</g>
</g>
<!-- Node13->Node140 -->
<g id="edge11" class="edge">
<title>Node13->Node140</title>
-<path fill="none" stroke="#191970" d="M933.7839,-465.0452C939.5487,-444.7585 944.7714,-412.9196 930,-391 880.8631,-318.0845 775.8012,-292.203 709.5496,-283.0272"/>
-<polygon fill="#191970" stroke="#191970" points="930.3966,-464.1498 930.7361,-474.7392 937.0743,-466.2493 930.3966,-464.1498"/>
+<path fill="none" stroke="#191970" d="M1066.3794,-473.0989C1073.9909,-471.6986 1081.6745,-470.303 1089,-469 1136.2628,-460.5935 1268.2908,-467.178 1302,-433 1341.9636,-392.4805 1332.6706,-314.8113 1327.9113,-287.6579"/>
+<polygon fill="#191970" stroke="#191970" points="1065.6475,-469.6748 1056.4508,-474.9351 1066.9205,-476.5581 1065.6475,-469.6748"/>
</g>
<!-- Node148 -->
<g id="node11" class="node">
<title>Node148</title>
<g id="a_node11"><a xlink:href="script_2ir__builder_2base_8h.html" target="_top" xlink:title="include/tvm/script\l/ir_builder/base.h">
-<polygon fill="#ffffff" stroke="#ff0000" points="1020,-201.5 1020,-231.5 1124,-231.5 1124,-201.5 1020,-201.5"/>
-<text text-anchor="start" x="1028" y="-219.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="middle" x="1072" y="-208.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/ir_builder/base.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1343,-201.5 1343,-231.5 1447,-231.5 1447,-201.5 1343,-201.5"/>
+<text text-anchor="start" x="1351" y="-219.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="middle" x="1395" y="-208.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/ir_builder/base.h</text>
</a>
</g>
</g>
<!-- Node13->Node148 -->
<g id="edge41" class="edge">
<title>Node13->Node148</title>
-<path fill="none" stroke="#191970" d="M975.0117,-472.4029C1020.5166,-460.4557 1082.9524,-442.6177 1090,-433 1135.5894,-370.7852 1096.8563,-269.8139 1079.4273,-231.7066"/>
-<polygon fill="#191970" stroke="#191970" points="974.0925,-469.0255 965.2969,-474.9321 975.8562,-475.7997 974.0925,-469.0255"/>
+<path fill="none" stroke="#191970" d="M1065.4681,-473.1334C1073.3732,-471.699 1081.3774,-470.2854 1089,-469 1194.9249,-451.1375 1242.4479,-497.9616 1328,-433 1393.0215,-383.6278 1396.6551,-272.055 1395.6982,-231.5515"/>
+<polygon fill="#191970" stroke="#191970" points="1064.7489,-469.7068 1055.5442,-474.9533 1066.0116,-476.592 1064.7489,-469.7068"/>
</g>
<!-- Node149 -->
<g id="node12" class="node">
<title>Node149</title>
<g id="a_node12"><a xlink:href="ir__builder_2ir_2frame_8h.html" target="_top" xlink:title="include/tvm/script\l/ir_builder/ir/frame.h">
-<polygon fill="#ffffff" stroke="#ff0000" points="1014,-134.5 1014,-164.5 1130,-164.5 1130,-134.5 1014,-134.5"/>
-<text text-anchor="start" x="1022" y="-152.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="middle" x="1072" y="-141.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/ir_builder/ir/frame.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1321,-134.5 1321,-164.5 1437,-164.5 1437,-134.5 1321,-134.5"/>
+<text text-anchor="start" x="1329" y="-152.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="middle" x="1379" y="-141.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/ir_builder/ir/frame.h</text>
</a>
</g>
</g>
<!-- Node13->Node149 -->
<g id="edge42" class="edge">
<title>Node13->Node149</title>
-<path fill="none" stroke="#191970" d="M979.0777,-472.5473C1030.4601,-460.3638 1102.7824,-442.0136 1113,-433 1146.7652,-403.2137 1152,-384.5256 1152,-339.5 1152,-339.5 1152,-339.5 1152,-278 1152,-242.7513 1151.5975,-230.9433 1133,-201 1123.8593,-186.2828 1109.1015,-173.6593 1096.3914,-164.5937"/>
-<polygon fill="#191970" stroke="#191970" points="978.0357,-469.1969 969.105,-474.8972 979.6412,-476.0103 978.0357,-469.1969"/>
+<path fill="none" stroke="#191970" d="M1085.0846,-481.5435C1175.8215,-476.7022 1317.5684,-464.2534 1360,-433 1404.9242,-399.9107 1482.2905,-250.2128 1456,-201 1447.3611,-184.8289 1431.1468,-172.8476 1415.8172,-164.5313"/>
+<polygon fill="#191970" stroke="#191970" points="1084.6369,-478.0618 1074.8296,-482.07 1084.9959,-485.0526 1084.6369,-478.0618"/>
</g>
<!-- Node150 -->
<g id="node13" class="node">
<title>Node150</title>
<g id="a_node13"><a xlink:href="ir_2ir_8h.html" target="_top" xlink:title="include/tvm/script\l/ir_builder/ir/ir.h">
-<polygon fill="#ffffff" stroke="#000000" points="1020,-67.5 1020,-97.5 1124,-97.5 1124,-67.5 1020,-67.5"/>
-<text text-anchor="start" x="1028" y="-85.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="middle" x="1072" y="-74.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/ir_builder/ir/ir.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1225,-67.5 1225,-97.5 1329,-97.5 1329,-67.5 1225,-67.5"/>
+<text text-anchor="start" x="1233" y="-85.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="middle" x="1277" y="-74.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/ir_builder/ir/ir.h</text>
</a>
</g>
</g>
<!-- Node13->Node150 -->
<g id="edge43" class="edge">
<title>Node13->Node150</title>
-<path fill="none" stroke="#191970" d="M981.5395,-472.7708C1043.6872,-459.2247 1137.5574,-438.1389 1144,-433 1180.2055,-404.121 1190,-385.8124 1190,-339.5 1190,-339.5 1190,-339.5 1190,-216.5 1190,-161.9431 1134.2266,-119.066 1099.3682,-97.5956"/>
-<polygon fill="#191970" stroke="#191970" points="980.5037,-469.4141 971.4756,-474.959 981.991,-476.2543 980.5037,-469.4141"/>
+<path fill="none" stroke="#191970" d="M1084.895,-478.0434C1147.4941,-471.1346 1227.8748,-457.6662 1249,-433 1296.7731,-377.2191 1252.2568,-341.2174 1258,-268 1262.962,-204.7425 1271.4634,-129.6114 1275.1758,-97.8685"/>
+<polygon fill="#191970" stroke="#191970" points="1084.1924,-474.5983 1074.619,-479.137 1084.9332,-481.559 1084.1924,-474.5983"/>
</g>
<!-- Node169 -->
<g id="node14" class="node">
<title>Node169</title>
<g id="a_node14"><a xlink:href="doc_8h.html" target="_top" xlink:title="include/tvm/script\l/printer/doc.h">
-<polygon fill="#ffffff" stroke="#ff0000" points="681,-201.5 681,-231.5 785,-231.5 785,-201.5 681,-201.5"/>
-<text text-anchor="start" x="689" y="-219.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="middle" x="733" y="-208.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/doc.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="674,-201.5 674,-231.5 778,-231.5 778,-201.5 674,-201.5"/>
+<text text-anchor="start" x="682" y="-219.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="middle" x="726" y="-208.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/doc.h</text>
</a>
</g>
</g>
<!-- Node13->Node169 -->
<g id="edge44" class="edge">
<title>Node13->Node169</title>
-<path fill="none" stroke="#191970" d="M937.7007,-466.0571C947.3794,-446.8669 958.7386,-416.1623 949,-391 917.4182,-309.4003 822.0441,-255.7924 769.208,-231.5017"/>
-<polygon fill="#191970" stroke="#191970" points="934.5706,-464.488 932.9133,-474.9524 940.7345,-467.8055 934.5706,-464.488"/>
+<path fill="none" stroke="#191970" d="M927.2875,-473.8704C874.6741,-465.2491 811.6315,-451.4924 792,-433 733.0782,-377.4969 726.2472,-271.1306 725.8266,-231.6985"/>
+<polygon fill="#191970" stroke="#191970" points="926.8888,-477.3511 937.3158,-475.4732 927.9936,-470.4388 926.8888,-477.3511"/>
</g>
<!-- Node171 -->
<g id="node15" class="node">
<title>Node171</title>
<g id="a_node15"><a xlink:href="printer_2frame_8h.html" target="_top" xlink:title="include/tvm/script\l/printer/frame.h">
-<polygon fill="#ffffff" stroke="#000000" points="784,-134.5 784,-164.5 888,-164.5 888,-134.5 784,-134.5"/>
-<text text-anchor="start" x="792" y="-152.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="middle" x="836" y="-141.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/frame.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="712,-134.5 712,-164.5 816,-164.5 816,-134.5 712,-134.5"/>
+<text text-anchor="start" x="720" y="-152.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="middle" x="764" y="-141.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/frame.h</text>
</a>
</g>
</g>
<!-- Node13->Node171 -->
<g id="edge45" class="edge">
<title>Node13->Node171</title>
-<path fill="none" stroke="#191970" d="M943.8124,-467.5718C951.9052,-458.2178 960.7564,-445.8984 965,-433 970.8337,-415.2683 969.1244,-409.2053 965,-391 943.7848,-297.355 875.7319,-201.1309 847.9661,-164.703"/>
-<polygon fill="#191970" stroke="#191970" points="941.2087,-465.2325 937.0305,-474.9687 946.3683,-469.9631 941.2087,-465.2325"/>
+<path fill="none" stroke="#191970" d="M942.9553,-472.8853C899.1361,-463.5809 845.8621,-449.401 830,-433 800.1741,-402.1607 806,-382.4027 806,-339.5 806,-339.5 806,-339.5 806,-278 806,-235.2838 784.7719,-188.3567 772.4993,-164.7962"/>
+<polygon fill="#191970" stroke="#191970" points="942.4984,-476.3649 953.0009,-474.9695 943.9205,-469.5109 942.4984,-476.3649"/>
</g>
<!-- Node172 -->
<g id="node16" class="node">
<title>Node172</title>
<g id="a_node16"><a xlink:href="ir__docsifier_8h.html" target="_top" xlink:title="include/tvm/script\l/printer/ir_docsifier.h">
-<polygon fill="#ffffff" stroke="#000000" points="698,-.5 698,-30.5 814,-30.5 814,-.5 698,-.5"/>
-<text text-anchor="start" x="706" y="-18.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="middle" x="756" y="-7.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/ir_docsifier.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="706,-.5 706,-30.5 822,-30.5 822,-.5 706,-.5"/>
+<text text-anchor="start" x="714" y="-18.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="middle" x="764" y="-7.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/ir_docsifier.h</text>
</a>
</g>
</g>
<!-- Node13->Node172 -->
<g id="edge46" class="edge">
<title>Node13->Node172</title>
-<path fill="none" stroke="#191970" d="M848.2999,-481.9234C684.6766,-476.0514 318.8064,-460.0523 269,-433 170.1066,-379.2862 151.1357,-338.58 115,-232 83.0811,-137.8572 171.3487,-97.3779 266,-67 345.0027,-41.6444 585.664,-25.143 697.6005,-18.6373"/>
-<polygon fill="#191970" stroke="#191970" points="848.1886,-485.4216 858.3066,-482.2789 848.4372,-478.426 848.1886,-485.4216"/>
+<path fill="none" stroke="#191970" d="M969.8488,-470.9465C922.104,-450.4224 844,-406.5942 844,-339.5 844,-339.5 844,-339.5 844,-149.5 844,-100.459 804.0085,-53.7794 780.5761,-30.6372"/>
+<polygon fill="#191970" stroke="#191970" points="968.6717,-474.2475 979.249,-474.8566 971.3601,-467.7844 968.6717,-474.2475"/>
</g>
<!-- Node173 -->
<g id="node17" class="node">
<title>Node173</title>
<g id="a_node17"><a xlink:href="var__table_8h.html" target="_top" xlink:title="include/tvm/script\l/printer/var_table.h">
-<polygon fill="#ffffff" stroke="#000000" points="784,-67.5 784,-97.5 894,-97.5 894,-67.5 784,-67.5"/>
-<text text-anchor="start" x="792" y="-85.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="middle" x="839" y="-74.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/var_table.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="546,-67.5 546,-97.5 656,-97.5 656,-67.5 546,-67.5"/>
+<text text-anchor="start" x="554" y="-85.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="middle" x="601" y="-74.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/var_table.h</text>
</a>
</g>
</g>
<!-- Node13->Node173 -->
<g id="edge49" class="edge">
<title>Node13->Node173</title>
-<path fill="none" stroke="#191970" d="M951.03,-468.6896C962.4032,-459.729 974.8237,-447.4251 981,-433 988.3472,-415.8401 983.4426,-409.5062 981,-391 971.4167,-318.3932 907.9245,-150.3334 897,-134 887.4173,-119.6727 872.9173,-106.7977 860.813,-97.5046"/>
-<polygon fill="#191970" stroke="#191970" points="948.6245,-466.115 942.6578,-474.87 952.7818,-471.7468 948.6245,-466.115"/>
+<path fill="none" stroke="#191970" d="M927.3829,-476.6333C824.9255,-465.9364 655.7272,-446.5821 630,-433 607.7388,-421.2477 613.3419,-404.3493 592,-391 543.347,-360.5678 503.3855,-400.1786 468,-355 407.3134,-277.518 417.1525,-208.0346 482,-134 498.5383,-115.1186 523.2447,-102.9151 545.8666,-95.1542"/>
+<polygon fill="#191970" stroke="#191970" points="927.0804,-480.1207 937.3886,-477.6726 927.8037,-473.1581 927.0804,-480.1207"/>
</g>
<!-- Node26 -->
<g id="node18" class="node">
<title>Node26</title>
<g id="a_node18"><a xlink:href="target_8h.html" target="_top" xlink:title="Compilation target object. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="1256.5,-201.5 1256.5,-231.5 1363.5,-231.5 1363.5,-201.5 1256.5,-201.5"/>
-<text text-anchor="start" x="1264.5" y="-219.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/target</text>
-<text text-anchor="middle" x="1310" y="-208.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/target.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="472.5,-201.5 472.5,-231.5 579.5,-231.5 579.5,-201.5 472.5,-201.5"/>
+<text text-anchor="start" x="480.5" y="-219.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/target</text>
+<text text-anchor="middle" x="526" y="-208.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/target.h</text>
</a>
</g>
</g>
<!-- Node13->Node26 -->
<g id="edge52" class="edge">
<title>Node13->Node26</title>
-<path fill="none" stroke="#191970" d="M982.3519,-473.0227C989.623,-471.6134 996.9831,-470.2362 1004,-469 1056.6793,-459.7194 1201.3754,-467.7977 1242,-433 1257.933,-419.3523 1294.4273,-278.5375 1306.2375,-231.6175"/>
-<polygon fill="#191970" stroke="#191970" points="981.6692,-469.5898 972.5327,-474.954 983.0202,-476.4582 981.6692,-469.5898"/>
+<path fill="none" stroke="#191970" d="M927.2684,-476.4433C854.1158,-467.9032 751.3151,-452.8989 716,-433 692.8,-419.9275 576.6738,-278.7307 538.3172,-231.6662"/>
+<polygon fill="#191970" stroke="#191970" points="927.0495,-479.9411 937.3837,-477.6053 927.8485,-472.9868 927.0495,-479.9411"/>
</g>
<!-- Node62 -->
<g id="node19" class="node">
<title>Node62</title>
<g id="a_node19"><a xlink:href="tag_8h.html" target="_top" xlink:title="Target tag registry. ">
-<polygon fill="#ffffff" stroke="#000000" points="1237.5,-134.5 1237.5,-164.5 1344.5,-164.5 1344.5,-134.5 1237.5,-134.5"/>
-<text text-anchor="start" x="1245.5" y="-152.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/target</text>
-<text text-anchor="middle" x="1291" y="-141.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/tag.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="491.5,-134.5 491.5,-164.5 598.5,-164.5 598.5,-134.5 491.5,-134.5"/>
+<text text-anchor="start" x="499.5" y="-152.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/target</text>
+<text text-anchor="middle" x="545" y="-141.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/tag.h</text>
</a>
</g>
</g>
<!-- Node13->Node62 -->
<g id="edge51" class="edge">
<title>Node13->Node62</title>
-<path fill="none" stroke="#191970" d="M983.7231,-472.9443C990.5514,-471.5939 997.4302,-470.2524 1004,-469 1092.7198,-452.0875 1137.8428,-494.4855 1204,-433 1235.4261,-403.7931 1228,-382.4027 1228,-339.5 1228,-339.5 1228,-339.5 1228,-278 1228,-242.7513 1230.896,-232.3549 1247,-201 1253.9845,-187.401 1265.2218,-174.333 1274.5916,-164.7635"/>
-<polygon fill="#191970" stroke="#191970" points="982.9214,-469.5351 973.797,-474.9199 984.2879,-476.4004 982.9214,-469.5351"/>
+<path fill="none" stroke="#191970" d="M927.2169,-474.3968C865.5058,-465.3451 785.6329,-450.7737 758,-433 704.2446,-398.4241 627.3908,-251.3339 588,-201 578.2372,-188.525 566.7116,-174.77 558.0516,-164.6158"/>
+<polygon fill="#191970" stroke="#191970" points="926.9451,-477.8937 937.3423,-475.8568 927.9442,-470.9654 926.9451,-477.8937"/>
</g>
<!-- Node13->Node176 -->
<g id="edge55" class="edge">
<title>Node13->Node176</title>
-<path fill="none" stroke="#191970" d="M848.1399,-481.5912C701.9631,-475.581 400.1796,-460.0432 361,-433 266.61,-367.8486 302.0133,-302.1768 248,-201 240.1057,-186.2125 230.1891,-169.5172 223.9709,-159.2503"/>
-<polygon fill="#191970" stroke="#191970" points="848.2561,-485.0987 858.39,-482.0075 848.5403,-478.1045 848.2561,-485.0987"/>
+<path fill="none" stroke="#191970" d="M972.2909,-470.1C957.0595,-461.6787 940.5824,-449.429 932,-433 915.3403,-401.1086 873.604,-420.6785 939,-201 943.764,-184.9967 954.6824,-168.8804 962.2012,-159.0435"/>
+<polygon fill="#191970" stroke="#191970" points="971.0381,-473.3914 981.5319,-474.8518 974.2393,-467.1662 971.0381,-473.3914"/>
</g>
<!-- Node177 -->
<g id="node21" class="node">
<title>Node177</title>
<g id="a_node21"><a xlink:href="relay_2base_8h.html" target="_top" xlink:title="Base classes for the Relay IR. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="275.5,-73 275.5,-92 412.5,-92 412.5,-73 275.5,-73"/>
-<text text-anchor="middle" x="344" y="-80" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/relay/base.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1056.5,-73 1056.5,-92 1193.5,-92 1193.5,-73 1056.5,-73"/>
+<text text-anchor="middle" x="1125" y="-80" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/relay/base.h</text>
</a>
</g>
</g>
<!-- Node13->Node177 -->
<g id="edge40" class="edge">
<title>Node13->Node177</title>
-<path fill="none" stroke="#191970" d="M848.3186,-479.0729C768.1556,-472.2451 650.4531,-458.2371 613,-433 572.6392,-405.8036 556,-388.1687 556,-339.5 556,-339.5 556,-339.5 556,-216.5 556,-141.4999 460.0214,-106.7619 397.2229,-92.06"/>
-<polygon fill="#191970" stroke="#191970" points="848.1335,-482.5695 858.3891,-479.9099 848.7134,-475.5935 848.1335,-482.5695"/>
+<path fill="none" stroke="#191970" d="M1084.8452,-476.8147C1142.1639,-469.4074 1212.7699,-455.8742 1231,-433 1242.634,-418.4022 1237.867,-408.3577 1231,-391 1218.334,-358.9843 1125,-312.4301 1125,-278 1125,-278 1125,-278 1125,-216.5 1125,-169.9722 1125,-114.357 1125,-92.2517"/>
+<polygon fill="#191970" stroke="#191970" points="1084.1683,-473.372 1074.6786,-478.0834 1085.0351,-480.3181 1084.1683,-473.372"/>
</g>
<!-- Node183 -->
<g id="node22" class="node">
<title>Node183</title>
<g id="a_node22"><a xlink:href="var_8h.html" target="_top" xlink:title="Variables in the TIR. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="124.5,-207 124.5,-226 239.5,-226 239.5,-207 124.5,-207"/>
-<text text-anchor="middle" x="182" y="-214" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/tir/var.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="948.5,-207 948.5,-226 1063.5,-226 1063.5,-207 948.5,-207"/>
+<text text-anchor="middle" x="1006" y="-214" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/tir/var.h</text>
</a>
</g>
</g>
<!-- Node13->Node183 -->
<g id="edge56" class="edge">
<title>Node13->Node183</title>
-<path fill="none" stroke="#191970" d="M848.1749,-480.8413C696.2451,-473.3358 373.7642,-455.0354 328,-433 238.5461,-389.9281 195.2955,-262.109 184.662,-226.0611"/>
-<polygon fill="#191970" stroke="#191970" points="848.2125,-484.3472 858.372,-481.3415 848.5555,-477.3557 848.2125,-484.3472"/>
+<path fill="none" stroke="#191970" d="M976.1551,-469.8367C962.1367,-461.2161 946.8181,-448.8702 939,-433 919.9174,-394.2635 982.815,-262.7893 1001.1386,-226.086"/>
+<polygon fill="#191970" stroke="#191970" points="974.6723,-473.0211 985.0876,-474.9632 978.1567,-466.9499 974.6723,-473.0211"/>
</g>
<!-- Node185 -->
<g id="node23" class="node">
<title>Node185</title>
<g id="a_node23"><a xlink:href="ir_2span_8h.html" target="_top" xlink:title="Span information for debugging purposes. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="370.5,-402.5 370.5,-421.5 489.5,-421.5 489.5,-402.5 370.5,-402.5"/>
-<text text-anchor="middle" x="430" y="-409.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/span.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1430.5,-402.5 1430.5,-421.5 1549.5,-421.5 1549.5,-402.5 1430.5,-402.5"/>
+<text text-anchor="middle" x="1490" y="-409.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/span.h</text>
</a>
</g>
</g>
<!-- Node13->Node185 -->
<g id="edge32" class="edge">
<title>Node13->Node185</title>
-<path fill="none" stroke="#191970" d="M848.2057,-478.5909C762.8112,-471.3325 622.8203,-456.9418 504,-433 489.2067,-430.0192 473.055,-425.5783 459.6264,-421.547"/>
-<polygon fill="#191970" stroke="#191970" points="848.0864,-482.0931 858.3439,-479.4409 848.6713,-475.1176 848.0864,-482.0931"/>
+<path fill="none" stroke="#191970" d="M1063.603,-473.1663C1072.1021,-471.6617 1080.7665,-470.2203 1089,-469 1233.6316,-447.5647 1272.5056,-461.0557 1416,-433 1430.9413,-430.0787 1447.2499,-425.5948 1460.7452,-421.5193"/>
+<polygon fill="#191970" stroke="#191970" points="1062.9543,-469.7268 1053.7376,-474.952 1064.2011,-476.6148 1062.9543,-469.7268"/>
</g>
<!-- Node186 -->
<g id="node24" class="node">
<title>Node186</title>
<g id="a_node24"><a xlink:href="ir_2type_8h.html" target="_top" xlink:title="IR/AST nodes for the unified type system in TVM. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="372,-330 372,-349 490,-349 490,-330 372,-330"/>
-<text text-anchor="middle" x="431" y="-337" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/type.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1470,-330 1470,-349 1588,-349 1588,-330 1470,-330"/>
+<text text-anchor="middle" x="1529" y="-337" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/type.h</text>
</a>
</g>
</g>
<!-- Node13->Node186 -->
<g id="edge38" class="edge">
<title>Node13->Node186</title>
-<path fill="none" stroke="#191970" d="M928.4439,-464.7766C928.8345,-443.2285 925.7458,-409.3415 905,-391 874.4579,-363.9974 610.2294,-348.0563 490.0589,-342.1427"/>
-<polygon fill="#191970" stroke="#191970" points="924.9472,-464.6263 927.9976,-474.7725 931.9403,-464.9386 924.9472,-464.6263"/>
+<path fill="none" stroke="#191970" d="M1084.8309,-478.5144C1235.4775,-466.8915 1550.119,-441.6505 1558,-433 1580.2049,-408.627 1552.1014,-367.4992 1537.1863,-349.0075"/>
+<polygon fill="#191970" stroke="#191970" points="1084.4138,-475.036 1074.7121,-479.2936 1084.9514,-482.0154 1084.4138,-475.036"/>
</g>
<!-- Node129 -->
<g id="node25" class="node">
<title>Node129</title>
<g id="a_node25"><a xlink:href="arg__info_8h.html" target="_top" xlink:title="include/tvm/meta_schedule\l/arg_info.h">
-<polygon fill="#ffffff" stroke="#ff0000" points="1370,-397 1370,-427 1522,-427 1522,-397 1370,-397"/>
-<text text-anchor="start" x="1378" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/meta_schedule</text>
-<text text-anchor="middle" x="1446" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/arg_info.h</text>
+<polygon fill="#ffffff" stroke="#ff0000" points="1860,-397 1860,-427 2012,-427 2012,-397 1860,-397"/>
+<text text-anchor="start" x="1868" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/meta_schedule</text>
+<text text-anchor="middle" x="1936" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/arg_info.h</text>
</a>
</g>
</g>
<!-- Node13->Node129 -->
<g id="edge39" class="edge">
<title>Node13->Node129</title>
-<path fill="none" stroke="#191970" d="M980.3361,-473.0689C988.2481,-471.5753 996.3213,-470.1627 1004,-469 1159.488,-445.4556 1200.8902,-458.9196 1356,-433 1365.8117,-431.3604 1376.1497,-429.2833 1386.1837,-427.0766"/>
-<polygon fill="#191970" stroke="#191970" points="979.5908,-469.6482 970.4421,-474.9916 980.9262,-476.5196 979.5908,-469.6482"/>
+<path fill="none" stroke="#191970" d="M1085.0121,-481.4841C1265.5335,-474.3006 1703.8794,-455.117 1850,-433 1860.0295,-431.4819 1870.6037,-429.3508 1880.7751,-427.0236"/>
+<polygon fill="#191970" stroke="#191970" points="1084.648,-477.9957 1074.7944,-481.8888 1084.9251,-484.9902 1084.648,-477.9957"/>
</g>
<!-- Node188 -->
<g id="node26" class="node">
<title>Node188</title>
<g id="a_node26"><a xlink:href="traced__object__functor_8h.html" target="_top" xlink:title="include/tvm/script\l/printer/traced_object\l_functor.h">
-<polygon fill="#ffffff" stroke="#000000" points="1578.5,-391.5 1578.5,-432.5 1699.5,-432.5 1699.5,-391.5 1578.5,-391.5"/>
-<text text-anchor="start" x="1586.5" y="-420.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="start" x="1586.5" y="-409.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/traced_object</text>
-<text text-anchor="middle" x="1639" y="-398.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_functor.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1720.5,-391.5 1720.5,-432.5 1841.5,-432.5 1841.5,-391.5 1720.5,-391.5"/>
+<text text-anchor="start" x="1728.5" y="-420.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="start" x="1728.5" y="-409.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer/traced_object</text>
+<text text-anchor="middle" x="1781" y="-398.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_functor.h</text>
</a>
</g>
</g>
<!-- Node13->Node188 -->
<g id="edge47" class="edge">
<title>Node13->Node188</title>
-<path fill="none" stroke="#191970" d="M979.1504,-473.0474C987.4334,-471.5047 995.929,-470.0833 1004,-469 1236.6815,-437.769 1298.2168,-463.4634 1531,-433 1546.4358,-430.98 1562.9895,-428.1177 1578.3847,-425.1543"/>
-<polygon fill="#191970" stroke="#191970" points="978.3477,-469.6377 969.1954,-474.9749 979.6784,-476.5101 978.3477,-469.6377"/>
+<path fill="none" stroke="#191970" d="M1084.9878,-479.4008C1203.5465,-471.426 1434.4715,-454.6479 1630,-433 1659.8583,-429.6942 1692.9223,-425.1918 1720.4804,-421.2037"/>
+<polygon fill="#191970" stroke="#191970" points="1084.6551,-475.9152 1074.9113,-480.0756 1085.1229,-482.8995 1084.6551,-475.9152"/>
</g>
<!-- Node189 -->
<g id="node27" class="node">
<title>Node189</title>
<g id="a_node27"><a xlink:href="printer_8h.html" target="_top" xlink:title="include/tvm/script\l/printer.h">
-<polygon fill="#ffffff" stroke="#000000" points="792,-397 792,-427 896,-427 896,-397 792,-397"/>
-<text text-anchor="start" x="800" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
-<text text-anchor="middle" x="844" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1118,-397 1118,-427 1222,-427 1222,-397 1118,-397"/>
+<text text-anchor="start" x="1126" y="-415" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/script</text>
+<text text-anchor="middle" x="1170" y="-404" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/printer.h</text>
</a>
</g>
</g>
<!-- Node13->Node189 -->
<g id="edge50" class="edge">
<title>Node13->Node189</title>
-<path fill="none" stroke="#191970" d="M908.1877,-468.0676C894.1309,-455.7891 875.0604,-439.1311 861.3648,-427.168"/>
-<polygon fill="#191970" stroke="#191970" points="905.9723,-470.7797 915.8063,-474.7223 910.5774,-465.5077 905.9723,-470.7797"/>
+<path fill="none" stroke="#191970" d="M1037.1533,-470.728C1065.3387,-458.268 1106.7644,-439.9548 1135.844,-427.0995"/>
+<polygon fill="#191970" stroke="#191970" points="1035.5089,-467.6281 1027.7779,-474.8726 1038.3393,-474.0304 1035.5089,-467.6281"/>
</g>
<!-- Node25 -->
<g id="node28" class="node">
<title>Node25</title>
<g id="a_node28"><a xlink:href="target__kind_8h.html" target="_top" xlink:title="Target kind registry. ">
-<polygon fill="#ffffff" stroke="#000000" points="1294.5,-324.5 1294.5,-354.5 1401.5,-354.5 1401.5,-324.5 1294.5,-324.5"/>
-<text text-anchor="start" x="1302.5" y="-342.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/target</text>
-<text text-anchor="middle" x="1348" y="-331.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/target_kind.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="477.5,-324.5 477.5,-354.5 584.5,-354.5 584.5,-324.5 477.5,-324.5"/>
+<text text-anchor="start" x="485.5" y="-342.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/target</text>
+<text text-anchor="middle" x="531" y="-331.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/target_kind.h</text>
</a>
</g>
</g>
<!-- Node13->Node25 -->
<g id="edge53" class="edge">
<title>Node13->Node25</title>
-<path fill="none" stroke="#191970" d="M981.7545,-472.996C989.2155,-471.5697 996.7859,-470.1951 1004,-469 1126.0424,-448.7819 1173.6336,-496.1608 1280,-433 1310.8923,-414.656 1331.8293,-376.2522 1341.6294,-354.8094"/>
-<polygon fill="#191970" stroke="#191970" points="980.8357,-469.6093 971.6921,-474.9612 982.1775,-476.4795 980.8357,-469.6093"/>
+<path fill="none" stroke="#191970" d="M946.5362,-473.1073C938.6305,-471.6772 930.6249,-470.2717 923,-469 814.4413,-450.8947 777.9304,-479.1162 678,-433 652.2615,-421.1221 653.4078,-406.9885 630,-391 609.3623,-376.9036 584.3853,-363.9368 564.6826,-354.5405"/>
+<polygon fill="#191970" stroke="#191970" points="945.9939,-476.5662 956.4607,-474.9239 947.2543,-469.6806 945.9939,-476.5662"/>
</g>
<!-- Node140->Node16 -->
<g id="edge12" class="edge">
<title>Node140->Node16</title>
-<path fill="none" stroke="#191970" d="M613.2464,-265.1009C579.107,-253.4366 529.625,-236.5302 498.9275,-226.0419"/>
-<polygon fill="#191970" stroke="#191970" points="612.2805,-268.4694 622.875,-268.3906 614.5437,-261.8454 612.2805,-268.4694"/>
+<path fill="none" stroke="#191970" d="M1376.3567,-266.1343C1425.8889,-254.463 1500.5576,-236.8686 1546.4101,-226.0643"/>
+<polygon fill="#191970" stroke="#191970" points="1375.4256,-262.7578 1366.4949,-268.4581 1377.0311,-269.5712 1375.4256,-262.7578"/>
</g>
<!-- Node140->Node148 -->
<g id="edge13" class="edge">
<title>Node140->Node148</title>
-<path fill="none" stroke="#191970" d="M719.9669,-271.1669C791.7321,-263.5653 907.2162,-249.9516 1006,-232 1010.5164,-231.1793 1015.1884,-230.2489 1019.8639,-229.2613"/>
-<polygon fill="#191970" stroke="#191970" points="719.2828,-267.7194 709.703,-272.2446 720.0138,-274.6811 719.2828,-267.7194"/>
+<path fill="none" stroke="#191970" d="M1344.5262,-261.4875C1354.9812,-252.1689 1367.948,-240.6116 1378.1207,-231.5446"/>
+<polygon fill="#191970" stroke="#191970" points="1341.9176,-259.1241 1336.7813,-268.3906 1346.5752,-264.3497 1341.9176,-259.1241"/>
</g>
<!-- Node140->Node149 -->
<g id="edge16" class="edge">
<title>Node140->Node149</title>
-<path fill="none" stroke="#191970" d="M720.1081,-272.2067C799.5899,-264.8068 924.6922,-250.6874 968,-232 977.681,-227.8226 1026.6654,-187.46 1053.8146,-164.7803"/>
-<polygon fill="#191970" stroke="#191970" points="719.3282,-268.7635 709.6898,-273.1626 719.9678,-275.7342 719.3282,-268.7635"/>
+<path fill="none" stroke="#191970" d="M1324.8256,-257.9447C1324.6379,-241.8928 1326.0958,-219.1446 1334,-201 1340.0696,-187.0668 1351.3196,-174.1918 1361.0522,-164.8001"/>
+<polygon fill="#191970" stroke="#191970" points="1321.3397,-258.4152 1325.1844,-268.2878 1328.3355,-258.1725 1321.3397,-258.4152"/>
</g>
<!-- Node140->Node150 -->
<g id="edge17" class="edge">
<title>Node140->Node150</title>
-<path fill="none" stroke="#191970" d="M696.3136,-265.7222C724.8242,-257.4375 762.1229,-245.5645 794,-232 891.3198,-190.5881 999.9077,-126.6455 1047.4595,-97.6818"/>
-<polygon fill="#191970" stroke="#191970" points="695.268,-262.3809 686.6188,-268.4998 697.196,-269.1101 695.268,-262.3809"/>
+<path fill="none" stroke="#191970" d="M1321.0995,-258.4481C1311.3233,-219.4428 1289.583,-132.7037 1280.7951,-97.6418"/>
+<polygon fill="#191970" stroke="#191970" points="1317.7438,-259.4561 1323.5701,-268.3051 1324.5338,-257.7542 1317.7438,-259.4561"/>
</g>
<!-- Node140->Node169 -->
<g id="edge18" class="edge">
<title>Node140->Node169</title>
-<path fill="none" stroke="#191970" d="M671.9175,-262.3119C684.5231,-252.8577 700.4921,-240.8809 712.9405,-231.5446"/>
-<polygon fill="#191970" stroke="#191970" points="669.7125,-259.5906 663.8125,-268.3906 673.9125,-265.1906 669.7125,-259.5906"/>
+<path fill="none" stroke="#191970" d="M1257.1047,-270.9382C1135.8237,-258.5069 886.7939,-232.9814 778.3711,-221.868"/>
+<polygon fill="#191970" stroke="#191970" points="1257.0722,-274.4532 1267.377,-271.9911 1257.7861,-267.4896 1257.0722,-274.4532"/>
</g>
<!-- Node140->Node26 -->
<g id="edge25" class="edge">
<title>Node140->Node26</title>
-<path fill="none" stroke="#191970" d="M719.7562,-271.5834C851.5923,-259.2801 1137.3853,-232.609 1256.3496,-221.5068"/>
-<polygon fill="#191970" stroke="#191970" points="719.403,-268.1011 709.7715,-272.5153 720.0535,-275.0708 719.403,-268.1011"/>
+<path fill="none" stroke="#191970" d="M1256.958,-274.143C1136.6195,-267.2085 880.6856,-251.5221 665,-232 636.8272,-229.45 605.5097,-226.0109 579.7441,-223.0201"/>
+<polygon fill="#191970" stroke="#191970" points="1256.946,-277.648 1267.1301,-274.7269 1257.3472,-270.6595 1256.946,-277.648"/>
</g>
<!-- Node140->Node176 -->
<g id="edge28" class="edge">
<title>Node140->Node176</title>
-<path fill="none" stroke="#191970" d="M632.4453,-261.8363C610.83,-243.9436 573.3,-215.6944 536,-201 490.2739,-182.9862 357.4539,-165.4335 278.6754,-156.1989"/>
-<polygon fill="#191970" stroke="#191970" points="630.2743,-264.584 640.1741,-268.3584 634.7888,-259.2342 630.2743,-264.584"/>
+<path fill="none" stroke="#191970" d="M1290.0345,-265.0181C1218.4843,-239.1917 1059.7633,-181.9005 996.6412,-159.1163"/>
+<polygon fill="#191970" stroke="#191970" points="1288.9386,-268.3435 1299.533,-268.4466 1291.3153,-261.7593 1288.9386,-268.3435"/>
</g>
<!-- Node140->Node183 -->
<g id="edge30" class="edge">
<title>Node140->Node183</title>
-<path fill="none" stroke="#191970" d="M581.9495,-268.9454C489.2003,-256.7832 326.0613,-235.3908 239.582,-224.0507"/>
-<polygon fill="#191970" stroke="#191970" points="581.7055,-272.4433 592.0757,-270.2733 582.6157,-265.5027 581.7055,-272.4433"/>
+<path fill="none" stroke="#191970" d="M1266.5127,-266.5673C1205.7211,-254.8839 1112.3949,-236.9478 1055.5021,-226.0137"/>
+<polygon fill="#191970" stroke="#191970" points="1265.8702,-270.0078 1276.3511,-268.4581 1267.1914,-263.1336 1265.8702,-270.0078"/>
</g>
<!-- Node148->Node149 -->
<g id="edge14" class="edge">
<title>Node148->Node149</title>
-<path fill="none" stroke="#191970" d="M1072,-191.0249C1072,-182.128 1072,-172.4287 1072,-164.6432"/>
-<polygon fill="#191970" stroke="#191970" points="1068.5001,-191.2966 1072,-201.2967 1075.5001,-191.2967 1068.5001,-191.2966"/>
+<path fill="none" stroke="#191970" d="M1388.9864,-191.3179C1386.8427,-182.3414 1384.4959,-172.5143 1382.6163,-164.6432"/>
+<polygon fill="#191970" stroke="#191970" points="1385.6423,-192.3831 1391.3694,-201.2967 1392.4508,-190.7572 1385.6423,-192.3831"/>
</g>
<!-- Node149->Node150 -->
<g id="edge15" class="edge">
<title>Node149->Node150</title>
-<path fill="none" stroke="#191970" d="M1072,-124.0249C1072,-115.128 1072,-105.4287 1072,-97.6432"/>
-<polygon fill="#191970" stroke="#191970" points="1068.5001,-124.2966 1072,-134.2967 1075.5001,-124.2967 1068.5001,-124.2966"/>
+<path fill="none" stroke="#191970" d="M1347.4791,-128.7951C1332.1955,-118.7558 1314.1652,-106.9124 1300.1673,-97.7177"/>
+<polygon fill="#191970" stroke="#191970" points="1345.8295,-131.8991 1356.1092,-134.4639 1349.6726,-126.0484 1345.8295,-131.8991"/>
</g>
<!-- Node169->Node171 -->
<g id="edge19" class="edge">
<title>Node169->Node171</title>
-<path fill="none" stroke="#191970" d="M764.8299,-195.7951C780.2634,-185.7558 798.4705,-173.9124 812.6056,-164.7177"/>
-<polygon fill="#191970" stroke="#191970" points="762.5893,-193.0772 756.1152,-201.4639 766.4063,-198.945 762.5893,-193.0772"/>
+<path fill="none" stroke="#191970" d="M739.6212,-192.4837C744.8868,-183.1996 750.753,-172.8565 755.4113,-164.6432"/>
+<polygon fill="#191970" stroke="#191970" points="736.5118,-190.8716 734.6228,-201.2967 742.6007,-194.325 736.5118,-190.8716"/>
</g>
<!-- Node169->Node172 -->
<g id="edge23" class="edge">
<title>Node169->Node172</title>
-<path fill="none" stroke="#191970" d="M735.8781,-191.348C740.7232,-149.0061 750.3629,-64.7637 754.2826,-30.5088"/>
-<polygon fill="#191970" stroke="#191970" points="732.3884,-191.0592 734.7287,-201.3923 739.343,-191.8551 732.3884,-191.0592"/>
+<path fill="none" stroke="#191970" d="M712.6119,-192.1504C708.7083,-183.7967 704.9661,-174.2238 703,-165 700.1277,-151.5249 699.7558,-147.3904 703,-134 712.698,-93.9712 738.4008,-52.3175 753.1631,-30.6508"/>
+<polygon fill="#191970" stroke="#191970" points="709.5797,-193.913 717.1718,-201.3029 715.8451,-190.7914 709.5797,-193.913"/>
</g>
<!-- Node169->Node173 -->
<g id="edge24" class="edge">
<title>Node169->Node173</title>
-<path fill="none" stroke="#191970" d="M742.2505,-191.9549C749.5143,-174.6471 760.7714,-151.5217 775,-134 786.5077,-119.8289 802.5077,-106.9539 815.6308,-97.6217"/>
-<polygon fill="#191970" stroke="#191970" points="738.9408,-190.8037 738.4428,-201.3869 745.4318,-193.4242 738.9408,-190.8037"/>
+<path fill="none" stroke="#191970" d="M700.7323,-194.5435C690.8703,-185.648 679.6404,-175.1049 670,-165 648.6043,-142.5734 625.9436,-114.5424 612.6548,-97.605"/>
+<polygon fill="#191970" stroke="#191970" points="698.4986,-197.2411 708.2915,-201.2845 703.1575,-192.0167 698.4986,-197.2411"/>
</g>
<!-- Node171->Node172 -->
<g id="edge20" class="edge">
<title>Node171->Node172</title>
-<path fill="none" stroke="#191970" d="M803.2614,-128.2715C792.8111,-119.9906 782.122,-109.6686 775,-98 762.1357,-76.9234 757.9732,-47.98 756.6316,-30.572"/>
-<polygon fill="#191970" stroke="#191970" points="801.4177,-131.2637 811.5184,-134.4618 805.6167,-125.6629 801.4177,-131.2637"/>
+<path fill="none" stroke="#191970" d="M764,-124.3415C764,-96.8131 764,-53.5714 764,-30.7614"/>
+<polygon fill="#191970" stroke="#191970" points="760.5001,-124.3889 764,-134.389 767.5001,-124.389 760.5001,-124.3889"/>
</g>
<!-- Node171->Node173 -->
<g id="edge21" class="edge">
<title>Node171->Node173</title>
-<path fill="none" stroke="#191970" d="M837.1407,-124.0249C837.539,-115.128 837.9733,-105.4287 838.3219,-97.6432"/>
-<polygon fill="#191970" stroke="#191970" points="833.6317,-124.1501 836.6807,-134.2967 840.6246,-124.4633 833.6317,-124.1501"/>
+<path fill="none" stroke="#191970" d="M717.9942,-130.5897C692.6006,-120.1518 661.4935,-107.3654 637.7399,-97.6017"/>
+<polygon fill="#191970" stroke="#191970" points="716.8398,-133.8992 727.4196,-134.4639 719.5011,-127.4248 716.8398,-133.8992"/>
</g>
<!-- Node173->Node172 -->
<g id="edge22" class="edge">
<title>Node173->Node172</title>
-<path fill="none" stroke="#191970" d="M812.0942,-60.7808C799.9012,-50.9383 785.7821,-39.541 774.7595,-30.6432"/>
-<polygon fill="#191970" stroke="#191970" points="810.1864,-63.7389 820.166,-67.2967 814.5832,-58.292 810.1864,-63.7389"/>
+<path fill="none" stroke="#191970" d="M647.0058,-63.5897C672.3994,-53.1518 703.5065,-40.3654 727.2601,-30.6017"/>
+<polygon fill="#191970" stroke="#191970" points="645.4989,-60.4248 637.5804,-67.4639 648.1602,-66.8992 645.4989,-60.4248"/>
</g>
<!-- Node26->Node8 -->
<g id="edge26" class="edge">
<title>Node26->Node8</title>
-<path fill="none" stroke="#191970" d="M1371.6904,-198.6842C1408.49,-188.0568 1454.7073,-174.7095 1489.7076,-164.6017"/>
-<polygon fill="#191970" stroke="#191970" points="1370.7016,-195.3267 1362.0653,-201.4639 1372.6438,-202.0519 1370.7016,-195.3267"/>
+<path fill="none" stroke="#191970" d="M464.911,-198.5484C428.8218,-187.9432 383.6413,-174.6665 349.3908,-164.6017"/>
+<polygon fill="#191970" stroke="#191970" points="464.2512,-202.0025 474.8323,-201.4639 466.2248,-195.2864 464.2512,-202.0025"/>
</g>
<!-- Node26->Node62 -->
<g id="edge27" class="edge">
<title>Node26->Node62</title>
-<path fill="none" stroke="#191970" d="M1302.9417,-191.6103C1300.3739,-182.5553 1297.5507,-172.5998 1295.2943,-164.6432"/>
-<polygon fill="#191970" stroke="#191970" points="1299.5931,-192.6309 1305.6886,-201.2967 1306.3275,-190.7211 1299.5931,-192.6309"/>
+<path fill="none" stroke="#191970" d="M533.0583,-191.6103C535.6261,-182.5553 538.4493,-172.5998 540.7057,-164.6432"/>
+<polygon fill="#191970" stroke="#191970" points="529.6725,-190.7211 530.3114,-201.2967 536.4069,-192.6309 529.6725,-190.7211"/>
</g>
<!-- Node176->Node177 -->
<g id="edge29" class="edge">
<title>Node176->Node177</title>
-<path fill="none" stroke="#191970" d="M245.1605,-135.0575C269.6098,-122.0567 304.85,-103.3179 325.9808,-92.0817"/>
-<polygon fill="#191970" stroke="#191970" points="243.2388,-132.1152 236.0527,-139.9005 246.5253,-138.2958 243.2388,-132.1152"/>
+<path fill="none" stroke="#191970" d="M1001.4353,-135.9118C1031.5615,-122.8895 1076.2642,-103.5664 1102.8335,-92.0817"/>
+<polygon fill="#191970" stroke="#191970" points="999.9981,-132.72 992.2077,-139.9005 1002.7756,-139.1454 999.9981,-132.72"/>
</g>
<!-- Node183->Node176 -->
<g id="edge31" class="edge">
<title>Node183->Node176</title>
-<path fill="none" stroke="#191970" d="M192.0814,-197.7374C198.8387,-185.1614 207.4554,-169.1246 212.8516,-159.0817"/>
-<polygon fill="#191970" stroke="#191970" points="188.808,-196.435 187.1579,-206.9005 194.9743,-199.7482 188.808,-196.435"/>
+<path fill="none" stroke="#191970" d="M995.9186,-197.7374C989.1613,-185.1614 980.5446,-169.1246 975.1484,-159.0817"/>
+<polygon fill="#191970" stroke="#191970" points="993.0257,-199.7482 1000.8421,-206.9005 999.192,-196.435 993.0257,-199.7482"/>
</g>
<!-- Node185->Node140 -->
<g id="edge33" class="edge">
<title>Node185->Node140</title>
-<path fill="none" stroke="#191970" d="M454.5795,-397.0966C500.3073,-369.3703 597.2507,-310.5901 635.2502,-287.5496"/>
-<polygon fill="#191970" stroke="#191970" points="452.6301,-394.1854 445.8938,-402.3631 456.2594,-400.1711 452.6301,-394.1854"/>
+<path fill="none" stroke="#191970" d="M1470.2482,-395.8613C1435.733,-367.6599 1365.4743,-310.2534 1337.6876,-287.5496"/>
+<polygon fill="#191970" stroke="#191970" points="1468.2472,-398.7461 1478.2055,-402.3631 1472.6763,-393.3255 1468.2472,-398.7461"/>
</g>
<!-- Node185->Node177 -->
<g id="edge37" class="edge">
<title>Node185->Node177</title>
-<path fill="none" stroke="#191970" d="M405.1529,-396.6355C390.7798,-386.5072 373.5107,-371.9946 363,-355 344.4591,-325.0216 344,-313.2487 344,-278 344,-278 344,-278 344,-216.5 344,-169.9722 344,-114.357 344,-92.2517"/>
-<polygon fill="#191970" stroke="#191970" points="403.3934,-399.6704 413.6417,-402.3582 407.3064,-393.8661 403.3934,-399.6704"/>
+<path fill="none" stroke="#191970" d="M1430.7161,-400.3694C1336.4686,-380.1612 1163,-335.1978 1163,-278 1163,-278 1163,-278 1163,-216.5 1163,-168.3794 1140.038,-114.1 1129.8127,-92.3143"/>
+<polygon fill="#191970" stroke="#191970" points="1430.1829,-403.834 1440.6907,-402.4792 1431.6315,-396.9855 1430.1829,-403.834"/>
</g>
<!-- Node185->Node186 -->
<g id="edge34" class="edge">
<title>Node185->Node186</title>
-<path fill="none" stroke="#191970" d="M430.2744,-392.1054C430.4673,-378.1237 430.7168,-360.0346 430.8673,-349.1228"/>
-<polygon fill="#191970" stroke="#191970" points="426.7732,-392.175 430.1349,-402.2223 433.7725,-392.2716 426.7732,-392.175"/>
+<path fill="none" stroke="#191970" d="M1500.0377,-393.3401C1507.6417,-379.2046 1517.7794,-360.3588 1523.8236,-349.1228"/>
+<polygon fill="#191970" stroke="#191970" points="1496.9148,-391.7576 1495.2597,-402.2223 1503.0795,-395.0738 1496.9148,-391.7576"/>
</g>
<!-- Node186->Node16 -->
<g id="edge35" class="edge">
<title>Node186->Node16</title>
-<path fill="none" stroke="#191970" d="M437.3129,-320.0879C446.0023,-293.3678 461.3169,-246.2756 467.8758,-226.1068"/>
-<polygon fill="#191970" stroke="#191970" points="433.903,-319.2562 434.1387,-329.8484 440.5598,-321.4211 433.903,-319.2562"/>
+<path fill="none" stroke="#191970" d="M1537.82,-320.7956C1550.3576,-294.2071 1572.8749,-246.4549 1582.47,-226.1068"/>
+<polygon fill="#191970" stroke="#191970" points="1534.6506,-319.3107 1533.5512,-329.8484 1540.982,-322.2963 1534.6506,-319.3107"/>
</g>
<!-- Node186->Node140 -->
<g id="edge36" class="edge">
<title>Node186->Node140</title>
-<path fill="none" stroke="#191970" d="M475.0111,-327.1969C516.7859,-315.5189 578.7032,-298.2103 616.8664,-287.5419"/>
-<polygon fill="#191970" stroke="#191970" points="474.0635,-323.8275 465.375,-329.8906 475.9481,-330.5691 474.0635,-323.8275"/>
+<path fill="none" stroke="#191970" d="M1487.4099,-326.9001C1448.8803,-315.2273 1392.4131,-298.1202 1357.496,-287.5419"/>
+<polygon fill="#191970" stroke="#191970" points="1486.696,-330.3408 1497.2813,-329.8906 1488.7256,-323.6415 1486.696,-330.3408"/>
</g>
<!-- Node188->Node172 -->
<g id="edge48" class="edge">
<title>Node188->Node172</title>
-<path fill="none" stroke="#191970" d="M1702.2049,-387.5738C1767.517,-360.3138 1860,-315.1536 1860,-278 1860,-278 1860,-278 1860,-149.5 1860,-43.1286 1041.8861,-20.5946 814.1883,-16.3829"/>
-<polygon fill="#191970" stroke="#191970" points="1700.689,-384.413 1692.7801,-391.4628 1703.3591,-390.8838 1700.689,-384.413"/>
+<path fill="none" stroke="#191970" d="M1851.1449,-390.3676C1946.4187,-359.9605 2105,-305.2548 2105,-278 2105,-278 2105,-278 2105,-149.5 2105,-84.0244 1079.4424,-30.5879 822.0515,-18.2076"/>
+<polygon fill="#191970" stroke="#191970" points="1850.0325,-387.0485 1841.5618,-393.4124 1852.1522,-393.7199 1850.0325,-387.0485"/>
</g>
<!-- Node25->Node26 -->
<g id="edge54" class="edge">
<title>Node25->Node26</title>
-<path fill="none" stroke="#191970" d="M1340.3159,-314.6276C1332.6439,-289.7948 1321.0997,-252.428 1314.6977,-231.7056"/>
-<polygon fill="#191970" stroke="#191970" points="1337.0446,-315.8965 1343.3405,-324.4178 1343.7327,-313.8302 1337.0446,-315.8965"/>
+<path fill="none" stroke="#191970" d="M529.9754,-314.2943C528.9659,-289.4615 527.4567,-252.3355 526.6181,-231.7056"/>
+<polygon fill="#191970" stroke="#191970" points="526.4835,-314.5682 530.3869,-324.4178 533.4778,-314.2838 526.4835,-314.5682"/>
</g>
<!-- Node190->Node13 -->
-<g id="edge60" class="edge">
+<g id="edge59" class="edge">
<title>Node190->Node13</title>
-<path fill="none" stroke="#191970" d="M936.9225,-531.9863C934.3194,-519.5286 931.0591,-503.9258 929.0021,-494.0817"/>
-<polygon fill="#191970" stroke="#191970" points="933.5227,-532.8279 938.9941,-541.9005 940.3747,-531.396 933.5227,-532.8279"/>
+<path fill="none" stroke="#191970" d="M1235.2781,-539.5596C1181.2205,-526.578 1095.8917,-506.0869 1045.9924,-494.1039"/>
+<polygon fill="#191970" stroke="#191970" points="1234.7815,-543.0398 1245.3224,-541.9717 1236.4161,-536.2333 1234.7815,-543.0398"/>
</g>
<!-- Node190->Node129 -->
-<g id="edge59" class="edge">
-<title>Node190->Node129</title>
-<path fill="none" stroke="#191970" d="M1003.7715,-539.7662C1040.7578,-531.4872 1087.8264,-518.5352 1127,-500 1148.6233,-489.7688 1148.9945,-478.3812 1171,-469 1248.0546,-436.1507 1274.0882,-450.5199 1356,-433 1364.6294,-431.1543 1373.7051,-429.1323 1382.6369,-427.0951"/>
-<polygon fill="#191970" stroke="#191970" points="1002.6471,-536.429 993.6217,-541.9782 1004.1377,-543.2685 1002.6471,-536.429"/>
-</g>
-<!-- Node191 -->
-<g id="node30" class="node">
-<title>Node191</title>
-<g id="a_node30"><a xlink:href="env__func_8h.html" target="_top" xlink:title="Serializable global function used in IR. ">
-<polygon fill="#ffffff" stroke="#ff0000" points="1013.5,-469.5 1013.5,-499.5 1118.5,-499.5 1118.5,-469.5 1013.5,-469.5"/>
-<text text-anchor="start" x="1021.5" y="-487.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/env</text>
-<text text-anchor="middle" x="1066" y="-476.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_func.h</text>
-</a>
-</g>
-</g>
-<!-- Node190->Node191 -->
<g id="edge58" class="edge">
-<title>Node190->Node191</title>
-<path fill="none" stroke="#191970" d="M967.8024,-537.1339C988.4787,-526.0514 1016.9968,-510.7657 1037.9743,-499.5218"/>
-<polygon fill="#191970" stroke="#191970" points="966.0697,-534.0915 958.9095,-541.9005 969.3767,-540.2611 966.0697,-534.0915"/>
+<title>Node190->Node129</title>
+<path fill="none" stroke="#191970" d="M1327.2158,-539.1354C1388.2815,-521.5969 1505.0173,-489.3677 1606,-469 1713.4545,-447.327 1742.4167,-454.0244 1850,-433 1858.8995,-431.2608 1868.2649,-429.2079 1877.4086,-427.0781"/>
+<polygon fill="#191970" stroke="#191970" points="1326.1859,-535.7897 1317.5471,-541.9234 1328.1254,-542.5157 1326.1859,-535.7897"/>
</g>
<!-- Node201->Node202 -->
-<g id="edge67" class="edge">
+<g id="edge66" class="edge">
<title>Node201->Node202</title>
-<path fill="none" stroke="#191970" d="M2336.2275,-661.7735C2342.2334,-652.4154 2348.9551,-641.9421 2354.2812,-633.6432"/>
-<polygon fill="#191970" stroke="#191970" points="2333.2131,-659.9903 2330.7574,-670.2967 2339.1042,-663.7712 2333.2131,-659.9903"/>
+<path fill="none" stroke="#191970" d="M2926.7725,-661.7735C2920.7666,-652.4154 2914.0449,-641.9421 2908.7188,-633.6432"/>
+<polygon fill="#191970" stroke="#191970" points="2923.8958,-663.7712 2932.2426,-670.2967 2929.7869,-659.9903 2923.8958,-663.7712"/>
</g>
<!-- Node204->Node201 -->
-<g id="edge70" class="edge">
+<g id="edge69" class="edge">
<title>Node204->Node201</title>
-<path fill="none" stroke="#191970" d="M2321,-727.0249C2321,-718.128 2321,-708.4287 2321,-700.6432"/>
-<polygon fill="#191970" stroke="#191970" points="2317.5001,-727.2966 2321,-737.2967 2324.5001,-727.2967 2317.5001,-727.2966"/>
+<path fill="none" stroke="#191970" d="M2942,-727.0249C2942,-718.128 2942,-708.4287 2942,-700.6432"/>
+<polygon fill="#191970" stroke="#191970" points="2938.5001,-727.2966 2942,-737.2967 2945.5001,-727.2967 2938.5001,-727.2966"/>
</g>
<!-- Node205->Node169 -->
<g id="edge92" class="edge">
<title>Node205->Node169</title>
-<path fill="none" stroke="#191970" d="M213.251,-733.0498C170.5408,-711.7967 114,-673.3105 114,-618.5 114,-618.5 114,-618.5 114,-484.5 114,-361.3932 531.5252,-259.7287 680.7579,-227.3233"/>
-<polygon fill="#191970" stroke="#191970" points="211.8723,-736.2705 222.4014,-737.4495 214.9057,-729.9618 211.8723,-736.2705"/>
+<path fill="none" stroke="#191970" d="M985.6415,-750.9363C814.2442,-746.5626 375.756,-732.4093 318,-701 277.7063,-679.0872 256,-664.3667 256,-618.5 256,-618.5 256,-618.5 256,-484.5 256,-434.6385 275.6942,-417.3893 318,-391 422.5861,-325.7619 482.9844,-410.5954 593,-355 650.7215,-325.831 697.3381,-261.3335 716.5323,-231.8034"/>
+<polygon fill="#191970" stroke="#191970" points="985.7209,-754.4394 995.8058,-751.1922 985.8971,-747.4416 985.7209,-754.4394"/>
</g>
<!-- Node205->Node176 -->
<g id="edge93" class="edge">
<title>Node205->Node176</title>
-<path fill="none" stroke="#191970" d="M191.6761,-739.4345C147.3062,-729.8094 94.5753,-715.8566 77,-701 46.0263,-674.8176 38,-659.0572 38,-618.5 38,-618.5 38,-618.5 38,-278 38,-209.847 124.5345,-174.4271 177.6045,-159.113"/>
-<polygon fill="#191970" stroke="#191970" points="191.0714,-742.8841 201.581,-741.5426 192.5287,-736.0375 191.0714,-742.8841"/>
+<path fill="none" stroke="#191970" d="M985.6852,-746.9812C805.6576,-731.5269 332,-684.5864 332,-618.5 332,-618.5 332,-618.5 332,-484.5 332,-439.1247 336.8398,-418.4109 373,-391 464.5633,-321.5914 520.7197,-393.4287 629,-355 770.0279,-304.9492 916.4082,-192.5782 958.2228,-159.0928"/>
+<polygon fill="#191970" stroke="#191970" points="985.5711,-750.4841 995.8321,-747.8456 986.1653,-743.5094 985.5711,-750.4841"/>
</g>
<!-- Node205->Node183 -->
<g id="edge94" class="edge">
<title>Node205->Node183</title>
-<path fill="none" stroke="#191970" d="M191.785,-737.3153C154.4965,-727.7394 113.2569,-714.6375 100,-701 73.3828,-673.6185 76,-656.6867 76,-618.5 76,-618.5 76,-618.5 76,-339.5 76,-285.9395 134.3893,-244.1292 164.4963,-226.1309"/>
-<polygon fill="#191970" stroke="#191970" points="191.1127,-740.7551 201.6641,-739.7959 192.8175,-733.9658 191.1127,-740.7551"/>
+<path fill="none" stroke="#191970" d="M985.6074,-743.4274C844.5479,-720.6127 523.3074,-646.5326 389,-433 329.2247,-337.9644 482.1175,-434.1981 744,-355 846.4353,-324.0217 955.3742,-251.9515 992.625,-226.0142"/>
+<polygon fill="#191970" stroke="#191970" points="985.4296,-746.9426 995.8532,-745.0452 986.5215,-740.0282 985.4296,-746.9426"/>
</g>
<!-- Node205->Node186 -->
-<g id="edge72" class="edge">
+<g id="edge71" class="edge">
<title>Node205->Node186</title>
-<path fill="none" stroke="#191970" d="M239.1931,-729.4738C218.5586,-704.1788 190,-661.4177 190,-618.5 190,-618.5 190,-618.5 190,-484.5 190,-434.6385 211.7901,-420.4845 252,-391 286.944,-365.3768 334.3496,-352.4676 371.5969,-345.9828"/>
-<polygon fill="#191970" stroke="#191970" points="236.7149,-731.9645 245.8345,-737.3575 242.0685,-727.4546 236.7149,-731.9645"/>
+<path fill="none" stroke="#191970" d="M1055.3119,-727.291C1056.6702,-710.7319 1059.4249,-688.7464 1065,-670 1086.1073,-599.0257 1099.9502,-573.1548 1164,-536 1262.0178,-479.1406 1299.9081,-495.838 1410,-469 1483.7669,-451.0173 1529.3025,-492.0753 1577,-433 1588.7264,-418.4764 1583.8391,-408.3687 1577,-391 1570.1038,-373.4864 1554.1345,-358.4072 1542.5037,-349.1712"/>
+<polygon fill="#191970" stroke="#191970" points="1051.8182,-727.0788 1054.6054,-737.3005 1058.8009,-727.5718 1051.8182,-727.0788"/>
</g>
<!-- Node205->Node129 -->
-<g id="edge73" class="edge">
+<g id="edge72" class="edge">
<title>Node205->Node129</title>
-<path fill="none" stroke="#191970" d="M297.2547,-732.8602C334.2835,-714.116 393.171,-686.3761 447,-670 572.6569,-631.7722 902.4517,-602.1645 1029,-567 1145.9823,-534.4936 1170.0557,-510.9312 1284,-469 1323.8725,-454.327 1369.5512,-438.3278 1402.2177,-427.0253"/>
-<polygon fill="#191970" stroke="#191970" points="295.6416,-729.7539 288.3279,-737.4195 298.8256,-735.9879 295.6416,-729.7539"/>
+<path fill="none" stroke="#191970" d="M1122.323,-747.2866C1220.2581,-739.2094 1395.3202,-722.3519 1455,-701 1507.8298,-682.0988 1517.509,-668.1876 1562,-634 1612.6168,-595.1052 1614.3165,-572.046 1667,-536 1723.5998,-497.2744 1743.6971,-497.6629 1806,-469 1837.9029,-454.3228 1874.5692,-438.3241 1900.8109,-427.0228"/>
+<polygon fill="#191970" stroke="#191970" points="1121.9059,-743.8089 1112.2238,-748.111 1122.4754,-750.7857 1121.9059,-743.8089"/>
</g>
<!-- Node205->Node190 -->
-<g id="edge74" class="edge">
+<g id="edge73" class="edge">
<title>Node205->Node190</title>
-<path fill="none" stroke="#191970" d="M282.76,-730.6287C303.7496,-711.7011 336.6241,-685.0677 370,-670 534.9441,-595.5351 747.9329,-567.2855 861.4556,-557.0171"/>
-<polygon fill="#191970" stroke="#191970" points="280.3156,-728.1216 275.3212,-737.4654 285.0524,-733.2756 280.3156,-728.1216"/>
+<path fill="none" stroke="#191970" d="M1059.5146,-727.5695C1068.1045,-694.4717 1087.8111,-636.3327 1126,-603 1151.3578,-580.8667 1186.3814,-568.2176 1217.1045,-561.0021"/>
+<polygon fill="#191970" stroke="#191970" points="1056.0932,-726.8262 1057.107,-737.3725 1062.8912,-728.4959 1056.0932,-726.8262"/>
</g>
<!-- Node205->Node208 -->
-<g id="edge75" class="edge">
+<g id="edge74" class="edge">
<title>Node205->Node208</title>
-<path fill="none" stroke="#191970" d="M328.0416,-746.8857C484.4593,-733.9793 867.4065,-702.3815 1013.7964,-690.3025"/>
-<polygon fill="#191970" stroke="#191970" points="327.7495,-743.3978 318.0712,-747.7084 328.3251,-750.3741 327.7495,-743.3978"/>
+<path fill="none" stroke="#191970" d="M1122.1244,-738.8343C1183.2774,-726.5671 1272.0736,-708.7547 1329.929,-697.149"/>
+<polygon fill="#191970" stroke="#191970" points="1121.2946,-735.431 1112.1784,-740.8295 1122.6714,-742.2942 1121.2946,-735.431"/>
</g>
-<!-- Node205->Node213 -->
+<!-- Node205->Node214 -->
<g id="edge91" class="edge">
-<title>Node205->Node213</title>
-<path fill="none" stroke="#191970" d="M328.2017,-750.8166C506.594,-746.0438 983.9158,-730.7991 1139,-701 1216.9804,-686.0162 1304.2375,-652.9658 1351.648,-633.5511"/>
-<polygon fill="#191970" stroke="#191970" points="328.0588,-747.3191 318.155,-751.0828 328.2442,-754.3166 328.0588,-747.3191"/>
+<title>Node205->Node214</title>
+<path fill="none" stroke="#191970" d="M1122.3449,-747.7493C1234.2959,-739.5139 1451.7374,-721.4905 1526,-701 1583.6596,-685.0906 1646.0937,-652.6951 1680.1747,-633.574"/>
+<polygon fill="#191970" stroke="#191970" points="1121.8946,-744.2728 1112.1761,-748.4918 1122.4044,-751.2542 1121.8946,-744.2728"/>
</g>
<!-- Node208->Node8 -->
-<g id="edge76" class="edge">
+<g id="edge75" class="edge">
<title>Node208->Node8</title>
-<path fill="none" stroke="#191970" d="M1069.8537,-660.4088C1069.5969,-642.9482 1071.8426,-619.8663 1083,-603 1132.5819,-528.0488 1179.7378,-545.8978 1257,-500 1280.1325,-486.2581 1283.7173,-478.2053 1309,-469 1402.9238,-434.803 1458.9312,-502.2619 1531,-433 1561.5743,-403.6165 1550,-381.9049 1550,-339.5 1550,-339.5 1550,-339.5 1550,-278 1550,-236.8719 1545.8672,-188.4822 1543.5425,-164.5162"/>
-<polygon fill="#191970" stroke="#191970" points="1066.3588,-660.6056 1070.2796,-670.4482 1073.3525,-660.3088 1066.3588,-660.6056"/>
+<path fill="none" stroke="#191970" d="M1377.5902,-660.9825C1368.2285,-642.0424 1352.6146,-616.805 1331,-603 1274.1779,-566.7083 799.3158,-512.1677 733,-500 579.7957,-471.8898 526.8058,-505.6037 389,-433 336.5876,-405.3863 294,-398.7417 294,-339.5 294,-339.5 294,-339.5 294,-278 294,-236.9315 296.0664,-188.5157 297.2288,-164.5303"/>
+<polygon fill="#191970" stroke="#191970" points="1374.5046,-662.6465 1381.9145,-670.219 1380.8442,-659.6784 1374.5046,-662.6465"/>
</g>
<!-- Node208->Node190 -->
-<g id="edge77" class="edge">
+<g id="edge76" class="edge">
<title>Node208->Node190</title>
-<path fill="none" stroke="#191970" d="M1051.3655,-662.6848C1036.0555,-645.9318 1014.538,-622.7497 995,-603 980.2401,-588.0801 962.5363,-571.42 951.5251,-561.1968"/>
-<polygon fill="#191970" stroke="#191970" points="1049.0253,-665.3132 1058.3467,-670.3493 1054.2003,-660.5995 1049.0253,-665.3132"/>
+<path fill="none" stroke="#191970" d="M1385.3419,-660.3486C1382.447,-642.8596 1376.4353,-619.7686 1364,-603 1349.6641,-583.6684 1325.7436,-569.557 1307.9728,-561.0449"/>
+<polygon fill="#191970" stroke="#191970" points="1381.8991,-660.9992 1386.7713,-670.4074 1388.8295,-660.0144 1381.8991,-660.9992"/>
</g>
<!-- Node208->Node210 -->
-<g id="edge78" class="edge">
+<g id="edge77" class="edge">
<title>Node208->Node210</title>
-<path fill="none" stroke="#191970" d="M1077.0352,-660.4614C1082.1311,-641.7576 1091.9301,-617.0863 1110,-603 1143.041,-577.2431 1256.4163,-562.8143 1326.8608,-556.1421"/>
-<polygon fill="#191970" stroke="#191970" points="1073.5814,-659.851 1074.6082,-670.396 1080.3815,-661.5123 1073.5814,-659.851"/>
+<path fill="none" stroke="#191970" d="M1456.0664,-679.6012C1550.4404,-670.8753 1715.8721,-653.4483 1773,-634 1820.2107,-617.9279 1869.3068,-585.5909 1895.9151,-566.5239"/>
+<polygon fill="#191970" stroke="#191970" points="1455.6807,-676.1218 1446.0419,-680.52 1456.3197,-683.0926 1455.6807,-676.1218"/>
</g>
-<!-- Node208->Node213 -->
+<!-- Node208->Node214 -->
<g id="edge80" class="edge">
-<title>Node208->Node213</title>
-<path fill="none" stroke="#191970" d="M1140.0319,-671.0297C1196.3631,-659.0482 1275.541,-642.2071 1328.9299,-630.8514"/>
-<polygon fill="#191970" stroke="#191970" points="1139.196,-667.6291 1130.1429,-673.1331 1140.6523,-674.476 1139.196,-667.6291"/>
+<title>Node208->Node214</title>
+<path fill="none" stroke="#191970" d="M1456.3349,-671.1024C1513.3638,-659.0869 1593.7651,-642.147 1647.7751,-630.7675"/>
+<polygon fill="#191970" stroke="#191970" points="1455.3897,-667.7246 1446.3261,-673.2112 1456.8329,-674.5742 1455.3897,-667.7246"/>
</g>
-<!-- Node208->Node219 -->
+<!-- Node208->Node220 -->
<g id="edge88" class="edge">
-<title>Node208->Node219</title>
-<path fill="none" stroke="#191970" d="M1098.4046,-664.6286C1113.3829,-654.5979 1132.0595,-642.7976 1147.5146,-633.6432"/>
-<polygon fill="#191970" stroke="#191970" points="1096.3684,-661.7807 1090.0653,-670.2967 1100.3033,-667.57 1096.3684,-661.7807"/>
+<title>Node208->Node220</title>
+<path fill="none" stroke="#191970" d="M1345.2778,-665.7743C1325.4833,-655.5335 1302.8324,-643.2132 1286.4016,-633.7177"/>
+<polygon fill="#191970" stroke="#191970" points="1343.9356,-669.0192 1354.4315,-670.4639 1347.1274,-662.7892 1343.9356,-669.0192"/>
</g>
-<!-- Node208->Node220 -->
+<!-- Node208->Node221 -->
<g id="edge90" class="edge">
-<title>Node208->Node220</title>
-<path fill="none" stroke="#191970" d="M1028.2171,-666.4516C1004.2841,-656.0392 975.0545,-643.3224 952.7113,-633.6017"/>
-<polygon fill="#191970" stroke="#191970" points="1026.8733,-669.6838 1037.4394,-670.4639 1029.666,-663.265 1026.8733,-669.6838"/>
+<title>Node208->Node221</title>
+<path fill="none" stroke="#191970" d="M1418.6246,-664.3469C1433.0578,-654.3776 1449.9467,-642.7121 1463.0763,-633.6432"/>
+<polygon fill="#191970" stroke="#191970" points="1416.2497,-661.7335 1410.0108,-670.2967 1420.228,-667.4932 1416.2497,-661.7335"/>
</g>
<!-- Node210->Node211 -->
-<g id="edge79" class="edge">
+<g id="edge78" class="edge">
<title>Node210->Node211</title>
-<path fill="none" stroke="#191970" d="M1381.6173,-526.3179C1380.4115,-517.3414 1379.0915,-507.5143 1378.0342,-499.6432"/>
-<polygon fill="#191970" stroke="#191970" points="1378.1575,-526.8517 1382.9578,-536.2967 1385.0952,-525.9197 1378.1575,-526.8517"/>
+<path fill="none" stroke="#191970" d="M1984.3908,-537.1807C2039.753,-525.5892 2117.1835,-509.3772 2171.4816,-498.0085"/>
+<polygon fill="#191970" stroke="#191970" points="1983.3913,-533.814 1974.3208,-539.2891 1984.8259,-540.6654 1983.3913,-533.814"/>
</g>
-<!-- Node213->Node6 -->
+<!-- Node210->Node212 -->
+<g id="edge79" class="edge">
+<title>Node210->Node212</title>
+<path fill="none" stroke="#191970" d="M1900.7725,-527.7735C1894.7666,-518.4154 1888.0449,-507.9421 1882.7188,-499.6432"/>
+<polygon fill="#191970" stroke="#191970" points="1897.8958,-529.7712 1906.2426,-536.2967 1903.7869,-525.9903 1897.8958,-529.7712"/>
+</g>
+<!-- Node214->Node6 -->
<g id="edge81" class="edge">
-<title>Node213->Node6</title>
-<path fill="none" stroke="#191970" d="M1421.1387,-598.301C1459.3686,-575.7171 1517.3898,-541.5617 1528,-536 1609.2301,-493.4206 1707.7367,-449.4148 1759.0658,-427.039"/>
-<polygon fill="#191970" stroke="#191970" points="1419.1887,-595.3879 1412.3604,-603.4888 1422.7501,-601.4142 1419.1887,-595.3879"/>
+<title>Node214->Node6</title>
+<path fill="none" stroke="#191970" d="M1646.1972,-600.5657C1580.0096,-581.4514 1469.868,-551.8105 1373,-536 1177.1674,-504.0368 1124.714,-525.9932 928,-500 764.922,-478.4514 725.03,-466.5874 564,-433 555.2043,-431.1654 545.95,-429.1286 536.8587,-427.0657"/>
+<polygon fill="#191970" stroke="#191970" points="1645.4558,-603.9949 1656.0352,-603.4229 1647.4081,-597.2727 1645.4558,-603.9949"/>
</g>
-<!-- Node213->Node188 -->
+<!-- Node214->Node188 -->
<g id="edge87" class="edge">
-<title>Node213->Node188</title>
-<path fill="none" stroke="#191970" d="M1419.794,-597.4362C1431.0781,-589 1443.0997,-578.5364 1452,-567 1481.0539,-529.3411 1460.7579,-502.0121 1495,-469 1507.6302,-456.8234 1544.8035,-442.2492 1578.2154,-430.886"/>
-<polygon fill="#191970" stroke="#191970" points="1417.4587,-594.8039 1411.3727,-603.4763 1421.5385,-600.4921 1417.4587,-594.8039"/>
+<title>Node214->Node188</title>
+<path fill="none" stroke="#191970" d="M1683.6665,-595.6413C1676.9688,-587.3061 1670.482,-577.3658 1667,-567 1662.6127,-553.9394 1661.9959,-548.8369 1667,-536 1684.0469,-492.2701 1724.5286,-454.6327 1752.3594,-432.6373"/>
+<polygon fill="#191970" stroke="#191970" points="1681.1594,-598.094 1690.3311,-603.3979 1686.4688,-593.5321 1681.1594,-598.094"/>
</g>
-<!-- Node213->Node190 -->
+<!-- Node214->Node190 -->
<g id="edge82" class="edge">
-<title>Node213->Node190</title>
-<path fill="none" stroke="#191970" d="M1318.5574,-608.2183C1232.9448,-595.3572 1087.5233,-573.5114 1004.3608,-561.0183"/>
-<polygon fill="#191970" stroke="#191970" points="1318.2843,-611.7164 1328.6933,-609.7409 1319.3242,-604.7941 1318.2843,-611.7164"/>
+<title>Node214->Node190</title>
+<path fill="none" stroke="#191970" d="M1637.993,-608.013C1571.6366,-597.7355 1467.8448,-581.5401 1378,-567 1366.2044,-565.0911 1353.5738,-563.0104 1341.5435,-561.0113"/>
+<polygon fill="#191970" stroke="#191970" points="1637.4915,-611.477 1647.9093,-609.548 1638.5624,-604.5594 1637.4915,-611.477"/>
</g>
-<!-- Node213->Node210 -->
+<!-- Node214->Node210 -->
<g id="edge83" class="edge">
-<title>Node213->Node210</title>
-<path fill="none" stroke="#191970" d="M1386.2395,-593.0249C1385.974,-584.128 1385.6844,-574.4287 1385.452,-566.6432"/>
-<polygon fill="#191970" stroke="#191970" points="1382.7493,-593.4056 1386.5462,-603.2967 1389.7461,-593.1967 1382.7493,-593.4056"/>
+<title>Node214->Node210</title>
+<path fill="none" stroke="#191970" d="M1762.6928,-600.4123C1795.8618,-589.8298 1837.2547,-576.6235 1868.6664,-566.6017"/>
+<polygon fill="#191970" stroke="#191970" points="1761.5911,-597.0899 1753.1281,-603.4639 1763.7188,-603.7587 1761.5911,-597.0899"/>
</g>
-<!-- Node213->Node211 -->
+<!-- Node214->Node212 -->
<g id="edge86" class="edge">
-<title>Node213->Node211</title>
-<path fill="none" stroke="#191970" d="M1346.6552,-597.8289C1335.2818,-589.8471 1324.3005,-579.5645 1318,-567 1311.8241,-554.6839 1312.3525,-548.5671 1318,-536 1324.7893,-520.8922 1338.4438,-508.4132 1350.7992,-499.5246"/>
-<polygon fill="#191970" stroke="#191970" points="1345.0656,-600.97 1355.3579,-603.4839 1348.8797,-595.1004 1345.0656,-600.97"/>
+<title>Node214->Node212</title>
+<path fill="none" stroke="#191970" d="M1686.8483,-595.3371C1674.6815,-577.6625 1663.5347,-553.4605 1677,-536 1693.8125,-514.1993 1763.3186,-499.8067 1814.809,-491.912"/>
+<polygon fill="#191970" stroke="#191970" points="1684.056,-597.4479 1692.7996,-603.4312 1689.6956,-593.3012 1684.056,-597.4479"/>
</g>
-<!-- Node213->Node214 -->
+<!-- Node214->Node215 -->
<g id="edge84" class="edge">
-<title>Node213->Node214</title>
-<path fill="none" stroke="#191970" d="M1437.4933,-600.2759C1469.2851,-589.7165 1510.1771,-576.5805 1542.3686,-566.6017"/>
-<polygon fill="#191970" stroke="#191970" points="1436.3168,-596.9787 1427.9387,-603.4639 1438.5324,-603.6188 1436.3168,-596.9787"/>
+<title>Node214->Node215</title>
+<path fill="none" stroke="#191970" d="M1713.2671,-594.1932C1718.0159,-584.9844 1724.0637,-574.771 1729.6578,-566.6432"/>
+<polygon fill="#191970" stroke="#191970" points="1710.0673,-592.7706 1708.862,-603.2967 1716.3684,-595.8197 1710.0673,-592.7706"/>
</g>
-<!-- Node214->Node213 -->
+<!-- Node215->Node214 -->
<g id="edge85" class="edge">
-<title>Node214->Node213</title>
-<path fill="none" stroke="#191970" d="M1544.2975,-569.7936C1512.4658,-580.3642 1471.573,-593.499 1439.4198,-603.4639"/>
-<polygon fill="#191970" stroke="#191970" points="1545.4872,-573.0864 1553.8657,-566.6017 1543.272,-566.4461 1545.4872,-573.0864"/>
+<title>Node215->Node214</title>
+<path fill="none" stroke="#191970" d="M1736.768,-575.7386C1732.0256,-584.9425 1725.9799,-595.1586 1720.3836,-603.2967"/>
+<polygon fill="#191970" stroke="#191970" points="1739.9637,-577.1697 1741.1649,-566.6432 1733.6614,-574.1231 1739.9637,-577.1697"/>
</g>
-<!-- Node219->Node208 -->
+<!-- Node220->Node208 -->
<g id="edge89" class="edge">
-<title>Node219->Node208</title>
-<path fill="none" stroke="#191970" d="M1150.5469,-639.4038C1135.5183,-649.4667 1116.778,-661.3044 1101.3046,-670.4639"/>
-<polygon fill="#191970" stroke="#191970" points="1152.6103,-642.2334 1158.9139,-633.7177 1148.6757,-636.4438 1152.6103,-642.2334"/>
+<title>Node220->Node208</title>
+<path fill="none" stroke="#191970" d="M1306.8329,-638.283C1326.7276,-648.5779 1349.4876,-660.9619 1365.9125,-670.4639"/>
+<polygon fill="#191970" stroke="#191970" points="1308.4223,-635.1647 1297.9267,-633.7177 1305.2292,-641.394 1308.4223,-635.1647"/>
</g>
-<!-- Node222->Node168 -->
-<g id="edge98" class="edge">
-<title>Node222->Node168</title>
-<path fill="none" stroke="#191970" d="M1725.3788,-728.4837C1720.1132,-719.1996 1714.247,-708.8565 1709.5887,-700.6432"/>
-<polygon fill="#191970" stroke="#191970" points="1722.3993,-730.325 1730.3772,-737.2967 1728.4882,-726.8716 1722.3993,-730.325"/>
-</g>
-<!-- Node224->Node13 -->
-<g id="edge107" class="edge">
-<title>Node224->Node13</title>
-<path fill="none" stroke="#191970" d="M886.061,-731.3073C875.3325,-722.9128 864.058,-712.5094 856,-701 830.2814,-664.2656 825.5186,-647.3718 832,-603 836.4916,-572.2501 833.7022,-561.1181 852,-536 865.8573,-516.9776 888.9909,-502.6523 905.9082,-494.0004"/>
-<polygon fill="#191970" stroke="#191970" points="883.9928,-734.1311 894.0963,-737.32 888.1867,-728.5265 883.9928,-734.1311"/>
+<!-- Node223->Node168 -->
+<g id="edge99" class="edge">
+<title>Node223->Node168</title>
+<path fill="none" stroke="#191970" d="M2322,-727.0249C2322,-718.128 2322,-708.4287 2322,-700.6432"/>
+<polygon fill="#191970" stroke="#191970" points="2318.5001,-727.2966 2322,-737.2967 2325.5001,-727.2967 2318.5001,-727.2966"/>
</g>
-<!-- Node224->Node16 -->
-<g id="edge102" class="edge">
-<title>Node224->Node16</title>
-<path fill="none" stroke="#191970" d="M849.8469,-744.989C806.1884,-738.0823 749.4117,-725.0892 704,-701 685.935,-691.4172 687.4797,-680.6128 670,-670 630.3368,-645.9185 614.511,-652.5983 572,-634 512.0875,-607.7887 493.2785,-606.2874 441,-567 408.3241,-542.444 400.0178,-534.4374 378,-500 335.7113,-433.8575 326.5756,-393.5444 363,-324 385.9532,-280.1759 433.0003,-243.0157 456.6942,-226.1678"/>
-<polygon fill="#191970" stroke="#191970" points="849.3878,-748.4593 859.7988,-746.4946 850.4349,-741.5381 849.3878,-748.4593"/>
+<!-- Node225->Node13 -->
+<g id="edge108" class="edge">
+<title>Node225->Node13</title>
+<path fill="none" stroke="#191970" d="M1681.8639,-750.9859C1523.1565,-745.9496 1135.8043,-724.4574 1050,-634 1011.4572,-593.3671 1006.4771,-520.6986 1005.9827,-494.3688"/>
+<polygon fill="#191970" stroke="#191970" points="1681.8793,-754.4879 1691.9814,-751.2944 1682.0927,-747.4911 1681.8793,-754.4879"/>
</g>
-<!-- Node224->Node140 -->
+<!-- Node225->Node16 -->
<g id="edge103" class="edge">
-<title>Node224->Node140</title>
-<path fill="none" stroke="#191970" d="M885.1955,-731.6442C818.1751,-686.7868 667.2925,-573.7056 613,-433 592.5531,-380.0094 629.2207,-312.4927 644.642,-287.6946"/>
-<polygon fill="#191970" stroke="#191970" points="883.428,-734.6715 893.6992,-737.27 887.2903,-728.8334 883.428,-734.6715"/>
+<title>Node225->Node16</title>
+<path fill="none" stroke="#191970" d="M1802.9694,-733.785C1824.9827,-725.0886 1850.3851,-713.8254 1872,-701 1891.0422,-689.7011 1892.5457,-682.2356 1911,-670 1938.8482,-651.536 1952.1961,-656.3878 1977,-634 2060.4325,-558.6945 2114.3446,-480.9811 2047,-391 1994.4767,-320.8221 1730.4802,-251.0042 1627.8738,-226.0805"/>
+<polygon fill="#191970" stroke="#191970" points="1801.5042,-730.599 1793.4426,-737.4738 1804.0318,-737.1268 1801.5042,-730.599"/>
</g>
-<!-- Node224->Node185 -->
+<!-- Node225->Node140 -->
<g id="edge104" class="edge">
-<title>Node224->Node185</title>
-<path fill="none" stroke="#191970" d="M873.4011,-733.2748C806.2114,-703.1899 676.6171,-640.733 580,-567 518.3783,-519.9736 458.3369,-447.655 437.6098,-421.6934"/>
-<polygon fill="#191970" stroke="#191970" points="872.2907,-736.6114 882.8502,-737.4761 875.1347,-730.2151 872.2907,-736.6114"/>
+<title>Node225->Node140</title>
+<path fill="none" stroke="#191970" d="M1681.8688,-747.4449C1578.3509,-739.2369 1386.5601,-721.815 1321,-701 1306.0739,-696.261 1205.1209,-647.3902 1197,-634 1174.4125,-596.7563 1169.1386,-570.2892 1196,-536 1237.2144,-483.3888 1277.2505,-516.5551 1342,-500 1454.6953,-471.1862 1522.8782,-524.2629 1595,-433 1625.0419,-394.985 1628.2632,-361.0171 1597,-324 1583.4325,-307.9354 1460.0928,-292.187 1384.5631,-283.934"/>
+<polygon fill="#191970" stroke="#191970" points="1681.6651,-750.9396 1691.9086,-748.2341 1682.2137,-743.9612 1681.6651,-750.9396"/>
</g>
-<!-- Node224->Node186 -->
+<!-- Node225->Node185 -->
<g id="edge105" class="edge">
-<title>Node224->Node186</title>
-<path fill="none" stroke="#191970" d="M850.1191,-739.3401C816.3787,-731.1564 775.7189,-718.7469 742,-701 722.7124,-690.8486 722.3356,-681.7842 704,-670 673.568,-650.4415 661.8583,-652.8788 631,-634 503.3869,-555.9274 428.142,-566.6877 361,-433 352.6223,-416.3189 352.5753,-407.6574 361,-391 370.8736,-371.478 392.0194,-357.5488 408.5005,-349.1397"/>
-<polygon fill="#191970" stroke="#191970" points="849.3594,-742.7569 859.8953,-741.6411 850.9632,-735.9431 849.3594,-742.7569"/>
+<title>Node225->Node185</title>
+<path fill="none" stroke="#191970" d="M1681.4069,-748.5673C1569.091,-741.5345 1351.1266,-725.2342 1278,-701 1222.2143,-682.5126 1193.1874,-685.009 1164,-634 1142.3684,-596.1958 1139.2751,-571.8576 1164,-536 1181.6595,-510.389 1381.0359,-445.9922 1459.0089,-421.5823"/>
+<polygon fill="#191970" stroke="#191970" points="1681.4122,-752.0742 1691.6094,-749.1988 1681.8448,-745.0876 1681.4122,-752.0742"/>
</g>
-<!-- Node224->Node129 -->
+<!-- Node225->Node186 -->
<g id="edge106" class="edge">
-<title>Node224->Node129</title>
-<path fill="none" stroke="#191970" d="M986.3249,-742.7372C1030.9694,-734.8812 1089.8838,-721.6765 1139,-701 1164.5106,-690.2608 1226.4015,-655.3639 1244,-634 1294.1133,-573.1641 1253.799,-525.26 1309,-469 1328.5756,-449.0488 1356.1672,-435.7705 1381.4636,-427.0925"/>
-<polygon fill="#191970" stroke="#191970" points="985.7161,-739.2905 976.4507,-744.4288 986.8982,-746.1899 985.7161,-739.2905"/>
+<title>Node225->Node186</title>
+<path fill="none" stroke="#191970" d="M1786.1114,-732.3471C1801.9772,-723.1821 1820.673,-711.954 1837,-701 1855.9706,-688.2724 1858.7698,-682.3319 1878,-670 1906.4556,-651.7519 1918.7963,-655.3561 1945,-634 1975.0837,-609.4818 1985.0977,-602.4019 2001,-567 2033.258,-495.1867 2073.6902,-449.4934 2021,-391 1992.5111,-359.3734 1713.0858,-345.8455 1588.4667,-341.3516"/>
+<polygon fill="#191970" stroke="#191970" points="1784.0745,-729.4803 1777.1363,-737.4873 1787.5534,-735.5547 1784.0745,-729.4803"/>
</g>
-<!-- Node224->Node190 -->
-<g id="edge108" class="edge">
-<title>Node224->Node190</title>
-<path fill="none" stroke="#191970" d="M900.3496,-729.2439C879.2718,-700.9099 846.0852,-654.3127 840,-634 836.0461,-620.8017 832.8029,-614.7486 840,-603 853.2119,-581.4329 878.608,-568.5413 900.6276,-561.0499"/>
-<polygon fill="#191970" stroke="#191970" points="897.6629,-731.4953 906.4588,-737.4015 903.2659,-727.2992 897.6629,-731.4953"/>
+<!-- Node225->Node129 -->
+<g id="edge107" class="edge">
+<title>Node225->Node129</title>
+<path fill="none" stroke="#191970" d="M1773.0295,-729.99C1790.8248,-712.9454 1816.3332,-689.2592 1840,-670 1900.7521,-620.5622 1947.2975,-636.7155 1983,-567 2007.6097,-518.9453 1968.8673,-455.7852 1947.9581,-427.2085"/>
+<polygon fill="#191970" stroke="#191970" points="1770.2016,-727.8546 1765.4303,-737.3143 1775.0593,-732.8947 1770.2016,-727.8546"/>
</g>
-<!-- Node224->Node196 -->
+<!-- Node225->Node190 -->
<g id="edge109" class="edge">
-<title>Node224->Node196</title>
-<path fill="none" stroke="#191970" d="M849.802,-739.1785C783.0102,-726.1317 682.8994,-706.5766 623.7409,-695.0208"/>
-<polygon fill="#191970" stroke="#191970" points="849.3292,-742.6522 859.8147,-741.1344 850.6712,-735.7821 849.3292,-742.6522"/>
-</g>
-<!-- Node224->Node208 -->
-<g id="edge112" class="edge">
-<title>Node224->Node208</title>
-<path fill="none" stroke="#191970" d="M961.7829,-733.4516C985.7159,-723.0392 1014.9455,-710.3224 1037.2887,-700.6017"/>
-<polygon fill="#191970" stroke="#191970" points="960.334,-730.265 952.5606,-737.4639 963.1267,-736.6838 960.334,-730.265"/>
+<title>Node225->Node190</title>
+<path fill="none" stroke="#191970" d="M1728.9733,-729.838C1696.5477,-696.3921 1630.8476,-634.3934 1562,-603 1551.9445,-598.4148 1418.0003,-574.631 1340.1246,-561.0449"/>
+<polygon fill="#191970" stroke="#191970" points="1726.6722,-732.4957 1736.1147,-737.3009 1731.7297,-727.6561 1726.6722,-732.4957"/>
</g>
-<!-- Node224->Node211 -->
-<g id="edge114" class="edge">
-<title>Node224->Node211</title>
-<path fill="none" stroke="#191970" d="M986.2521,-744.3566C1057.5278,-735.0998 1164.003,-718.9299 1201,-701 1241.5216,-681.362 1252.0773,-671.5035 1277,-634 1302.437,-595.7226 1283.4277,-573.4981 1310,-536 1320.504,-521.1771 1336.4858,-508.5511 1350.0917,-499.5107"/>
-<polygon fill="#191970" stroke="#191970" points="985.4771,-740.9273 976.0039,-745.6716 986.3681,-747.8704 985.4771,-740.9273"/>
+<!-- Node225->Node196 -->
+<g id="edge110" class="edge">
+<title>Node225->Node196</title>
+<path fill="none" stroke="#191970" d="M1817.5642,-734.886C1866.5296,-722.1207 1930.8564,-705.3507 1970.5394,-695.0053"/>
+<polygon fill="#191970" stroke="#191970" points="1816.4694,-731.5543 1807.6758,-737.4639 1818.2354,-738.3279 1816.4694,-731.5543"/>
</g>
-<!-- Node224->Node213 -->
+<!-- Node225->Node208 -->
<g id="edge113" class="edge">
-<title>Node224->Node213</title>
-<path fill="none" stroke="#191970" d="M986.2213,-749.4542C1084.3351,-744.1036 1259.582,-730.6397 1315,-701 1344.1683,-685.3996 1366.9482,-652.8931 1378.5488,-633.6692"/>
-<polygon fill="#191970" stroke="#191970" points="985.9034,-745.966 976.1033,-749.9921 986.2751,-752.9562 985.9034,-745.966"/>
-</g>
-<!-- Node224->Node214 -->
-<g id="edge111" class="edge">
-<title>Node224->Node214</title>
-<path fill="none" stroke="#191970" d="M986.4176,-747.8405C1090.3492,-740.1666 1282.8158,-723.454 1348,-701 1441.3623,-668.8395 1537.3022,-597.6453 1576.5655,-566.5375"/>
-<polygon fill="#191970" stroke="#191970" points="986.0558,-744.3575 976.3374,-748.5769 986.5659,-751.3389 986.0558,-744.3575"/>
+<title>Node225->Node208</title>
+<path fill="none" stroke="#191970" d="M1681.6439,-739.8485C1613.5705,-727.2492 1510.2132,-708.1196 1446.0172,-696.238"/>
+<polygon fill="#191970" stroke="#191970" points="1681.3684,-743.3568 1691.8384,-741.7353 1682.6424,-736.4737 1681.3684,-743.3568"/>
</g>
-<!-- Node224->Node220 -->
+<!-- Node225->Node212 -->
<g id="edge115" class="edge">
-<title>Node224->Node220</title>
-<path fill="none" stroke="#191970" d="M918,-727.3415C918,-699.8131 918,-656.5714 918,-633.7614"/>
-<polygon fill="#191970" stroke="#191970" points="914.5001,-727.3889 918,-737.389 921.5001,-727.389 914.5001,-727.3889"/>
+<title>Node225->Node212</title>
+<path fill="none" stroke="#191970" d="M1761.138,-727.9907C1779.4027,-687.8336 1816.9013,-605.5319 1849,-536 1854.6228,-523.8199 1861.0448,-510.0465 1865.83,-499.8102"/>
+<polygon fill="#191970" stroke="#191970" points="1757.877,-726.7068 1756.9239,-737.2587 1764.2492,-729.6043 1757.877,-726.7068"/>
</g>
-<!-- Node224->Node168 -->
-<g id="edge110" class="edge">
-<title>Node224->Node168</title>
-<path fill="none" stroke="#191970" d="M986.3372,-746.6525C1138.1482,-733.6623 1501.2168,-702.5951 1642.9363,-690.4684"/>
-<polygon fill="#191970" stroke="#191970" points="985.8777,-743.1789 976.2126,-747.5188 986.4746,-750.1534 985.8777,-743.1789"/>
+<!-- Node225->Node214 -->
+<g id="edge114" class="edge">
+<title>Node225->Node214</title>
+<path fill="none" stroke="#191970" d="M1737.3016,-727.8496C1733.3157,-719.4752 1729.1411,-709.9662 1726,-701 1717.9601,-678.0502 1711.9757,-650.4119 1708.7371,-633.6627"/>
+<polygon fill="#191970" stroke="#191970" points="1734.2732,-729.6232 1741.8207,-737.0586 1740.5573,-726.5394 1734.2732,-729.6232"/>
+</g>
+<!-- Node225->Node215 -->
+<g id="edge112" class="edge">
+<title>Node225->Node215</title>
+<path fill="none" stroke="#191970" d="M1753.7576,-727.3244C1756.3302,-710.9363 1759.992,-689.1289 1764,-670 1767.3821,-653.8581 1770.9849,-650.3689 1773,-634 1774.6834,-620.3255 1776.4944,-616.3273 1773,-603 1769.5648,-589.8983 1761.9619,-576.6154 1755.4026,-566.8216"/>
+<polygon fill="#191970" stroke="#191970" points="1750.2954,-726.8121 1752.2317,-737.2285 1757.2137,-727.8781 1750.2954,-726.8121"/>
+</g>
+<!-- Node225->Node221 -->
+<g id="edge116" class="edge">
+<title>Node225->Node221</title>
+<path fill="none" stroke="#191970" d="M1682.0137,-740.3296C1648.0724,-732.3767 1607.2672,-719.8752 1574,-701 1542.4816,-683.117 1513.14,-652.0225 1497.2781,-633.5421"/>
+<polygon fill="#191970" stroke="#191970" points="1681.3247,-743.7621 1691.8504,-742.5536 1682.8684,-736.9344 1681.3247,-743.7621"/>
+</g>
+<!-- Node225->Node168 -->
+<g id="edge111" class="edge">
+<title>Node225->Node168</title>
+<path fill="none" stroke="#191970" d="M1818.2269,-744.5084C1931.7642,-731.2094 2157.4961,-704.7688 2263.9176,-692.3034"/>
+<polygon fill="#191970" stroke="#191970" points="1817.7847,-741.0362 1808.2598,-745.6759 1818.5991,-747.9886 1817.7847,-741.0362"/>
</g>
</g>
</svg>
diff --git a/docs/reference/api/doxygen/classes.html b/docs/reference/api/doxygen/classes.html
index 1f5fae3c9..bf6841e79 100644
--- a/docs/reference/api/doxygen/classes.html
+++ b/docs/reference/api/doxygen/classes.html
@@ -65,255 +65,255 @@ $(function() {
<div class="qindex"><a class="qindex" href="#letter_a">a</a> | <a class="qindex" href="#letter_b">b</a> | <a class="qindex" href="#letter_c">c</a> | <a class="qindex" href="#letter_d">d</a> | <a class="qindex" href="#letter_e">e</a> | <a class="qindex" href="#letter_f">f</a> | <a class="qindex" href="#letter_g">g</a> | <a class="qindex" href="#letter_h">h</a> | <a class="qindex" href="#letter_i">i</a> |& [...]
<table class="classindex">
<tr><td rowspan="2" valign="bottom"><a name="letter_a"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  a  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1CostModel.html">CostModel</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1is__specialized_3_01Container_3_01Args_8_8_8_01_4_00_01Container_01_4.html">is_specialized< Container< Args... >, Container ></a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)    [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CostModel.html">CostModel</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1IterAdapter.html">IterAdapter</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1PragmaStep.html">PragmaStep</a> [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AccessAnalyzer.html">AccessAnalyzer</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1CostModelNode.html">CostModelNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1MapNo [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AccessAnalyzerNode.html">AccessAnalyzerNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CostModelNode.html">CostModelNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runti [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AdaptivePool1DAttrs.html">AdaptivePool1DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1CountNode.html">CountNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1Iterat [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AdaptivePool2DAttrs.html">AdaptivePool2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CropAndResizeAttrs.html">CropAndResizeAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1support_1_1Span_1_1iterator__base.html">Span::ite [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1CostModel.html">CostModel</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1is__specialized_3_01Container_3_01Args_8_8_8_01_4_00_01Container_01_4.html">is_specialized< Container< Args... >, Container ></a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)    [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CostModel.html">CostModel</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1IterAdapter.html">IterAdapter</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1PragmaStep.html">PragmaStep</a> [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AccessAnalyzer.html">AccessAnalyzer</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1CostModelNode.html">CostModelNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1MapNo [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AccessAnalyzerNode.html">AccessAnalyzerNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CostModelNode.html">CostModelNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runti [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AdaptivePool1DAttrs.html">AdaptivePool1DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1CountNode.html">CountNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1Iterat [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AdaptivePool2DAttrs.html">AdaptivePool2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CropAndResizeAttrs.html">CropAndResizeAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1support_1_1Span_1_1iterator__base.html">Span::ite [...]
<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AdaptivePool3DAttrs.html">AdaptivePool3DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_d"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  d  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1IteratorNode.html">IteratorNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1PreloadMeasuredStatesNode.html">PreloadMeasuredStatesNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="clas [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Add.html">Add</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1auto__scheduler_1_1AttachMapNode_1_1IterKeyHash.html">AttachMapNode::IterKeyHash</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1PReluAttrs.html">PReluAttrs</a> [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AddNode.html">AddNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1Database.html">Database</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMapExpr.html">IterMapExpr</a> (<a class="el" href="namespacetv [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ADT.html">ADT</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1DatabaseNode.html">DatabaseNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMapExprNode.html">IterMapExprNode</a> (<a class="e [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ADTObj.html">ADTObj</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DataProducer.html">DataProducer</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMapResult.html">IterMapResult</a> (<a class="el" href="namespacetvm_1_1arith [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AffineGridAttrs.html">AffineGridAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DataProducerNode.html">DataProducerNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMapResultNode.html">IterMapResultNode</a> (<a class="el [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AffineType.html">AffineType</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1DataType.html">DataType</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMark.html">IterMark</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>) & [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AffineTypeNode.html">AffineTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DataTypePattern.html">DataTypePattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMarkNode.html">IterMarkNode</a> (<a class="el" href="namespacetvm_1_1arith.html [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AllClassNonMaximumSuppressionAttrs.html">AllClassNonMaximumSuppressionAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DataTypePatternNode.html">DataTypePatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterSplit [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Allocate.html">Allocate</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DebugAttrs.html">DebugAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterSplitExprNode.html">IterSplitExprNode</a> (<a class="el" href="namespacetvm_1_1ar [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AllocateConst.html">AllocateConst</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DeclBuffer.html">DeclBuffer</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterSumExpr.html">IterSumExpr</a> (<a class="el" href="namespacetvm_1_1arith.html" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AllocateConstNode.html">AllocateConstNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DeclBufferNode.html">DeclBufferNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterSumExprNode.html">IterSumExprNode</a> (<a class="el" href="nam [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1usmp_1_1AllocatedPoolInfo.html">AllocatedPoolInfo</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DeformableConv2DAttrs.html">DeformableConv2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IterVar.html">IterVar</a> [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1tir_1_1usmp_1_1AllocatedPoolInfoNode.html">AllocatedPoolInfoNode</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DenseAttrs.html">DenseAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1IterVarAttr.html">IterVarAttr</a> (<a c [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AllocateNode.html">AllocateNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1DenseMapNode.html">DenseMapNode</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1IterVarAttrNode.html">IterVarAttrNode</a> (<a class="el" href="namespac [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Allocator.html">Allocator</a> (<a class="el" href="namespacetvm_1_1runtime_1_1vm.html">tvm::runtime::vm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DensePackAttrs.html">DensePackAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IterVarNode.html">IterVarNode</a> (<a class="el" [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AllocStorageAttrs.html">AllocStorageAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Dependency.html">Dependency</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1IterVarRelation.html">IterVarRelation</a> (<a class="el" href="namespa [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AllocTensorAttrs.html">AllocTensorAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DependencyNode.html">DependencyNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1IterVarRelationNode.html">IterVarRelationNode</a> (<a class="el" [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1IteratorNode.html">IteratorNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1PreloadMeasuredStatesNode.html">PreloadMeasuredStatesNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="clas [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Add.html">Add</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1auto__scheduler_1_1AttachMapNode_1_1IterKeyHash.html">AttachMapNode::IterKeyHash</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1PReluAttrs.html">PReluAttrs</a> [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AddNode.html">AddNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1Database.html">Database</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMapExpr.html">IterMapExpr</a> (<a class="el" href="namespacetv [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ADT.html">ADT</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1DatabaseNode.html">DatabaseNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMapExprNode.html">IterMapExprNode</a> (<a class="e [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ADTObj.html">ADTObj</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DataProducer.html">DataProducer</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMapResult.html">IterMapResult</a> (<a class="el" href="namespacetvm_1_1arith [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AffineGridAttrs.html">AffineGridAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DataProducerNode.html">DataProducerNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMapResultNode.html">IterMapResultNode</a> (<a class="el [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AffineType.html">AffineType</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1DataType.html">DataType</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMark.html">IterMark</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>) & [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AffineTypeNode.html">AffineTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DataTypePattern.html">DataTypePattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterMarkNode.html">IterMarkNode</a> (<a class="el" href="namespacetvm_1_1arith.html [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AllClassNonMaximumSuppressionAttrs.html">AllClassNonMaximumSuppressionAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DataTypePatternNode.html">DataTypePatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterSplit [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Allocate.html">Allocate</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DebugAttrs.html">DebugAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterSplitExprNode.html">IterSplitExprNode</a> (<a class="el" href="namespacetvm_1_1ar [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AllocateConst.html">AllocateConst</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DeclBuffer.html">DeclBuffer</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterSumExpr.html">IterSumExpr</a> (<a class="el" href="namespacetvm_1_1arith.html" [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AllocateConstNode.html">AllocateConstNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DeclBufferNode.html">DeclBufferNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IterSumExprNode.html">IterSumExprNode</a> (<a class="el" href="nam [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1usmp_1_1AllocatedPoolInfo.html">AllocatedPoolInfo</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DeformableConv2DAttrs.html">DeformableConv2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IterVar.html">IterVar</a> [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1tir_1_1usmp_1_1AllocatedPoolInfoNode.html">AllocatedPoolInfoNode</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DenseAttrs.html">DenseAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1IterVarAttr.html">IterVarAttr</a> (<a c [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AllocateNode.html">AllocateNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1DenseMapNode.html">DenseMapNode</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1IterVarAttrNode.html">IterVarAttrNode</a> (<a class="el" href="namespac [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Allocator.html">Allocator</a> (<a class="el" href="namespacetvm_1_1runtime_1_1vm.html">tvm::runtime::vm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DensePackAttrs.html">DensePackAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IterVarNode.html">IterVarNode</a> (<a class="el" [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AllocStorageAttrs.html">AllocStorageAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Dependency.html">Dependency</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1IterVarRelation.html">IterVarRelation</a> (<a class="el" href="namespa [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AllocTensorAttrs.html">AllocTensorAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DependencyNode.html">DependencyNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1IterVarRelationNode.html">IterVarRelationNode</a> (<a class="el" [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1AltPattern.html">AltPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1qnn_1_1DequantizeAttrs.html">DequantizeAttrs</a> (<a class="el" href="namespacetvm_1_1relay_1_1qnn.html">tvm::relay::qnn</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_l"></a><table border="0" cellspacing="0" cellpadding="0" [...]
-</td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ProfilerNode.html">ProfilerNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1String.html">String</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td></tr>
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1AltPatternNode.html">AltPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1DeviceAPI.html">DeviceAPI</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ProgramBuilder.html">ProgramBuilder</a> (<a class="el" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1arith_1_1Analyzer.html">Analyzer</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DeviceCopyAttrs.html">DeviceCopyAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1L2NormalizeAttrs.html">L2NormalizeAttrs</a> (<a class="el" href="nam [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1And.html">And</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1DeviceWrapper.html">DeviceWrapper</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1LambdaDoc.html">LambdaDoc</a> (<a c [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AndNode.html">AndNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1profiling_1_1DeviceWrapperNode.html">DeviceWrapperNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1LambdaDocNode.html [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AnnotationStep.html">AnnotationStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPattern.html">DFPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1LayerNormAttrs.html">LayerNormAttrs</a> [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AnnotationStepNode.html">AnnotationStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternCallback.html">DFPatternCallback</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Layout.html">Layout [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Any.html">Any</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternCallbackNode.html">DFPatternCallbackNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LayoutAxis.html">LayoutAxis</a> (<a class="el" href="namespacetvm_1_1tir.htm [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AnyNode.html">AnyNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternFunctor.html">DFPatternFunctor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LayoutNode.html">LayoutNode</a> (<a class="el" href="namespacetvm_1_1tir.html" [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ArangeAttrs.html">ArangeAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternFunctor_3_01R_07const_01DFPattern_01_6n_00_01Args_8_8_8_08_4.html">DFPatternFunctor< R(const DFPattern &n, Args...)></a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a cla [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ProfilerNode.html">ProfilerNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1StridedSliceAttrs.html">StridedSliceAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td></tr>
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1AltPatternNode.html">AltPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1DeviceAPI.html">DeviceAPI</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ProgramBuilder.html">ProgramBuilder</a> (<a class="el" [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1arith_1_1Analyzer.html">Analyzer</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DeviceCopyAttrs.html">DeviceCopyAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1L2NormalizeAttrs.html">L2NormalizeAttrs</a> (<a class="el" href="nam [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1And.html">And</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1DeviceWrapper.html">DeviceWrapper</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1LambdaDoc.html">LambdaDoc</a> (<a c [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AndNode.html">AndNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1profiling_1_1DeviceWrapperNode.html">DeviceWrapperNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1LambdaDocNode.html [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AnnotationStep.html">AnnotationStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPattern.html">DFPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1LayerNormAttrs.html">LayerNormAttrs</a> [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AnnotationStepNode.html">AnnotationStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternCallback.html">DFPatternCallback</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Layout.html">Layout [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Any.html">Any</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternCallbackNode.html">DFPatternCallbackNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LayoutAxis.html">LayoutAxis</a> (<a class="el" href="namespacetvm_1_1tir.htm [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AnyNode.html">AnyNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternFunctor.html">DFPatternFunctor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LayoutNode.html">LayoutNode</a> (<a class="el" href="namespacetvm_1_1tir.html" [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ArangeAttrs.html">ArangeAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternFunctor_3_01R_07const_01DFPattern_01_6n_00_01Args_8_8_8_08_4.html">DFPatternFunctor< R(const DFPattern &n, Args...)></a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a cla [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ArgInfo.html">ArgInfo</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternNode.html">DFPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LE.html">LE</a> (<a class="el" href="namespacetvm_1_ [...]
</td></tr>
-<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ArgInfo.html">ArgInfo</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternNode.html">DFPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LE.html">LE</a> (<a class="el" href="namespacetvm_1_ [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ArgInfoNode.html">ArgInfoNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternVisitor.html">DFPatternVisitor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1LeakyReluAttrs.html">LeakyReluAttrs [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ArgReduceAttrs.html">ArgReduceAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1Diagnostic.html">Diagnostic</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1tir_1_1LENode.html">LENode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)  [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ArgsortAttrs.html">ArgsortAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticBuilder.html">DiagnosticBuilder</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Let.html">Let</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)  [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Array.html">Array</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticContext.html">DiagnosticContext</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Let.html">Let</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)  [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayAccessor.html">ArrayAccessor</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticContextNode.html">DiagnosticContextNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1LetNode.html">LetNode</a> (<a cla [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayAccessor_3_01const_01char_01_5_00_01_1_1tvm_1_1runtime_1_1String_01_4.html">ArrayAccessor< const char *, ::tvm::runtime::String ></a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticNode.html">DiagnosticNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </t [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator_1_1ArrayHandler.html">SimpleObjAllocator::ArrayHandler</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticRenderer.html">DiagnosticRenderer</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1LetPattern.html">LetPattern</a> ( [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1ArrayIndexPath.html">ArrayIndexPath</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticRendererNode.html">DiagnosticRendererNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1LetPatternNode.html">LetPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm:: [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1ArrayIndexPathNode.html">ArrayIndexPathNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DictAttrs.html">DictAttrs</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LetStmt.html">LetStmt</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td v [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayIterator.html">ArrayIterator</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DictAttrsNode.html">DictAttrsNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LetStmtNode.html">LetStmtNode</a> (<a class="el" hr [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ArrayNode.html">ArrayNode</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1DictDoc.html">DictDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1support_1_1LinearCongruentialEngine.html">LinearCongr [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AssertDoc.html">AssertDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1DictDocNode.html">DictDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AssertDocNode.html">AssertDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DilateAttrs.html">DilateAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ListDocNode.html">Lis [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ArgInfoNode.html">ArgInfoNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DFPatternVisitor.html">DFPatternVisitor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1LeakyReluAttrs.html">LeakyReluAttrs [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ArgReduceAttrs.html">ArgReduceAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1Diagnostic.html">Diagnostic</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1tir_1_1LENode.html">LENode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)  [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ArgsortAttrs.html">ArgsortAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticBuilder.html">DiagnosticBuilder</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Let.html">Let</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)  [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Array.html">Array</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticContext.html">DiagnosticContext</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Let.html">Let</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)  [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayAccessor.html">ArrayAccessor</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticContextNode.html">DiagnosticContextNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LetNode.html">LetNode</a> (<a class [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayAccessor_3_01const_01char_01_5_00_01_1_1tvm_1_1runtime_1_1String_01_4.html">ArrayAccessor< const char *, ::tvm::runtime::String ></a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticNode.html">DiagnosticNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </t [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator_1_1ArrayHandler.html">SimpleObjAllocator::ArrayHandler</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticRenderer.html">DiagnosticRenderer</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1LetPattern.html">LetPattern</a> ( [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1ArrayIndexPath.html">ArrayIndexPath</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DiagnosticRendererNode.html">DiagnosticRendererNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1LetPatternNode.html">LetPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm:: [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1ArrayIndexPathNode.html">ArrayIndexPathNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DictAttrs.html">DictAttrs</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LetStmt.html">LetStmt</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td v [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ArrayIterator.html">ArrayIterator</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1DictAttrsNode.html">DictAttrsNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LetStmtNode.html">LetStmtNode</a> (<a class="el" hr [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ArrayNode.html">ArrayNode</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1DictDoc.html">DictDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1support_1_1LinearCongruentialEngine.html">LinearCongr [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AssertDoc.html">AssertDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1DictDocNode.html">DictDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AssertDocNode.html">AssertDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DilateAttrs.html">DilateAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ListDocNode.html">Lis [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1ir__builder_1_1tir_1_1AssertFrameNode.html">AssertFrameNode</a> (<a class="el" href="namespacetvm_1_1script_1_1ir__builder_1_1tir.html">tvm::script::ir_builder::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1Dilation2DAttrs.html">Dilation2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1 [...]
-</td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1TempExprNode.html">TempExprNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td></tr>
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AssertStmt.html">AssertStmt</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Div.html">Div</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1LiteralDocNode.html">LiteralDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1pri [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AssertStmtNode.html">AssertStmtNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DivNode.html">DivNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Load.html">Load</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>) &# [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1TempExpr.html">TempExpr</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td></tr>
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AssertStmt.html">AssertStmt</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Div.html">Div</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1LiteralDocNode.html">LiteralDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1pri [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AssertStmtNode.html">AssertStmtNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1DivNode.html">DivNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Load.html">Load</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>) &# [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AssignDoc.html">AssignDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1Doc.html">Doc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LoadNode.html">LoadNode</a [...]
-</td><td valign="top"><a class="el" href="classtvm_1_1TensorAffineTypeNode.html">TensorAffineTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td></tr>
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AssignDocNode.html">AssignDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1DocNode.html">DocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1 [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AttachMap.html">AttachMap</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DominatorPattern.html">DominatorPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1LocalBuilderNode.html">Loca [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AttachMapNode.html">AttachMapNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DominatorPatternNode.html">DominatorPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1LocalRunner [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AttrAccessDoc.html">AttrAccessDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DropoutAttrs.html">DropoutAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1LocalRunnerNode.html" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AttrAccessDocNode.html">AttrAccessDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1DurationNode.html">DurationNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td valign="top"><a class="el" href="cl [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrDocEntry.html">AttrDocEntry</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DynExpandDimsAttrs.html">DynExpandDimsAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LoopRVNode.html">LoopRVNode</a> (<a class="el" href="n [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1TensorAffineType.html">TensorAffineType</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td></tr>
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AssignDocNode.html">AssignDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1DocNode.html">DocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1 [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AttachMap.html">AttachMap</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DominatorPattern.html">DominatorPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1LocalBuilderNode.html">Loca [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1AttachMapNode.html">AttachMapNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1DominatorPatternNode.html">DominatorPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1LocalRunner [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AttrAccessDoc.html">AttrAccessDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DropoutAttrs.html">DropoutAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1LocalRunnerNode.html" [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1AttrAccessDocNode.html">AttrAccessDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1DurationNode.html">DurationNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td valign="top"><a class="el" href="cl [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrDocEntry.html">AttrDocEntry</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1DynExpandDimsAttrs.html">DynExpandDimsAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LoopRVNode.html">LoopRVNode</a> (<a class="el" href="n [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrDocVisitor.html">AttrDocVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_e"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  e  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1LRNAttrs.html">LRNAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1RangeNode.html">RangeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1TensorInfoNode.html">TensorInfoNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1met [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1AttrError.html">AttrError</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LT.html">LT</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1RatioNode.html">RatioNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::p [...]
+</td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1LRNAttrs.html">LRNAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1RangeNode.html">RangeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1TensorInfoNode.html">TensorInfoNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1AttrError.html">AttrError</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LT.html">LT</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1RatioNode.html">RatioNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::p [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrExistVisitor.html">AttrExistVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1EinsumAttrs.html">EinsumAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1LTNode.html">LTNode</a> (<a class="el" href="namespacetvm_1_ [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1AttrFieldInfo.html">AttrFieldInfo</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1EnvFunc.html">EnvFunc</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_m"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  m  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="classtvm_1_1te_1_1RebaseNode.html">RebaseNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1TensorIntrinCall.html">TensorIntrinCall</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td></tr>
-<tr><td valign="top"><a class="el" href="classtvm_1_1AttrFieldInfoNode.html">AttrFieldInfoNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1EnvFuncNode.html">EnvFuncNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1RecClosure.html">RecClosure</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)  &# [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AttributeAccessPath.html">AttributeAccessPath</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1EQ.html">EQ</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Map.html">Map</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)    [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AttributeAccessPathNode.html">AttributeAccessPathNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1EQNode.html">EQNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1MapNode.html">MapNode</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runti [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1detail_1_1AttrInitEntry.html">AttrInitEntry</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1ErrorBuilder.html">ErrorBuilder</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MapValuePath.html">MapValuePath</a> (<a class="el" href="namespacetvm.html">tvm</a>)   [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrInitVisitor.html">AttrInitVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1ErrorReporter.html">ErrorReporter</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MapValuePathNode.html">MapValuePathNode</a> (<a class="el" href="namespacetvm.html">tvm</a> [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrNonDefaultVisitor.html">AttrNonDefaultVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Evaluate.html">Evaluate</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Match.html">Match</a> (<a class="el" href="namespacetvm_1_1re [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1detail_1_1AttrNopEntry.html">AttrNopEntry</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1EvaluateNode.html">EvaluateNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MatchBufferRegion.html">MatchBufferRegion</a> (<a class="el" href="names [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrNormalVisitor.html">AttrNormalVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Executable.html">Executable</a> (<a class="el" href="namespacetvm_1_1runtime_1_1vm.html">tvm::runtime::vm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MatchBufferRegionNode.html">MatchBuffer [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1AttrPattern.html">AttrPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Executor.html">Executor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1MatchNode.html">MatchNode</a> (<a class="el" href="namespacetvm_1_1relay.html" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1AttrPatternNode.html">AttrPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExecutorNode.html">ExecutorNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MatmulAttrs.html">MatmulAttrs</a> (<a class="el" href="namesp [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AttrRegistry.html">AttrRegistry</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExecutorRegEntry.html">ExecutorRegEntry</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MatrixSetDiagAttrs.html">MatrixSetDiagAttrs</a> (<a class="el" href="namespacetvm_1_ [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AttrRegistryMap.html">AttrRegistryMap</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ExpandDimsAttrs.html">ExpandDimsAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Max.html">Max</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>) [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AttrRegistryMapContainerMap.html">AttrRegistryMapContainerMap</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1tir_1_1ExprDeepEqual.html">ExprDeepEqual</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MaxNode.html">MaxNode</a> (<a class="el" href="namespacetvm_1_1tir.h [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1Attrs.html">Attrs</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ExprDoc.html">ExprDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MaxPool1DAttrs.html">MaxPool1DAttrs</a> (<a class="el" href="namespacetvm_1_1relay [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AttrsNode.html">AttrsNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ExprDocNode.html">ExprDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MaxPool2DAttrs.html">MaxPool2DAttrs</a> (<a class="el" href="names [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrsSEqualVisitor.html">AttrsSEqualVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ExprFunctor.html">ExprFunctor</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MaxPool3DAttrs.html">MaxPool3DAttrs</a> (<a class="el" href= [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrsSHashVisitor.html">AttrsSHashVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprFunctor.html">ExprFunctor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1MeasureCallback.html">MeasureCallback</a> (<a cl [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AttrStmt.html">AttrStmt</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprFunctor_3_01R_07const_01Expr_01_6n_00_01Args_8_8_8_08_4.html">ExprFunctor< R(const Expr &n, Args...)></a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__s [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AttrStmtNode.html">AttrStmtNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ExprFunctor_3_01R_07const_01PrimExpr_01_6n_00_01Args_8_8_8_08_4.html">ExprFunctor< R(const PrimExpr &n, Args...)></a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_ [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1detail_1_1AttrTriggerNonDefaultEntry.html">AttrTriggerNonDefaultEntry</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ExprMutator.html">ExprMutator</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureCallbackNode.html">MeasureC [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1AttrVisitor.html">AttrVisitor</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprMutator.html">ExprMutator</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1MeasureCandidate.html">MeasureCandidate</a> (<a class="el" href="namespacetvm_1_1meta__s [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AutoSchedulerLayoutTransformAttrs.html">AutoSchedulerLayoutTransformAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprPattern.html">ExprPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1MeasureCandidateNo [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AvgPool1DAttrs.html">AvgPool1DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprPatternNode.html">ExprPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureInput.html">MeasureInput</a> (<a class=" [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AvgPool2DAttrs.html">AvgPool2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprRewriter.html">ExprRewriter</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureInputNode.html">MeasureInputNode</a> (<a class [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AvgPool3DAttrs.html">AvgPool3DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ExprStmtDoc.html">ExprStmtDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureResult.html">M [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1te_1_1RebaseNode.html">RebaseNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1TensorIntrin.html">TensorIntrin</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td></tr>
+<tr><td valign="top"><a class="el" href="classtvm_1_1AttrFieldInfoNode.html">AttrFieldInfoNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1EnvFuncNode.html">EnvFuncNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1RecClosure.html">RecClosure</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)  &# [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AttributeAccessPath.html">AttributeAccessPath</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1EQ.html">EQ</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Map.html">Map</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)    [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AttributeAccessPathNode.html">AttributeAccessPathNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1EQNode.html">EQNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1MapNode.html">MapNode</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runti [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1detail_1_1AttrInitEntry.html">AttrInitEntry</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1ErrorBuilder.html">ErrorBuilder</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MapValuePath.html">MapValuePath</a> (<a class="el" href="namespacetvm.html">tvm</a>)   [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrInitVisitor.html">AttrInitVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1ErrorReporter.html">ErrorReporter</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MapValuePathNode.html">MapValuePathNode</a> (<a class="el" href="namespacetvm.html">tvm</a> [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrNonDefaultVisitor.html">AttrNonDefaultVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Evaluate.html">Evaluate</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Match.html">Match</a> (<a class="el" href="namespacetvm_1_1re [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1detail_1_1AttrNopEntry.html">AttrNopEntry</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1EvaluateNode.html">EvaluateNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MatchBufferRegion.html">MatchBufferRegion</a> (<a class="el" href="names [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrNormalVisitor.html">AttrNormalVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1vm_1_1Executable.html">Executable</a> (<a class="el" href="namespacetvm_1_1runtime_1_1vm.html">tvm::runtime::vm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MatchBufferRegionNode.html">MatchBuffer [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1AttrPattern.html">AttrPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Executor.html">Executor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1MatchNode.html">MatchNode</a> (<a class="el" href="namespacetvm_1_1relay.html" [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1AttrPatternNode.html">AttrPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExecutorNode.html">ExecutorNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MatmulAttrs.html">MatmulAttrs</a> (<a class="el" href="namesp [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AttrRegistry.html">AttrRegistry</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExecutorRegEntry.html">ExecutorRegEntry</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MatrixSetDiagAttrs.html">MatrixSetDiagAttrs</a> (<a class="el" href="namespacetvm_1_ [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AttrRegistryMap.html">AttrRegistryMap</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ExpandDimsAttrs.html">ExpandDimsAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Max.html">Max</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>) [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AttrRegistryMapContainerMap.html">AttrRegistryMapContainerMap</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1tir_1_1ExprDeepEqual.html">ExprDeepEqual</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MaxNode.html">MaxNode</a> (<a class="el" href="namespacetvm_1_1tir.h [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1Attrs.html">Attrs</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ExprDoc.html">ExprDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MaxPool1DAttrs.html">MaxPool1DAttrs</a> (<a class="el" href="namespacetvm_1_1relay [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AttrsNode.html">AttrsNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ExprDocNode.html">ExprDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MaxPool2DAttrs.html">MaxPool2DAttrs</a> (<a class="el" href="names [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrsSEqualVisitor.html">AttrsSEqualVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ExprFunctor.html">ExprFunctor</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MaxPool3DAttrs.html">MaxPool3DAttrs</a> (<a class="el" href= [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1detail_1_1AttrsSHashVisitor.html">AttrsSHashVisitor</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprFunctor.html">ExprFunctor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1MeasureCallback.html">MeasureCallback</a> (<a cl [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AttrStmt.html">AttrStmt</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprFunctor_3_01R_07const_01Expr_01_6n_00_01Args_8_8_8_08_4.html">ExprFunctor< R(const Expr &n, Args...)></a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__s [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1AttrStmtNode.html">AttrStmtNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ExprFunctor_3_01R_07const_01PrimExpr_01_6n_00_01Args_8_8_8_08_4.html">ExprFunctor< R(const PrimExpr &n, Args...)></a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_ [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1detail_1_1AttrTriggerNonDefaultEntry.html">AttrTriggerNonDefaultEntry</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ExprMutator.html">ExprMutator</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureCallbackNode.html">MeasureC [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1AttrVisitor.html">AttrVisitor</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprMutator.html">ExprMutator</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1MeasureCandidate.html">MeasureCandidate</a> (<a class="el" href="namespacetvm_1_1meta__s [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AutoSchedulerLayoutTransformAttrs.html">AutoSchedulerLayoutTransformAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprPattern.html">ExprPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1MeasureCandidateNo [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AvgPool1DAttrs.html">AvgPool1DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprPatternNode.html">ExprPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureInput.html">MeasureInput</a> (<a class=" [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AvgPool2DAttrs.html">AvgPool2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprRewriter.html">ExprRewriter</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureInputNode.html">MeasureInputNode</a> (<a class [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1AvgPool3DAttrs.html">AvgPool3DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ExprStmtDoc.html">ExprStmtDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureResult.html">M [...]
<tr><td rowspan="2" valign="bottom"><a name="letter_b"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  b  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ExprStmtDocNode.html">ExprStmtDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureResultNode.html">MeasureResultNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_ [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ExprVisitor.html">ExprVisitor</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MemoryInfo.html">MemoryInfo</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ReorderStep.html">ReorderStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tv [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1BaseAttrsNode.html">BaseAttrsNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprVisitor.html">ExprVisitor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MemoryInfoNode.html">MemoryInfoNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)  &# [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1te_1_1BaseComputeOpNode.html">BaseComputeOpNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1ExternOp.html">ExternOp</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1vm_1_1MemoryManager.html">MemoryManager</a> (<a class="el" href="namespacetvm_1_1r [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1BaseExpr.html">BaseExpr</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1ExternOpNode.html">ExternOpNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="structMemoryManagerInterface.html">MemoryManagerInterface</a>   </td><td valign="top"><a class="el" href="classtvm_1 [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1BaseExprNode.html">BaseExprNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ExtractedTask.html">ExtractedTask</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MeshgridAttrs.html">MeshgridAttrs</a> (<a class="el" href="names [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1BaseFunc.html">BaseFunc</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ExtractedTaskNode.html">ExtractedTaskNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1Metadata.html">Metadata</a> (<a class="el" href="na [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1BaseFuncNode.html">BaseFuncNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1PackedFuncObj_1_1Extractor.html">PackedFuncObj::Extractor</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArray.html">MetadataArray</a> (<a clas [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ExprStmtDocNode.html">ExprStmtDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1MeasureResultNode.html">MeasureResultNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_ [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ExprVisitor.html">ExprVisitor</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MemoryInfo.html">MemoryInfo</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ReorderStep.html">ReorderStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tv [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1BaseAttrsNode.html">BaseAttrsNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ExprVisitor.html">ExprVisitor</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MemoryInfoNode.html">MemoryInfoNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)  &# [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1te_1_1BaseComputeOpNode.html">BaseComputeOpNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1ExternOp.html">ExternOp</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1vm_1_1MemoryManager.html">MemoryManager</a> (<a class="el" href="namespacetvm_1_1r [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1BaseExpr.html">BaseExpr</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1ExternOpNode.html">ExternOpNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="structMemoryManagerInterface.html">MemoryManagerInterface</a>   </td><td valign="top"><a class="el" href="classtvm_1 [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1BaseExprNode.html">BaseExprNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ExtractedTask.html">ExtractedTask</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MeshgridAttrs.html">MeshgridAttrs</a> (<a class="el" href="names [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1BaseFunc.html">BaseFunc</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ExtractedTaskNode.html">ExtractedTaskNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1Metadata.html">Metadata</a> (<a class="el" href="na [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1BaseFuncNode.html">BaseFuncNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1PackedFuncObj_1_1Extractor.html">PackedFuncObj::Extractor</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArray.html">MetadataArray</a> (<a clas [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1BaseTensorType.html">BaseTensorType</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_f"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  f  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArrayNode.html">MetadataArrayNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ReshapeAttrs.html">ReshapeAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1TriluAttrs.html [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1BaseTensorTypeNode.html">BaseTensorTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBase.html">MetadataBase</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ReshapeLikeAttrs.html">ReshapeLikeAttrs< [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1BaseValueEqual.html">BaseValueEqual</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1FeatureExtractor.html">FeatureExtractor</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBaseNode.html">MetadataBaseNode</ [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1BaseValueHash.html">BaseValueHash</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1FeatureExtractorNode.html">FeatureExtractorNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1MetadataFrame.html">MetadataFrame</a> [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BatchMatmulAttrs.html">BatchMatmulAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1FeatureSet.html">FeatureSet</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1MetadataFrameNode.html">MetadataFrameNode</a> (<a c [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BatchNormAttrs.html">BatchNormAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1FIFOBufferAttrs.html">FIFOBufferAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataNode.html">MetadataNode</a> (<a cl [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BatchToSpaceNDAttrs.html">BatchToSpaceNDAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1FixedPointMultiplyAttrs.html">FixedPointMultiplyAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MetaScheduleLayoutTransformAt [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BiasAddAttrs.html">BiasAddAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1SeqStmt_1_1Flattener.html">SeqStmt::Flattener</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollector.html">MetricCollector</a> (< [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BijectiveLayout.html">BijectiveLayout</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1FloatImm.html">FloatImm</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollectorNode.html">MetricCollectorNode</a> (<a class="el" href="namespacetvm_1_ [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BijectiveLayoutNode.html">BijectiveLayoutNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1FloatImmNode.html">FloatImmNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Min.html">Min</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)  [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BinaryConv2DAttrs.html">BinaryConv2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1FloorDiv.html">FloorDiv</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MinNode.html">MinNode</a> (<a class="el" href="namespacetvm_1_1tir.html"> [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BinaryDenseAttrs.html">BinaryDenseAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1FloorDivNode.html">FloorDivNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MirrorPadAttrs.html">MirrorPadAttrs</a> (<a class="el" href="nam [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BinaryOpNode.html">BinaryOpNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1FloorMod.html">FloorMod</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MissingArrayElementPath.html">MissingArrayElementPath</a> (<a class="el" href="namespacetvm.html" [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BitPackAttrs.html">BitPackAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1FloorModNode.html">FloorModNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MissingArrayElementPathNode.html">MissingArrayElementPathNode</a> (<a class="el" h [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Block.html">Block</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FollowFusedSplitStep.html">FollowFusedSplitStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MissingMapEntryPath.html">MissingMapEntryPath</a> (< [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1ir__builder_1_1tir_1_1BlockFrame.html">BlockFrame</a> (<a class="el" href="namespacetvm_1_1script_1_1ir__builder_1_1tir.html">tvm::script::ir_builder::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FollowFusedSplitStepNode.html">FollowFusedSplitStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="t [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1ir__builder_1_1tir_1_1BlockFrameNode.html">BlockFrameNode</a> (<a class="el" href="namespacetvm_1_1script_1_1ir__builder_1_1tir.html">tvm::script::ir_builder::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FollowSplitStep.html">FollowSplitStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a cla [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1tir_1_1BlockInfo.html">BlockInfo</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FollowSplitStepNode.html">FollowSplitStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1MixedModeVisitor.html">MixedModeVisi [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockNode.html">BlockNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1For.html">For</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Mod.html">Mod</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td v [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockRealize.html">BlockRealize</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ForDoc.html">ForDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ModNode.html">ModNode</a> (<a class="el" href="namespa [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockRealizeNode.html">BlockRealizeNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ForDocNode.html">ForDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1ModularSet.html">ModularSet</a> (<a [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockRV.html">BlockRV</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ForNode.html">ForNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1ModularSetAnalyzer.html">ModularSetAnalyzer</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockRVNode.html">BlockRVNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1Frame.html">Frame</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1ModularSetNode.html">ModularSetNode</a> (<a class="el" h [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockScope.html">BlockScope</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1FrameBuffer.html">FrameBuffer</a> (<a class="el" href="namespacetvm_1_1runtime_1_1micro__rpc.html">tvm::runtime::micro_rpc</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Module.html">Module</a> (<a class [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockScopeNode.html">BlockScopeNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1FrameNode.html">FrameNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ModuleNode.html">ModuleNode</a> (<a clas [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1Bool.html">Bool</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1Framer.html">Framer</a> (<a class="el" href="namespacetvm_1_1runtime_1_1micro__rpc.html">tvm::runtime::micro_rpc</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Mul.html">Mul</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>) [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Broadcast.html">Broadcast</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ShapeTupleObj_1_1FromStd.html">ShapeTupleObj::FromStd</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MulNode.html">MulNode</a> (<a class="el" href="namespa [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1qnn_1_1BroadcastAttrs.html">BroadcastAttrs</a> (<a class="el" href="namespacetvm_1_1relay_1_1qnn.html">tvm::relay::qnn</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1StringObj_1_1FromStd.html">StringObj::FromStd</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MultiBoxPriorAttrs.html [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BroadcastNode.html">BroadcastNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Function.html">Function</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MultiBoxTransformLocAttrs.html">MultiBoxTransformLocAttrs</a> (<a class="el" hr [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1vm_1_1Buffer.html">Buffer</a> (<a class="el" href="namespacetvm_1_1runtime_1_1vm.html">tvm::runtime::vm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1FunctionDoc.html">FunctionDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MultinomialAttrs.html">M [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Buffer.html">Buffer</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1FunctionDocNode.html">FunctionDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1Mutator.html">Mutator</a> (<a class=" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1usmp_1_1BufferInfo.html">BufferInfo</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1FunctionNode.html">FunctionNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1MutatorNode.html">MutatorNode</a> (<a class="el [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataArrayNode.html">MetadataArrayNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ReshapeAttrs.html">ReshapeAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1TransposeAttrs. [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1BaseTensorTypeNode.html">BaseTensorTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBase.html">MetadataBase</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ReshapeLikeAttrs.html">ReshapeLikeAttrs< [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1BaseValueEqual.html">BaseValueEqual</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1FeatureExtractor.html">FeatureExtractor</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataBaseNode.html">MetadataBaseNode</ [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1BaseValueHash.html">BaseValueHash</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1FeatureExtractorNode.html">FeatureExtractorNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1MetadataFrame.html">MetadataFrame</a> [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BatchMatmulAttrs.html">BatchMatmulAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1FeatureSet.html">FeatureSet</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1MetadataFrameNode.html">MetadataFrameNode</a> (<a c [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BatchNormAttrs.html">BatchNormAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1FIFOBufferAttrs.html">FIFOBufferAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1MetadataNode.html">MetadataNode</a> (<a cl [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BatchToSpaceNDAttrs.html">BatchToSpaceNDAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1FixedPointMultiplyAttrs.html">FixedPointMultiplyAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MetaScheduleLayoutTransformAt [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BiasAddAttrs.html">BiasAddAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1SeqStmt_1_1Flattener.html">SeqStmt::Flattener</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollector.html">MetricCollector</a> (< [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BijectiveLayout.html">BijectiveLayout</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1FloatImm.html">FloatImm</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1profiling_1_1MetricCollectorNode.html">MetricCollectorNode</a> (<a class="el" href="namespacetvm_1_ [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BijectiveLayoutNode.html">BijectiveLayoutNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1FloatImmNode.html">FloatImmNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Min.html">Min</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)  [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BinaryConv2DAttrs.html">BinaryConv2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1FloorDiv.html">FloorDiv</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MinNode.html">MinNode</a> (<a class="el" href="namespacetvm_1_1tir.html"> [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BinaryDenseAttrs.html">BinaryDenseAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1FloorDivNode.html">FloorDivNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MirrorPadAttrs.html">MirrorPadAttrs</a> (<a class="el" href="nam [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BinaryOpNode.html">BinaryOpNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1FloorMod.html">FloorMod</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MissingArrayElementPath.html">MissingArrayElementPath</a> (<a class="el" href="namespacetvm.html" [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1BitPackAttrs.html">BitPackAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1FloorModNode.html">FloorModNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MissingArrayElementPathNode.html">MissingArrayElementPathNode</a> (<a class="el" h [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Block.html">Block</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FollowFusedSplitStep.html">FollowFusedSplitStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1MissingMapEntryPath.html">MissingMapEntryPath</a> (< [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1ir__builder_1_1tir_1_1BlockFrame.html">BlockFrame</a> (<a class="el" href="namespacetvm_1_1script_1_1ir__builder_1_1tir.html">tvm::script::ir_builder::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FollowFusedSplitStepNode.html">FollowFusedSplitStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="t [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1ir__builder_1_1tir_1_1BlockFrameNode.html">BlockFrameNode</a> (<a class="el" href="namespacetvm_1_1script_1_1ir__builder_1_1tir.html">tvm::script::ir_builder::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FollowSplitStep.html">FollowSplitStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a cla [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1tir_1_1BlockInfo.html">BlockInfo</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FollowSplitStepNode.html">FollowSplitStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1MixedModeVisitor.html">MixedModeVisi [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockNode.html">BlockNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1For.html">For</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Mod.html">Mod</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td v [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockRealize.html">BlockRealize</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ForDoc.html">ForDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ModNode.html">ModNode</a> (<a class="el" href="namespa [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockRealizeNode.html">BlockRealizeNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ForDocNode.html">ForDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1ModularSet.html">ModularSet</a> (<a [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockRV.html">BlockRV</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ForNode.html">ForNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1ModularSetAnalyzer.html">ModularSetAnalyzer</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockRVNode.html">BlockRVNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1Frame.html">Frame</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1ModularSetNode.html">ModularSetNode</a> (<a class="el" h [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockScope.html">BlockScope</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1FrameBuffer.html">FrameBuffer</a> (<a class="el" href="namespacetvm_1_1runtime_1_1micro__rpc.html">tvm::runtime::micro_rpc</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Module.html">Module</a> (<a class [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BlockScopeNode.html">BlockScopeNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1FrameNode.html">FrameNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ModuleNode.html">ModuleNode</a> (<a clas [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1Bool.html">Bool</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1Framer.html">Framer</a> (<a class="el" href="namespacetvm_1_1runtime_1_1micro__rpc.html">tvm::runtime::micro_rpc</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Mul.html">Mul</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>) [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Broadcast.html">Broadcast</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1StringObj_1_1FromStd.html">StringObj::FromStd</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1MulNode.html">MulNode</a> (<a class="el" href="namespacetvm_1_ [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1qnn_1_1BroadcastAttrs.html">BroadcastAttrs</a> (<a class="el" href="namespacetvm_1_1relay_1_1qnn.html">tvm::relay::qnn</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ShapeTupleObj_1_1FromStd.html">ShapeTupleObj::FromStd</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MultiBoxPriorAt [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BroadcastNode.html">BroadcastNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Function.html">Function</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MultiBoxTransformLocAttrs.html">MultiBoxTransformLocAttrs</a> (<a class="el" hr [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1vm_1_1Buffer.html">Buffer</a> (<a class="el" href="namespacetvm_1_1runtime_1_1vm.html">tvm::runtime::vm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1FunctionDoc.html">FunctionDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1MultinomialAttrs.html">M [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Buffer.html">Buffer</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1FunctionDocNode.html">FunctionDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1Mutator.html">Mutator</a> (<a class=" [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1usmp_1_1BufferInfo.html">BufferInfo</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1FunctionNode.html">FunctionNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1MutatorNode.html">MutatorNode</a> (<a class="el [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1usmp_1_1BufferInfoAnalysis.html">BufferInfoAnalysis</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1FunctionPattern.html">FunctionPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_n"></a><table border="0" cellspacing="0" cel [...]
-</td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1RuntimeRegEntry.html">RuntimeRegEntry</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structTVMModule.html">TVMModule</a>   </td></tr>
+</td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1RuntimeRegEntry.html">RuntimeRegEntry</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structTVMMetadata.html">TVMMetadata</a>   </td></tr>
<tr><td valign="top"><a class="el" href="structtvm_1_1tir_1_1usmp_1_1BufferInfoAnalysisNode.html">BufferInfoAnalysisNode</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1FunctionPatternNode.html">FunctionPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_s"></a><table border="0" ce [...]
-</td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1TVMMovableArgValue__.html">TVMMovableArgValue_</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td></tr>
-<tr><td valign="top"><a class="el" href="structtvm_1_1tir_1_1usmp_1_1BufferInfoNode.html">BufferInfoNode</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1FuncType.html">FuncType</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1ir__builder_1_1details_1_1Namer.html">Namer</a> (<a class="el" href="namespacet [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferLoad.html">BufferLoad</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1FuncTypeNode.html">FuncTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1NameSupply.html">NameSupply</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valig [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferLoadNode.html">BufferLoadNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1Fuse.html">Fuse</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1NameSupplyNode.html">NameSupplyNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferNode.html">BufferNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1FuseNode.html">FuseNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1NDArray.html">NDArray</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a> [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferRealize.html">BufferRealize</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FuseStep.html">FuseStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1NDArrayContainerTrait.html">NDArrayContainerTrait</a> (<a c [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferRealizeNode.html">BufferRealizeNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FuseStepNode.html">FuseStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1NdarraySizeAttrs.html">NdarraySize [...]
+</td><td valign="top"><a class="el" href="structTVMModule.html">TVMModule</a>   </td></tr>
+<tr><td valign="top"><a class="el" href="structtvm_1_1tir_1_1usmp_1_1BufferInfoNode.html">BufferInfoNode</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp.html">tvm::tir::usmp</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1FuncType.html">FuncType</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1ir__builder_1_1details_1_1Namer.html">Namer</a> (<a class="el" href="namespacet [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferLoad.html">BufferLoad</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1FuncTypeNode.html">FuncTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1NameSupply.html">NameSupply</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valig [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferLoadNode.html">BufferLoadNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1Fuse.html">Fuse</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1NameSupplyNode.html">NameSupplyNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferNode.html">BufferNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1FuseNode.html">FuseNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1NDArray.html">NDArray</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a> [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferRealize.html">BufferRealize</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FuseStep.html">FuseStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1NDArrayContainerTrait.html">NDArrayContainerTrait</a> (<a c [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferRealizeNode.html">BufferRealizeNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1FuseStepNode.html">FuseStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1NdarraySizeAttrs.html">NdarraySize [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferRegion.html">BufferRegion</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_g"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  g  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1NE.html">NE</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ScatterNDAttrs.html">ScatterNDAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1TVMRetValue.html">TVMRetValue</a> (<a class="el" href="namespacetvm_1_1runtime.html">t [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferRegionNode.html">BufferRegionNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1NENode.html">NENode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Schedule.html">Schedule</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferStore.html">BufferStore</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GatherAttrs.html">GatherAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1NLLLossAttrs.html">NLLLossAttrs</a> (<a class="el" href="namespacetvm_1_1rel [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferStoreNode.html">BufferStoreNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GatherNDAttrs.html">GatherNDAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1NodeFunctor.html">NodeFunctor</a> (<a class="el" href="namespacetvm.html"> [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1Builder.html">Builder</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1GE.html">GE</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1NodeFunctor_3_01R_07const_01ObjectRef_01_6n_00_01Args_8_8_8_08_4.html">NodeFunctor< [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderInput.html">BuilderInput</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GenericFunc.html">GenericFunc</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1NonMaximumSuppressionAttrs.html">NonMaximumSuppressionAttrs</a> (<a [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderInputNode.html">BuilderInputNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GenericFuncNode.html">GenericFuncNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1NormalAttrs.html">NormalAttrs</a> (<a class="el" hre [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderNode.html">BuilderNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1GENode.html">GENode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Not.html">Not</a> (<a class="el" href="namespacetvm_1_1tir.html" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderResult.html">BuilderResult</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GetValidCountsAttrs.html">GetValidCountsAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1NotNode.html">NotNode</a> ( [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderResultNode.html">BuilderResultNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GlobalPool2DAttrs.html">GlobalPool2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1NullOptType.html"> [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1NE.html">NE</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ScatterNDAttrs.html">ScatterNDAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1TVMPODValue__.html">TVMPODValue_</a> (<a class="el" href="namespacetvm_1_1runtime.html [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferRegionNode.html">BufferRegionNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1NENode.html">NENode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Schedule.html">Schedule</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferStore.html">BufferStore</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GatherAttrs.html">GatherAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1NLLLossAttrs.html">NLLLossAttrs</a> (<a class="el" href="namespacetvm_1_1rel [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1BufferStoreNode.html">BufferStoreNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GatherNDAttrs.html">GatherNDAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1NodeFunctor.html">NodeFunctor</a> (<a class="el" href="namespacetvm.html"> [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1Builder.html">Builder</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1GE.html">GE</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1NodeFunctor_3_01R_07const_01ObjectRef_01_6n_00_01Args_8_8_8_08_4.html">NodeFunctor< [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderInput.html">BuilderInput</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GenericFunc.html">GenericFunc</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1NonMaximumSuppressionAttrs.html">NonMaximumSuppressionAttrs</a> (<a [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderInputNode.html">BuilderInputNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GenericFuncNode.html">GenericFuncNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1NormalAttrs.html">NormalAttrs</a> (<a class="el" hre [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderNode.html">BuilderNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1GENode.html">GENode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Not.html">Not</a> (<a class="el" href="namespacetvm_1_1tir.html" [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderResult.html">BuilderResult</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GetValidCountsAttrs.html">GetValidCountsAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1NotNode.html">NotNode</a> ( [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1BuilderResultNode.html">BuilderResultNode</a> (<a class="el" href="namespacetvm_1_1meta__schedule.html">tvm::meta_schedule</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GlobalPool2DAttrs.html">GlobalPool2DAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1NullOptType.html"> [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1BuildResult.html">BuildResult</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalTypeVar.html">GlobalTypeVar</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_o"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div clas [...]
-</td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ScopeDocNode.html">ScopeDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1TypedEnvFunc.html">TypedEnvFunc</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td></tr>
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1BuildResultNode.html">BuildResultNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalTypeVarNode.html">GlobalTypeVarNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ScopedTimer.html">ScopedTimer</a> (<a c [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ScopeDocNode.html">ScopeDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1TypeDataNode.html">TypeDataNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td></tr>
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1BuildResultNode.html">BuildResultNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalTypeVarNode.html">GlobalTypeVarNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1meta__schedule_1_1ScopedTimer.html">ScopedTimer</a> (<a c [...]
<tr><td rowspan="2" valign="bottom"><a name="letter_c"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  c  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="classGlobalVar.html">GlobalVar</a>   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ObjAllocatorBase.html">ObjAllocatorBase</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1SearchCallback.html">SearchCallback</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>) & [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1GlobalVar.html">GlobalVar</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Object.html">Object</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1SearchCallbackNode.html">SearchCallbackNode</a> (<a class="el" href="namespacetvm_1_1auto__sche [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CacheReadStep.html">CacheReadStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalVarNode.html">GlobalVarNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectEqual.html">ObjectEqual</a> (<a class="el" href="na [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CacheReadStepNode.html">CacheReadStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalVarSupply.html">GlobalVarSupply</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectHash.html">ObjectHash</a> (<a class="el [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CacheWriteStep.html">CacheWriteStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalVarSupplyNode.html">GlobalVarSupplyNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1ObjectPath.html">ObjectPath</a> (<a class="el" href="na [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CacheWriteStepNode.html">CacheWriteStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1usmp_1_1algo_1_1GreedyBase.html">GreedyBase</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp_1_1algo.html">tvm::tir::usmp::algo</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1Ob [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Call.html">Call</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GridSampleAttrs.html">GridSampleAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1ObjectPathPair.html">ObjectPathPair</a> (<a class="el" href="namespacetvm.html">tvm</a>) [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Call.html">Call</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GroupNormAttrs.html">GroupNormAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1ObjectPathPairNode.html">ObjectPathPairNode</a> (<a class="el" href="namespacetvm.html"> [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1CallDoc.html">CallDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1GT.html">GT</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ObjectPtr.html">ObjectPtr</a> (<a class="el" href="namespacetvm_1_1r [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1CallDocNode.html">CallDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1GTNode.html">GTNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectPtrEqual.html">ObjectPtrEqual</a> (<a class=" [...]
+</td><td valign="top"><a class="el" href="classGlobalVar.html">GlobalVar</a>   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ObjAllocatorBase.html">ObjAllocatorBase</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1hexagon_1_1SDLTensor.html">SDLTensor</a> (<a class="el" href="namespacetvm_1_1runtime_1_1hexagon.html">tvm::runtime::hexagon</a>)  [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1GlobalVar.html">GlobalVar</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Object.html">Object</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1SearchCallback.html">SearchCallback</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.ht [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CacheReadStep.html">CacheReadStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalVarNode.html">GlobalVarNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectEqual.html">ObjectEqual</a> (<a class="el" href="na [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CacheReadStepNode.html">CacheReadStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalVarSupply.html">GlobalVarSupply</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectHash.html">ObjectHash</a> (<a class="el [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CacheWriteStep.html">CacheWriteStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1GlobalVarSupplyNode.html">GlobalVarSupplyNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1ObjectPath.html">ObjectPath</a> (<a class="el" href="na [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1CacheWriteStepNode.html">CacheWriteStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1usmp_1_1algo_1_1GreedyBase.html">GreedyBase</a> (<a class="el" href="namespacetvm_1_1tir_1_1usmp_1_1algo.html">tvm::tir::usmp::algo</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1Ob [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Call.html">Call</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GridSampleAttrs.html">GridSampleAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1ObjectPathPair.html">ObjectPathPair</a> (<a class="el" href="namespacetvm.html">tvm</a>) [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Call.html">Call</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1GroupNormAttrs.html">GroupNormAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1ObjectPathPairNode.html">ObjectPathPairNode</a> (<a class="el" href="namespacetvm.html"> [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1CallDoc.html">CallDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1GT.html">GT</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ObjectPtr.html">ObjectPtr</a> (<a class="el" href="namespacetvm_1_1r [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1CallDocNode.html">CallDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1GTNode.html">GTNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectPtrEqual.html">ObjectPtrEqual</a> (<a class=" [...]
<tr><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1profiling_1_1CallFrame.html">CallFrame</a> (<a class="el" href="namespacetvm_1_1runtime_1_1profiling.html">tvm::runtime::profiling</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_h"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  h  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectPtrHash.html">ObjectPtrHash</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Select.html">Select</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1TypeName_3_01int_01_4.html">TypeName< int ></a> (<a class="el" href= [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CallLoweredAttrs.html">CallLoweredAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html">ObjectRef</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1SelectNode.html">SelectNode</a> (<a class="el" href="namespac [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CallNode.html">CallNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator_1_1Handler.html">SimpleObjAllocator::Handler</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker.html">ObjectTypeChecker [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1CallNode.html">CallNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1SEqualReducer_1_1Handler.html">SEqualReducer::Handler</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker_3_01Array_3_01T_01_4_01_4.html">ObjectTypeChecker< Arra [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1CallPattern.html">CallPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1SHashReducer_1_1Handler.html">SHashReducer::Handler</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker_3_01Map_3_01K_00_01V_01_4_01_4.html">ObjectTypeChecker [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1CallPatternNode.html">CallPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structdmlc_1_1serializer_1_1Handler_3_01DLDataType_01_4.html">Handler< DLDataType ></a> (<a class="el" href="namespacedmlc_1_1serializer.html">dmlc::serializer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1OnDeviceAttr [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1arith_1_1CanonicalSimplifier.html">CanonicalSimplifier</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="structdmlc_1_1serializer_1_1Handler_3_01DLDevice_01_4.html">Handler< DLDevice ></a> (<a class="el" href="namespacedmlc_1_1serializer.html">dmlc::serializer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1OneHotAt [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Cast.html">Cast</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1HardwareParams.html">HardwareParams</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1Op.html">Op</a> (<a class="el" href="namespacetvm.html">tvm</a>)&# [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CastAttrs.html">CastAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1HardwareParamsNode.html">HardwareParamsNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1OpAttrMap.html">OpAttrMap</a> (<a class=" [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CastHintAttrs.html">CastHintAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1HybridOp.html">HybridOp</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1Operation.html">Operation</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</ [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CastNode.html">CastNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1HybridOpNode.html">HybridOpNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1OperationDoc.html">OperationDoc</a> (<a class="el" href="namespacetvm_1_1script [...]
+</td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectPtrHash.html">ObjectPtrHash</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1SearchTaskNode.html">SearchTaskNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1TypeName_3_01double_01_ [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CallLoweredAttrs.html">CallLoweredAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ObjectRef.html">ObjectRef</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Select.html">Select</a> (<a class="el" href="namespacetvm_1_1 [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CallNode.html">CallNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1SimpleObjAllocator_1_1Handler.html">SimpleObjAllocator::Handler</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker.html">ObjectTypeChecker [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1CallNode.html">CallNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1SHashReducer_1_1Handler.html">SHashReducer::Handler</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker_3_01Array_3_01T_01_4_01_4.html">ObjectTypeChecker< Array& [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1CallPattern.html">CallPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1SEqualReducer_1_1Handler.html">SEqualReducer::Handler</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1ObjectTypeChecker_3_01Map_3_01K_00_01V_01_4_01_4.html">ObjectTypeCheck [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1CallPatternNode.html">CallPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="structdmlc_1_1serializer_1_1Handler_3_01DLDataType_01_4.html">Handler< DLDataType ></a> (<a class="el" href="namespacedmlc_1_1serializer.html">dmlc::serializer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1OnDeviceAttr [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1arith_1_1CanonicalSimplifier.html">CanonicalSimplifier</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="structdmlc_1_1serializer_1_1Handler_3_01DLDevice_01_4.html">Handler< DLDevice ></a> (<a class="el" href="namespacedmlc_1_1serializer.html">dmlc::serializer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1OneHotAt [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Cast.html">Cast</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1HardwareParams.html">HardwareParams</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1Op.html">Op</a> (<a class="el" href="namespacetvm.html">tvm</a>)&# [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CastAttrs.html">CastAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1HardwareParamsNode.html">HardwareParamsNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1OpAttrMap.html">OpAttrMap</a> (<a class=" [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CastHintAttrs.html">CastHintAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1HybridOp.html">HybridOp</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1Operation.html">Operation</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</ [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CastNode.html">CastNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1HybridOpNode.html">HybridOpNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1OperationDoc.html">OperationDoc</a> (<a class="el" href="namespacetvm_1_1script [...]
<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ClassDoc.html">ClassDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_i"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  i  </div></td></tr></table>
-</td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1OperationDocNode.html">OperationDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1SequenceMaskAttrs.html">SequenceMaskAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1TypeVar.html">TypeVar</ [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ClassDocNode.html">ClassDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1OperationNode.html">OperationNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1Sequential.html">Sequential</a> (<a c [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Clause.html">Clause</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Id.html">Id</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpImplementation.html">OpImplementation</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::re [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ClauseNode.html">ClauseNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IdDoc.html">IdDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpImplementationNode.html">OpImplementationNode</a> [...]
-</td></tr>
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ClipAttrs.html">ClipAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IdDocNode.html">IdDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1OpNode.html">OpNode</a> (<a class="el" href="namespace [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Closure.html">Closure</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1IdNode.html">IdNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1OpRegEntry.html">OpRegEntry</a> (<a class="el" href="namespacetvm.html">tvm</a>)  & [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ClosureObj.html">ClosureObj</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1If.html">If</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpSpecialization.html">OpSpecialization</a> (<a class="el" href="namespacetvm_1_1relay [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CmpOpNode.html">CmpOpNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IfDoc.html">IfDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpSpecializationNode.html">OpSpecializationNode</a> (<a clas [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CommReducer.html">CommReducer</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IfDocNode.html">IfDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpStrategy.html">OpStrategy</a> (<a class="el" h [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CommReducerNode.html">CommReducerNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1IfNode.html">IfNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpStrategyNode.html">OpStrategyNode</a> (<a class="el" href="namespacetvm_1_1rel [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1CompilationConfig.html">CompilationConfig</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1IfPattern.html">IfPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Optional.html">Optional</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::run [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1CompilationConfigNode.html">CompilationConfigNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1IfPatternNode.html">IfPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Or.html">Or</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a> [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1CompileError.html">CompileError</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IfThenElse.html">IfThenElse</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1OrNode.html">OrNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </t [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1OperationDocNode.html">OperationDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1SEqualReducer.html">SEqualReducer</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1TypeReporterNode.html">TypeReporterNode</a> (<a class="el [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1ClassDocNode.html">ClassDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1te_1_1OperationNode.html">OperationNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1SequenceMaskAttrs.html">SequenceMaskAttr [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Clause.html">Clause</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Id.html">Id</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpImplementation.html">OpImplementation</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::re [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ClauseNode.html">ClauseNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IdDoc.html">IdDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpImplementationNode.html">OpImplementationNode</a> [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ClipAttrs.html">ClipAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IdDocNode.html">IdDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1OpNode.html">OpNode</a> (<a class="el" href="namespace [...]
</td></tr>
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Closure.html">Closure</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1IdNode.html">IdNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1OpRegEntry.html">OpRegEntry</a> (<a class="el" href="namespacetvm.html">tvm</a>)  & [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1ClosureObj.html">ClosureObj</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1If.html">If</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpSpecialization.html">OpSpecialization</a> (<a class="el" href="namespacetvm_1_1relay [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CmpOpNode.html">CmpOpNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IfDoc.html">IfDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpSpecializationNode.html">OpSpecializationNode</a> (<a clas [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CommReducer.html">CommReducer</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IfDocNode.html">IfDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpStrategy.html">OpStrategy</a> (<a class="el" h [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1tir_1_1CommReducerNode.html">CommReducerNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1IfNode.html">IfNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1OpStrategyNode.html">OpStrategyNode</a> (<a class="el" href="namespacetvm_1_1rel [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1CompilationConfig.html">CompilationConfig</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1IfPattern.html">IfPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1Optional.html">Optional</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::run [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1CompilationConfigNode.html">CompilationConfigNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1IfPatternNode.html">IfPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Or.html">Or</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a> [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1CompileError.html">CompileError</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IfThenElse.html">IfThenElse</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1OrNode.html">OrNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </t [...]
<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1CompilerAttrs.html">CompilerAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IfThenElseNode.html">IfThenElseNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_p"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class=" [...]
-</td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ShuffleNode.html">ShuffleNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td></tr>
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeAtStep.html">ComputeAtStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplSEqualReduce.html">ImplSEqualReduce</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1SignaturePrinter.html" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeAtStepNode.html">ComputeAtStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplSEqualReduce_3_01T_00_01true_01_4.html">ImplSEqualReduce< T, true ></a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href=" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeDAG.html">ComputeDAG</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplSHashReduce.html">ImplSHashReduce</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1PackedFuncObj.html">PackedFuncO [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeDAGNode.html">ComputeDAGNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplSHashReduce_3_01T_00_01true_01_4.html">ImplSHashReduce< T, true ></a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeInlineStep.html">ComputeInlineStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplVisitAttrs.html">ImplVisitAttrs</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueCon [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeInlineStepNode.html">ComputeInlineStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplVisitAttrs_3_01T_00_01true_01_4.html">ImplVisitAttrs< T, true ></a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" hr [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1te_1_1ComputeOp.html">ComputeOp</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1IncompleteType.html">IncompleteType</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01PrimExpr_01_4.html">PackedFuncValueConverter< PrimExpr ></a> (<a c [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1te_1_1ComputeOpNode.html">ComputeOpNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1IncompleteTypeNode.html">IncompleteTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01tvm_1_1Bool_01_4.html">PackedFuncValueConverter< tvm: [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeRootStep.html">ComputeRootStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IndexDoc.html">IndexDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1Pack [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeRootStepNode.html">ComputeRootStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IndexDocNode.html">IndexDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_ [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ConcatenateAttrs.html">ConcatenateAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IndexMap.html">IndexMap</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1PacketFieldSizeBytes.html">PacketFieldSizeBytes</a> (<a c [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Constant.html">Constant</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IndexMapNode.html">IndexMapNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1PadAttrs.html">PadAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm: [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1ConstantInfo.html">ConstantInfo</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1InitOpAttrs.html">InitOpAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1Pass.html">Pass</a> (<a class="el" href="namespacetvm_1_1transform.html">tvm::transform</a [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ConstantInfoMetadata.html">ConstantInfoMetadata</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1InplaceArrayBase.html">InplaceArrayBase</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transfo [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ConstantInfoMetadataNode.html">ConstantInfoMetadataNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1InstanceNormAttrs.html">InstanceNormAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tr [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1ConstantInfoNode.html">ConstantInfoNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Instruction.html">Instruction</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1PassInfo.html">PassInfo</a> (<a class="el" href="namespacetvm_1_1transform.html">tvm::tr [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1ConstantMemoryPools.html">ConstantMemoryPools</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1vm_1_1Instruction.html">Instruction</a> (<a class="el" href="namespacetvm_1_1runtime_1_1vm.html">tvm::runtime::vm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1PassInfoNode.html">PassInfoNode</a> (<a class="el" hre [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1ConstantMemoryPoolsNode.html">ConstantMemoryPoolsNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1InstructionKind.html">InstructionKind</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1instrument_1_1PassInstrument.html">PassInstrument</a> (<a class="el" href="name [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ConstantNode.html">ConstantNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1InstructionKindNode.html">InstructionKindNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1instrument_1_1PassInstrumentNode.html">PassInstrumentNode</a> (<a cla [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ConstantPattern.html">ConstantPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1InstructionKindRegEntry.html">InstructionKindRegEntry</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1PassNode.html">PassNode</a> (<a class="el" [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ConstantPatternNode.html">ConstantPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1InstructionNode.html">InstructionNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Pattern.html">Pattern</a> (<a class="el" href="namespa [...]
-<tr><td valign="top"><a class="el" href="classtvm_1_1ConstantPoolInfo.html">ConstantPoolInfo</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IntConstraints.html">IntConstraints</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1PatternConstructor.html">PatternConstructor</a> (<a class="el" href="namespacetvm [...]
-<tr><td valign="top"><a class="el" href="structtvm_1_1ConstantPoolInfoNode.html">ConstantPoolInfoNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IntConstraintsNode.html">IntConstraintsNode</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1PatternConstructorNode.html">PatternConstructorNode</a> (<a clas [...]
+</td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Shuffle.html">Shuffle</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td rowspan="2" valign="bottom"><a name="letter_v"></a><table border="0" cellspacing="0" cellpadding="0"><tr><td><div class="ah">  v  </div></td></tr></table>
+</td></tr>
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeAtStep.html">ComputeAtStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplSEqualReduce.html">ImplSEqualReduce</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1ShuffleNode.html">ShuffleNo [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeAtStepNode.html">ComputeAtStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplSEqualReduce_3_01T_00_01true_01_4.html">ImplSEqualReduce< T, true ></a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href=" [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeDAG.html">ComputeDAG</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplSHashReduce.html">ImplSHashReduce</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1PackedFuncObj.html">PackedFuncO [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeDAGNode.html">ComputeDAGNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplSHashReduce_3_01T_00_01true_01_4.html">ImplSHashReduce< T, true ></a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="classtvm [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeInlineStep.html">ComputeInlineStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplVisitAttrs.html">ImplVisitAttrs</a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueCon [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeInlineStepNode.html">ComputeInlineStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1detail_1_1ImplVisitAttrs_3_01T_00_01true_01_4.html">ImplVisitAttrs< T, true ></a> (<a class="el" href="namespacetvm_1_1detail.html">tvm::detail</a>)   </td><td valign="top"><a class="el" hr [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1te_1_1ComputeOp.html">ComputeOp</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1IncompleteType.html">IncompleteType</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01PrimExpr_01_4.html">PackedFuncValueConverter< PrimExpr ></a> (<a c [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1te_1_1ComputeOpNode.html">ComputeOpNode</a> (<a class="el" href="namespacetvm_1_1te.html">tvm::te</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1IncompleteTypeNode.html">IncompleteTypeNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1PackedFuncValueConverter_3_01tvm_1_1Bool_01_4.html">PackedFuncValueConverter< tvm: [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeRootStep.html">ComputeRootStep</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IndexDoc.html">IndexDoc</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1Pack [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1auto__scheduler_1_1ComputeRootStepNode.html">ComputeRootStepNode</a> (<a class="el" href="namespacetvm_1_1auto__scheduler.html">tvm::auto_scheduler</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1script_1_1printer_1_1IndexDocNode.html">IndexDocNode</a> (<a class="el" href="namespacetvm_1_1script_1_1printer.html">tvm::script::printer</a>)   </td><td valign="top"><a class="el" href="structtvm_1_ [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1relay_1_1ConcatenateAttrs.html">ConcatenateAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IndexMap.html">IndexMap</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1micro__rpc_1_1PacketFieldSizeBytes.html">PacketFieldSizeBytes</a> (<a c [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Constant.html">Constant</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1IndexMapNode.html">IndexMapNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1PadAttrs.html">PadAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm: [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1ConstantInfo.html">ConstantInfo</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1InitOpAttrs.html">InitOpAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1Pass.html">Pass</a> (<a class="el" href="namespacetvm_1_1transform.html">tvm::transform</a [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ConstantInfoMetadata.html">ConstantInfoMetadata</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1InplaceArrayBase.html">InplaceArrayBase</a> (<a class="el" href="namespacetvm_1_1runtime.html">tvm::runtime</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transfo [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1runtime_1_1metadata_1_1ConstantInfoMetadataNode.html">ConstantInfoMetadataNode</a> (<a class="el" href="namespacetvm_1_1runtime_1_1metadata.html">tvm::runtime::metadata</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1relay_1_1InstanceNormAttrs.html">InstanceNormAttrs</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tr [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1ConstantInfoNode.html">ConstantInfoNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1Instruction.html">Instruction</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1PassInfo.html">PassInfo</a> (<a class="el" href="namespacetvm_1_1transform.html">tvm::tr [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1ConstantMemoryPools.html">ConstantMemoryPools</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="structtvm_1_1runtime_1_1vm_1_1Instruction.html">Instruction</a> (<a class="el" href="namespacetvm_1_1runtime_1_1vm.html">tvm::runtime::vm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1PassInfoNode.html">PassInfoNode</a> (<a class="el" hre [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1ConstantMemoryPoolsNode.html">ConstantMemoryPoolsNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1InstructionKind.html">InstructionKind</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1instrument_1_1PassInstrument.html">PassInstrument</a> (<a class="el" href="name [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ConstantNode.html">ConstantNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1InstructionKindNode.html">InstructionKindNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1instrument_1_1PassInstrumentNode.html">PassInstrumentNode</a> (<a cla [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ConstantPattern.html">ConstantPattern</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1InstructionKindRegEntry.html">InstructionKindRegEntry</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1transform_1_1PassNode.html">PassNode</a> (<a class="el" [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1relay_1_1ConstantPatternNode.html">ConstantPatternNode</a> (<a class="el" href="namespacetvm_1_1relay.html">tvm::relay</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1tir_1_1InstructionNode.html">InstructionNode</a> (<a class="el" href="namespacetvm_1_1tir.html">tvm::tir</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1Pattern.html">Pattern</a> (<a class="el" href="namespa [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1ConstantPoolInfo.html">ConstantPoolInfo</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IntConstraints.html">IntConstraints</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1PatternConstructor.html">PatternConstructor</a> (<a class="el" href="namespacetvm [...]
+<tr><td valign="top"><a class="el" href="structtvm_1_1ConstantPoolInfoNode.html">ConstantPoolInfoNode</a> (<a class="el" href="namespacetvm.html">tvm</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IntConstraintsNode.html">IntConstraintsNode</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1PatternConstructorNode.html">PatternConstructorNode</a> (<a clas [...]
+<tr><td valign="top"><a class="el" href="classtvm_1_1arith_1_1ConstIntBound.html">ConstIntBound</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1arith_1_1IntConstraintsTransform.html">IntConstraintsTransform</a> (<a class="el" href="namespacetvm_1_1arith.html">tvm::arith</a>)   </td><td valign="top"><a class="el" href="classtvm_1_1relay_1_1PatternFunctor.html">PatternFunctor</a> (<a [...]
</td></tr>
... 62645 lines suppressed ...