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 2022/08/17 07:02:03 UTC
[tvm-site] branch asf-site updated: deploying docs (apache/tvm@d2f9f254d275df256dbcbc5a9f8b3a07cee1d81f)
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 54092cb98 deploying docs (apache/tvm@d2f9f254d275df256dbcbc5a9f8b3a07cee1d81f)
54092cb98 is described below
commit 54092cb98d34b07b5ce48b31c6e3b64d40463d9e
Author: tvm-bot <95...@users.noreply.github.com>
AuthorDate: Wed Aug 17 07:01:53 2022 +0000
deploying docs (apache/tvm@d2f9f254d275df256dbcbc5a9f8b3a07cee1d81f)
---
.../how_to/compile_models/from_darknet.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 | 4 +-
.../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 | 1469 +++---------
.../tune_network_cuda.rst.txt | 2 +-
.../tune_network_x86.rst.txt | 4 +-
.../tune_sparse_x86.rst.txt | 86 +-
.../tune_with_autotvm/sg_execution_times.rst.txt | 6 +-
.../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 | 14 +-
.../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 | 4 +-
.../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 | 9 +-
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 | 22 +-
.../tutorial/tensor_expr_get_started.rst.txt | 49 +-
docs/commit_hash | 2 +-
docs/how_to/compile_models/from_darknet.html | 2 +-
docs/how_to/compile_models/from_mxnet.html | 2 +-
docs/how_to/compile_models/from_oneflow.html | 19 +-
docs/how_to/compile_models/from_pytorch.html | 17 +-
docs/how_to/compile_models/from_tensorflow.html | 2 +-
docs/how_to/compile_models/sg_execution_times.html | 22 +-
.../deploy_models/deploy_model_on_android.html | 2 +-
.../deploy_object_detection_pytorch.html | 80 +-
docs/how_to/deploy_models/deploy_prequantized.html | 12 +-
.../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 | 40 +-
docs/how_to/deploy_models/sg_execution_times.html | 18 +-
.../extend_tvm/bring_your_own_datatypes.html | 4 +-
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 | 18 +-
.../tune_conv2d_layer_cuda.html | 1469 +++---------
.../tune_with_autoscheduler/tune_network_cuda.html | 2 +-
.../tune_with_autoscheduler/tune_network_x86.html | 4 +-
.../tune_with_autoscheduler/tune_sparse_x86.html | 86 +-
.../tune_with_autotvm/sg_execution_times.html | 6 +-
.../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 | 14 +-
.../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 +-
.../api/doxygen/affine__type_8h__incl.svg | 1504 ++++++------
docs/reference/api/doxygen/algorithm_8h__incl.svg | 766 +++---
docs/reference/api/doxygen/algorithms_8h.html | 2 +-
docs/reference/api/doxygen/algorithms_8h__incl.svg | 1680 +++++++-------
docs/reference/api/doxygen/analyzer_8h.html | 2 +-
.../api/doxygen/analyzer_8h__dep__incl.svg | 644 +++---
docs/reference/api/doxygen/analyzer_8h__incl.svg | 1655 +++++++------
docs/reference/api/doxygen/annotation_8h.html | 2 +-
docs/reference/api/doxygen/annotation_8h__incl.svg | 1386 +++++------
.../api/doxygen/apply__history__best_8h.html | 2 +-
.../api/doxygen/apply__history__best_8h__incl.svg | 2139 +++++++++--------
.../doxygen/apply__history__best_8h_source.html | 2 +-
docs/reference/api/doxygen/arg__info_8h.html | 2 +-
.../api/doxygen/arg__info_8h__dep__incl.svg | 140 +-
docs/reference/api/doxygen/arg__info_8h__incl.svg | 1826 ++++++++-------
docs/reference/api/doxygen/array__utils_8h.html | 2 +-
.../api/doxygen/array__utils_8h__dep__incl.svg | 44 +-
.../api/doxygen/array__utils_8h__incl.svg | 1748 +++++++-------
.../api/doxygen/attr__registry__map_8h.html | 2 +-
.../doxygen/attr__registry__map_8h__dep__incl.svg | 536 ++---
.../api/doxygen/attr__registry__map_8h__incl.svg | 290 +--
.../api/doxygen/attr__registry__map_8h_source.html | 2 +-
.../api/doxygen/auto__schedule_8h__incl.svg | 598 ++---
.../auto__scheduler_2cost__model_8h__incl.svg | 552 ++---
.../doxygen/auto__scheduler_2feature_8h__incl.svg | 690 +++---
docs/reference/api/doxygen/autodiff_8h.html | 2 +-
docs/reference/api/doxygen/autodiff_8h__incl.svg | 1768 +++++++-------
docs/reference/api/doxygen/bias__add_8h__incl.svg | 770 +++----
docs/reference/api/doxygen/bitserial_8h.html | 2 +-
docs/reference/api/doxygen/bitserial_8h__incl.svg | 1670 +++++++-------
docs/reference/api/doxygen/block__scope_8h.html | 2 +-
.../api/doxygen/block__scope_8h__dep__incl.svg | 152 +-
.../api/doxygen/block__scope_8h__incl.svg | 1584 ++++++-------
docs/reference/api/doxygen/bound_8h.html | 2 +-
docs/reference/api/doxygen/bound_8h__dep__incl.svg | 628 ++---
docs/reference/api/doxygen/bound_8h__incl.svg | 1763 +++++++-------
.../api/doxygen/broadcast_8h__dep__incl.svg | 92 +-
docs/reference/api/doxygen/broadcast_8h__incl.svg | 786 +++----
docs/reference/api/doxygen/buffer_8h.html | 2 +-
.../reference/api/doxygen/buffer_8h__dep__incl.svg | 640 +++---
docs/reference/api/doxygen/buffer_8h__incl.svg | 1458 ++++++------
docs/reference/api/doxygen/buffer_8h_source.html | 2 +-
docs/reference/api/doxygen/builder_8h.html | 2 +-
.../api/doxygen/builder_8h__dep__incl.svg | 40 +-
docs/reference/api/doxygen/builder_8h__incl.svg | 1960 ++++++++--------
.../api/doxygen/builtin_8h__dep__incl.svg | 108 +-
docs/reference/api/doxygen/builtin_8h__incl.svg | 530 ++---
docs/reference/api/doxygen/call_8h.html | 2 +-
docs/reference/api/doxygen/call_8h__incl.svg | 1386 +++++------
docs/reference/api/doxygen/codegen_8h.html | 2 +-
docs/reference/api/doxygen/codegen_8h__incl.svg | 1870 +++++++--------
.../api/doxygen/compilation__config_8h.html | 2 +-
.../doxygen/compilation__config_8h__dep__incl.svg | 20 +-
.../api/doxygen/compilation__config_8h__incl.svg | 1843 ++++++++-------
.../api/doxygen/compute__dag_8h__dep__incl.svg | 76 +-
.../api/doxygen/compute__dag_8h__incl.svg | 610 ++---
.../api/doxygen/compute__dag_8h_source.html | 2 +-
.../api/doxygen/constant__utils_8h__dep__incl.svg | 168 +-
.../api/doxygen/constant__utils_8h__incl.svg | 834 +++----
docs/reference/api/doxygen/cublas_8h.html | 2 +-
.../reference/api/doxygen/cublas_8h__dep__incl.svg | 20 +-
docs/reference/api/doxygen/cublas_8h__incl.svg | 2154 +++++++++--------
.../api/doxygen/cuda_2dense_8h__dep__incl.svg | 12 +-
.../reference/api/doxygen/cuda_2dense_8h__incl.svg | 734 +++---
.../api/doxygen/cuda_2injective_8h__dep__incl.svg | 12 +-
.../api/doxygen/cuda_2injective_8h__incl.svg | 758 +++---
.../api/doxygen/cuda_2pooling_8h__dep__incl.svg | 12 +-
.../api/doxygen/cuda_2pooling_8h__incl.svg | 730 +++---
.../api/doxygen/cuda_2reduction_8h__dep__incl.svg | 12 +-
.../api/doxygen/cuda_2reduction_8h__incl.svg | 758 +++---
.../api/doxygen/cuda_2softmax_8h__dep__incl.svg | 12 +-
.../api/doxygen/cuda_2softmax_8h__incl.svg | 758 +++---
docs/reference/api/doxygen/data__layout_8h.html | 2 +-
.../api/doxygen/data__layout_8h__dep__incl.svg | 112 +-
.../api/doxygen/data__layout_8h__incl.svg | 1753 +++++++-------
.../api/doxygen/data__layout_8h_source.html | 2 +-
docs/reference/api/doxygen/database_8h.html | 2 +-
.../api/doxygen/database_8h__dep__incl.svg | 56 +-
docs/reference/api/doxygen/database_8h__incl.svg | 1990 ++++++++--------
docs/reference/api/doxygen/database_8h_source.html | 2 +-
.../api/doxygen/dataflow__matcher_8h__incl.svg | 794 +++----
.../doxygen/dataflow__pattern_8h__dep__incl.svg | 24 +-
.../api/doxygen/dataflow__pattern_8h__incl.svg | 822 +++----
.../api/doxygen/dataflow__pattern_8h_source.html | 2 +-
.../dataflow__pattern__functor_8h__dep__incl.svg | 12 +-
.../dataflow__pattern__functor_8h__incl.svg | 798 +++----
.../doxygen/detail_2broadcast_8h__dep__incl.svg | 104 +-
.../api/doxygen/detail_2broadcast_8h__incl.svg | 686 +++---
.../api/doxygen/detail_2extern_8h__dep__incl.svg | 40 +-
.../api/doxygen/detail_2extern_8h__incl.svg | 642 +++---
docs/reference/api/doxygen/device__api_8h.html | 2 +-
.../api/doxygen/device__api_8h__dep__incl.svg | 64 +-
.../reference/api/doxygen/device__api_8h__incl.svg | 936 ++++----
docs/reference/api/doxygen/device__copy_8h.html | 2 +-
.../api/doxygen/device__copy_8h__incl.svg | 1964 ++++++++--------
docs/reference/api/doxygen/diagnostic_8h.html | 2 +-
.../api/doxygen/diagnostic_8h__dep__incl.svg | 532 ++---
docs/reference/api/doxygen/diagnostic_8h__incl.svg | 1644 ++++++-------
.../api/doxygen/diagnostic_8h_source.html | 2 +-
docs/reference/api/doxygen/dilate_8h__incl.svg | 604 ++---
docs/reference/api/doxygen/doc_8h.html | 2 +-
docs/reference/api/doxygen/doc_8h__dep__incl.svg | 48 +-
docs/reference/api/doxygen/doc_8h__incl.svg | 1516 ++++++------
docs/reference/api/doxygen/doc_8h_source.html | 2 +-
docs/reference/api/doxygen/doc__printer_8h.html | 2 +-
.../api/doxygen/doc__printer_8h__incl.svg | 1422 ++++++------
.../reference/api/doxygen/driver__api_8h__incl.svg | 886 +++----
docs/reference/api/doxygen/einsum_8h__incl.svg | 762 +++---
.../api/doxygen/elemwise_8h__dep__incl.svg | 48 +-
docs/reference/api/doxygen/elemwise_8h__incl.svg | 448 ++--
docs/reference/api/doxygen/env__func_8h.html | 2 +-
.../api/doxygen/env__func_8h__dep__incl.svg | 548 ++---
docs/reference/api/doxygen/env__func_8h__incl.svg | 1226 +++++-----
.../reference/api/doxygen/env__func_8h_source.html | 2 +-
docs/reference/api/doxygen/error_8h.html | 2 +-
docs/reference/api/doxygen/error_8h__dep__incl.svg | 520 ++---
docs/reference/api/doxygen/error_8h__incl.svg | 2191 +++++++++---------
docs/reference/api/doxygen/executable_8h.html | 2 +-
.../api/doxygen/executable_8h__dep__incl.svg | 12 +-
docs/reference/api/doxygen/executable_8h__incl.svg | 1098 ++++-----
docs/reference/api/doxygen/executor_8h.html | 2 +-
docs/reference/api/doxygen/executor_8h__incl.svg | 1899 ++++++++-------
docs/reference/api/doxygen/executor_8h_source.html | 2 +-
docs/reference/api/doxygen/extracted__task_8h.html | 2 +-
.../api/doxygen/extracted__task_8h__incl.svg | 1718 +++++++-------
.../api/doxygen/extracted__task_8h_source.html | 2 +-
.../api/doxygen/feature__extractor_8h.html | 2 +-
.../api/doxygen/feature__extractor_8h__incl.svg | 1462 ++++++------
docs/reference/api/doxygen/flatten_8h__incl.svg | 690 +++---
docs/reference/api/doxygen/frame_8h.html | 2 +-
docs/reference/api/doxygen/frame_8h__dep__incl.svg | 24 +-
docs/reference/api/doxygen/frame_8h__incl.svg | 1498 ++++++------
docs/reference/api/doxygen/fuse_8h.html | 2 +-
docs/reference/api/doxygen/fuse_8h__dep__incl.svg | 156 +-
docs/reference/api/doxygen/fuse_8h__incl.svg | 1748 +++++++-------
.../api/doxygen/generic_2default_8h__incl.svg | 758 +++---
.../api/doxygen/generic_2extern_8h__dep__incl.svg | 24 +-
.../api/doxygen/generic_2extern_8h__incl.svg | 746 +++---
.../doxygen/generic_2injective_8h__dep__incl.svg | 32 +-
.../api/doxygen/generic_2injective_8h__incl.svg | 758 +++---
.../api/doxygen/generic__func_8h__dep__incl.svg | 196 +-
.../api/doxygen/generic__func_8h__incl.svg | 606 ++---
.../api/doxygen/global__var__supply_8h.html | 2 +-
.../doxygen/global__var__supply_8h__dep__incl.svg | 12 +-
.../api/doxygen/global__var__supply_8h__incl.svg | 1696 +++++++-------
.../api/doxygen/global__var__supply_8h_source.html | 2 +-
docs/reference/api/doxygen/globals_defs.html | 6 -
docs/reference/api/doxygen/globals_t.html | 6 -
docs/reference/api/doxygen/greedy_8h__incl.svg | 682 +++---
docs/reference/api/doxygen/image_8h.html | 2 +-
docs/reference/api/doxygen/image_8h__incl.svg | 1670 +++++++-------
docs/reference/api/doxygen/image_8h_source.html | 2 +-
docs/reference/api/doxygen/index__map_8h.html | 2 +-
.../api/doxygen/index__map_8h__dep__incl.svg | 592 ++---
docs/reference/api/doxygen/index__map_8h__incl.svg | 1460 ++++++------
.../api/doxygen/index__map_8h_source.html | 2 +-
docs/reference/api/doxygen/instruction_8h.html | 2 +-
.../api/doxygen/instruction_8h__dep__incl.svg | 180 +-
.../reference/api/doxygen/instruction_8h__incl.svg | 1220 +++++-----
.../api/doxygen/instruction_8h_source.html | 2 +-
docs/reference/api/doxygen/instrument_8h.html | 2 +-
.../api/doxygen/instrument_8h__dep__incl.svg | 512 ++---
docs/reference/api/doxygen/instrument_8h__incl.svg | 1230 +++++-----
.../api/doxygen/instrument_8h_source.html | 2 +-
docs/reference/api/doxygen/int__set_8h.html | 2 +-
.../api/doxygen/int__set_8h__dep__incl.svg | 660 +++---
docs/reference/api/doxygen/int__set_8h__incl.svg | 1682 +++++++-------
docs/reference/api/doxygen/int__solver_8h.html | 2 +-
.../reference/api/doxygen/int__solver_8h__incl.svg | 1658 +++++++------
docs/reference/api/doxygen/interpreter_8h.html | 2 +-
.../reference/api/doxygen/interpreter_8h__incl.svg | 1720 +++++++-------
docs/reference/api/doxygen/ir_2adt_8h.html | 2 +-
.../api/doxygen/ir_2adt_8h__dep__incl.svg | 672 +++---
docs/reference/api/doxygen/ir_2adt_8h__incl.svg | 1472 ++++++------
docs/reference/api/doxygen/ir_2adt_8h_source.html | 2 +-
.../api/doxygen/ir_2attrs_8h__dep__incl.svg | 584 ++---
docs/reference/api/doxygen/ir_2attrs_8h__incl.svg | 1376 +++++------
.../reference/api/doxygen/ir_2attrs_8h_source.html | 2 +-
docs/reference/api/doxygen/ir_2expr_8h.html | 2 +-
.../api/doxygen/ir_2expr_8h__dep__incl.svg | 632 ++---
docs/reference/api/doxygen/ir_2expr_8h__incl.svg | 1474 ++++++------
docs/reference/api/doxygen/ir_2expr_8h_source.html | 2 +-
docs/reference/api/doxygen/ir_2function_8h.html | 2 +-
.../api/doxygen/ir_2function_8h__dep__incl.svg | 648 +++---
.../api/doxygen/ir_2function_8h__incl.svg | 1532 ++++++------
docs/reference/api/doxygen/ir_2module_8h.html | 2 +-
.../api/doxygen/ir_2module_8h__dep__incl.svg | 680 +++---
docs/reference/api/doxygen/ir_2module_8h__incl.svg | 2039 ++++++++--------
.../api/doxygen/ir_2module_8h_source.html | 2 +-
docs/reference/api/doxygen/ir_2op_8h.html | 2 +-
.../reference/api/doxygen/ir_2op_8h__dep__incl.svg | 560 ++---
docs/reference/api/doxygen/ir_2op_8h__incl.svg | 1893 ++++++++-------
docs/reference/api/doxygen/ir_2op_8h_source.html | 2 +-
docs/reference/api/doxygen/ir_2span_8h.html | 2 +-
.../api/doxygen/ir_2span_8h__dep__incl.svg | 700 +++---
docs/reference/api/doxygen/ir_2span_8h__incl.svg | 1352 +++++------
docs/reference/api/doxygen/ir_2span_8h_source.html | 2 +-
docs/reference/api/doxygen/ir_2transform_8h.html | 2 +-
.../api/doxygen/ir_2transform_8h__dep__incl.svg | 520 ++---
.../api/doxygen/ir_2transform_8h__incl.svg | 1908 ++++++++-------
.../api/doxygen/ir_2transform_8h_source.html | 2 +-
docs/reference/api/doxygen/ir_2type_8h.html | 2 +-
.../api/doxygen/ir_2type_8h__dep__incl.svg | 688 +++---
docs/reference/api/doxygen/ir_2type_8h__incl.svg | 1394 +++++------
docs/reference/api/doxygen/ir_2type_8h_source.html | 4 +-
.../api/doxygen/ir__docsifier_8h__incl.svg | 602 ++---
.../api/doxygen/ir__docsifier_8h_source.html | 2 +-
.../api/doxygen/iter__affine__map_8h.html | 2 +-
.../api/doxygen/iter__affine__map_8h__incl.svg | 1731 +++++++-------
.../api/doxygen/libtorch__runtime_8h.html | 2 +-
.../api/doxygen/libtorch__runtime_8h__incl.svg | 956 ++++----
.../api/doxygen/libtorch__runtime_8h_source.html | 2 +-
.../api/doxygen/local__response__norm_8h.html | 2 +-
.../api/doxygen/local__response__norm_8h__incl.svg | 1886 ++++++++-------
.../api/doxygen/loop__state_8h__dep__incl.svg | 88 +-
.../reference/api/doxygen/loop__state_8h__incl.svg | 634 ++---
.../api/doxygen/loop__state_8h_source.html | 2 +-
docs/reference/api/doxygen/mapping_8h.html | 2 +-
docs/reference/api/doxygen/mapping_8h__incl.svg | 1886 ++++++++-------
.../api/doxygen/measure_8h__dep__incl.svg | 48 +-
docs/reference/api/doxygen/measure_8h__incl.svg | 500 ++--
docs/reference/api/doxygen/measure_8h_source.html | 2 +-
.../api/doxygen/measure__callback_8h.html | 2 +-
.../doxygen/measure__callback_8h__dep__incl.svg | 12 +-
.../api/doxygen/measure__callback_8h__incl.svg | 2427 ++++++++++----------
.../doxygen/measure__candidate_8h__dep__incl.svg | 72 +-
.../api/doxygen/measure__candidate_8h__incl.svg | 594 ++---
.../api/doxygen/measure__record_8h__incl.svg | 502 ++--
.../api/doxygen/measure__record_8h_source.html | 2 +-
docs/reference/api/doxygen/memory__manager_8h.html | 2 +-
.../api/doxygen/memory__manager_8h__dep__incl.svg | 12 +-
.../api/doxygen/memory__manager_8h__incl.svg | 544 ++---
.../api/doxygen/memory__pools_8h__dep__incl.svg | 44 +-
.../api/doxygen/memory__pools_8h__incl.svg | 602 ++---
.../api/doxygen/memory__pools_8h_source.html | 2 +-
.../doxygen/meta__schedule_2cost__model_8h.html | 2 +-
.../meta__schedule_2cost__model_8h__dep__incl.svg | 48 +-
.../meta__schedule_2cost__model_8h__incl.svg | 1852 +++++++--------
.../meta__schedule_2cost__model_8h_source.html | 2 +-
docs/reference/api/doxygen/metadata_8h.html | 2 +-
docs/reference/api/doxygen/metadata_8h__incl.svg | 1458 ++++++------
docs/reference/api/doxygen/metadata__base_8h.html | 2 +-
.../api/doxygen/metadata__base_8h__dep__incl.svg | 12 +-
.../api/doxygen/metadata__base_8h__incl.svg | 1530 ++++++------
.../api/doxygen/metadata__base_8h_source.html | 4 +-
.../api/doxygen/mutator_8h__dep__incl.svg | 32 +-
docs/reference/api/doxygen/mutator_8h__incl.svg | 506 ++--
docs/reference/api/doxygen/name__supply_8h.html | 2 +-
.../api/doxygen/name__supply_8h__dep__incl.svg | 20 +-
.../api/doxygen/name__supply_8h__incl.svg | 1518 ++++++------
.../api/doxygen/name__supply_8h_source.html | 2 +-
.../api/doxygen/namespacemembers_func_p.html | 6 +-
docs/reference/api/doxygen/namespacemembers_p.html | 6 +-
docs/reference/api/doxygen/ndarray_8h.html | 2 +-
.../api/doxygen/ndarray_8h__dep__incl.svg | 696 +++---
docs/reference/api/doxygen/ndarray_8h__incl.svg | 480 ++--
docs/reference/api/doxygen/nn_2bnn_8h__incl.svg | 754 +++---
docs/reference/api/doxygen/nn_2dense_8h.html | 2 +-
.../api/doxygen/nn_2dense_8h__dep__incl.svg | 24 +-
docs/reference/api/doxygen/nn_2dense_8h__incl.svg | 1886 ++++++++-------
.../reference/api/doxygen/nn_2pooling_8h__incl.svg | 850 +++----
.../reference/api/doxygen/nn_2softmax_8h__incl.svg | 826 +++----
docs/reference/api/doxygen/node_8h.html | 2 +-
docs/reference/api/doxygen/node_8h__dep__incl.svg | 668 +++---
docs/reference/api/doxygen/node_8h__incl.svg | 1332 +++++------
docs/reference/api/doxygen/object__path_8h.html | 2 +-
.../api/doxygen/object__path_8h__dep__incl.svg | 652 +++---
.../api/doxygen/object__path_8h__incl.svg | 278 +--
.../api/doxygen/object__path_8h_source.html | 2 +-
docs/reference/api/doxygen/on__device_8h.html | 2 +-
docs/reference/api/doxygen/on__device_8h__incl.svg | 1964 ++++++++--------
.../api/doxygen/op__strategy_8h__incl.svg | 902 ++++----
.../api/doxygen/op__strategy_8h_source.html | 2 +-
docs/reference/api/doxygen/operation_8h.html | 2 +-
.../api/doxygen/operation_8h__dep__incl.svg | 592 ++---
docs/reference/api/doxygen/operation_8h__incl.svg | 1744 +++++++-------
.../reference/api/doxygen/operation_8h_source.html | 2 +-
docs/reference/api/doxygen/packed__func_8h.html | 2 +-
.../api/doxygen/packed__func_8h__dep__incl.svg | 588 ++---
.../api/doxygen/packed__func_8h__incl.svg | 764 +++---
.../api/doxygen/packed__func_8h_source.html | 2 +-
.../api/doxygen/pad__utils_8h__dep__incl.svg | 12 +-
docs/reference/api/doxygen/pad__utils_8h__incl.svg | 562 ++---
docs/reference/api/doxygen/papi_8h.html | 2 +-
docs/reference/api/doxygen/papi_8h__incl.svg | 1038 ++++-----
docs/reference/api/doxygen/parser_8h.html | 2 +-
docs/reference/api/doxygen/parser_8h__incl.svg | 1780 +++++++-------
docs/reference/api/doxygen/pattern_8h__incl.svg | 1624 +++++++------
.../api/doxygen/pattern__functor_8h__incl.svg | 726 +++---
.../api/doxygen/postproc_8h__dep__incl.svg | 32 +-
docs/reference/api/doxygen/postproc_8h__incl.svg | 550 ++---
docs/reference/api/doxygen/profiler_8h.html | 2 +-
docs/reference/api/doxygen/profiler_8h__incl.svg | 2213 +++++++++---------
docs/reference/api/doxygen/profiler_8h_source.html | 2 +-
docs/reference/api/doxygen/profiling_8h.html | 2 +-
.../api/doxygen/profiling_8h__dep__incl.svg | 12 +-
docs/reference/api/doxygen/profiling_8h__incl.svg | 1024 +++++----
.../reference/api/doxygen/profiling_8h_source.html | 2 +-
docs/reference/api/doxygen/random_8h.html | 2 +-
docs/reference/api/doxygen/random_8h__incl.svg | 1378 +++++------
docs/reference/api/doxygen/ravel__unravel_8h.html | 2 +-
.../api/doxygen/ravel__unravel_8h__dep__incl.svg | 84 +-
.../api/doxygen/ravel__unravel_8h__incl.svg | 1752 +++++++-------
docs/reference/api/doxygen/reduce_8h.html | 2 +-
docs/reference/api/doxygen/reduce_8h__incl.svg | 1386 +++++------
.../api/doxygen/reduction_8h__dep__incl.svg | 40 +-
docs/reference/api/doxygen/reduction_8h__incl.svg | 850 +++----
docs/reference/api/doxygen/reflection_8h.html | 2 +-
.../api/doxygen/reflection_8h__dep__incl.svg | 716 +++---
docs/reference/api/doxygen/reflection_8h__incl.svg | 1200 +++++-----
.../api/doxygen/reflection_8h_source.html | 2 +-
docs/reference/api/doxygen/registry_8h.html | 2 +-
.../api/doxygen/registry_8h__dep__incl.svg | 580 ++---
docs/reference/api/doxygen/registry_8h__incl.svg | 798 +++----
docs/reference/api/doxygen/relay_2adt_8h.html | 2 +-
.../api/doxygen/relay_2adt_8h__dep__incl.svg | 36 +-
docs/reference/api/doxygen/relay_2adt_8h__incl.svg | 2263 +++++++++---------
docs/reference/api/doxygen/relay_2analysis_8h.html | 2 +-
.../api/doxygen/relay_2analysis_8h__incl.svg | 1806 +++++++--------
.../api/doxygen/relay_2attrs_2debug_8h.html | 2 +-
.../api/doxygen/relay_2attrs_2debug_8h__incl.svg | 1410 ++++++------
.../api/doxygen/relay_2attrs_2memory_8h__incl.svg | 834 +++----
.../reference/api/doxygen/relay_2attrs_2nn_8h.html | 2 +-
.../api/doxygen/relay_2attrs_2nn_8h__incl.svg | 1668 +++++++-------
.../api/doxygen/relay_2attrs_2nn_8h_source.html | 2 +-
.../relay_2attrs_2transform_8h__dep__incl.svg | 20 +-
.../doxygen/relay_2attrs_2transform_8h__incl.svg | 786 +++----
.../doxygen/relay_2attrs_2transform_8h_source.html | 2 +-
.../reference/api/doxygen/relay_2attrs_2vm_8h.html | 2 +-
.../api/doxygen/relay_2attrs_2vm_8h__incl.svg | 1376 +++++------
docs/reference/api/doxygen/relay_2base_8h.html | 2 +-
.../api/doxygen/relay_2base_8h__dep__incl.svg | 296 +--
.../reference/api/doxygen/relay_2base_8h__incl.svg | 1716 +++++++-------
.../api/doxygen/relay_2base_8h_source.html | 2 +-
.../api/doxygen/relay_2expr_8h__dep__incl.svg | 212 +-
.../reference/api/doxygen/relay_2expr_8h__incl.svg | 822 +++----
.../api/doxygen/relay_2expr_8h_source.html | 2 +-
.../api/doxygen/relay_2expr__functor_8h__incl.svg | 726 +++---
docs/reference/api/doxygen/relay_2feature_8h.html | 2 +-
.../api/doxygen/relay_2feature_8h__incl.svg | 1834 ++++++++-------
docs/reference/api/doxygen/relay_2function_8h.html | 2 +-
.../api/doxygen/relay_2function_8h__dep__incl.svg | 36 +-
.../api/doxygen/relay_2function_8h__incl.svg | 1985 ++++++++--------
.../api/doxygen/relay_2op_8h__dep__incl.svg | 36 +-
docs/reference/api/doxygen/relay_2op_8h__incl.svg | 814 +++----
.../relay_2op__attr__types_8h__dep__incl.svg | 28 +-
.../doxygen/relay_2op__attr__types_8h__incl.svg | 886 +++----
.../api/doxygen/relay_2qnn_2attrs_8h.html | 2 +-
.../api/doxygen/relay_2qnn_2attrs_8h__incl.svg | 1386 +++++------
.../api/doxygen/relay_2qnn_2transform_8h__incl.svg | 720 +++---
.../api/doxygen/relay_2transform_8h__dep__incl.svg | 12 +-
.../api/doxygen/relay_2transform_8h__incl.svg | 736 +++---
.../api/doxygen/relay_2transform_8h_source.html | 2 +-
.../api/doxygen/relay_2type_8h__dep__incl.svg | 240 +-
.../reference/api/doxygen/relay_2type_8h__incl.svg | 734 +++---
docs/reference/api/doxygen/reorg_8h__incl.svg | 894 +++----
docs/reference/api/doxygen/rocblas_8h.html | 2 +-
.../api/doxygen/rocblas_8h__dep__incl.svg | 12 +-
docs/reference/api/doxygen/rocblas_8h__incl.svg | 2154 +++++++++--------
.../reference/api/doxygen/rocm_2dense_8h__incl.svg | 732 +++---
.../api/doxygen/rocm_2injective_8h__incl.svg | 758 +++---
.../api/doxygen/rocm_2pooling_8h__incl.svg | 734 +++---
.../api/doxygen/rocm_2reduction_8h__incl.svg | 758 +++---
.../api/doxygen/rocm_2softmax_8h__incl.svg | 758 +++---
docs/reference/api/doxygen/runner_8h.html | 2 +-
.../reference/api/doxygen/runner_8h__dep__incl.svg | 72 +-
docs/reference/api/doxygen/runner_8h__incl.svg | 2315 +++++++++----------
docs/reference/api/doxygen/runner_8h_source.html | 2 +-
.../runtime_2container_2base_8h_source.html | 2 +-
docs/reference/api/doxygen/runtime_2debug_8h.html | 2 +-
.../api/doxygen/runtime_2debug_8h__incl.svg | 540 ++---
docs/reference/api/doxygen/runtime_2module_8h.html | 2 +-
.../api/doxygen/runtime_2module_8h__dep__incl.svg | 588 ++---
.../api/doxygen/runtime_2module_8h__incl.svg | 942 ++++----
.../api/doxygen/runtime_2module_8h_source.html | 2 +-
docs/reference/api/doxygen/runtime_2vm_2vm_8h.html | 2 +-
.../api/doxygen/runtime_2vm_2vm_8h__incl.svg | 1382 +++++------
docs/reference/api/doxygen/runtime_8h.html | 2 +-
docs/reference/api/doxygen/runtime_8h__incl.svg | 1899 ++++++++-------
docs/reference/api/doxygen/runtime_8h_source.html | 2 +-
.../api/doxygen/schedule__pass_8h__dep__incl.svg | 136 +-
.../api/doxygen/schedule__pass_8h__incl.svg | 638 ++---
docs/reference/api/doxygen/schedule__rule_8h.html | 2 +-
.../api/doxygen/schedule__rule_8h__dep__incl.svg | 32 +-
.../api/doxygen/schedule__rule_8h__incl.svg | 1585 +++++++------
.../api/doxygen/schedule__rule_8h_source.html | 2 +-
docs/reference/api/doxygen/search/all_11.js | 2 +-
docs/reference/api/doxygen/search/all_13.js | 2 +-
docs/reference/api/doxygen/search/all_15.js | 2 -
docs/reference/api/doxygen/search/defines_8.js | 2 -
docs/reference/api/doxygen/search/functions_10.js | 2 +-
.../api/doxygen/search__policy_8h__dep__incl.svg | 12 +-
.../api/doxygen/search__policy_8h__incl.svg | 594 ++---
.../api/doxygen/search__policy_8h_source.html | 2 +-
.../api/doxygen/search__strategy_8h__dep__incl.svg | 36 +-
.../api/doxygen/search__strategy_8h__incl.svg | 742 +++---
docs/reference/api/doxygen/search__task_8h.html | 2 +-
.../api/doxygen/search__task_8h__dep__incl.svg | 60 +-
.../api/doxygen/search__task_8h__incl.svg | 1561 ++++++-------
.../api/doxygen/search__task_8h_source.html | 2 +-
docs/reference/api/doxygen/serializer_8h.html | 2 +-
.../api/doxygen/serializer_8h__dep__incl.svg | 696 +++---
docs/reference/api/doxygen/serializer_8h__incl.svg | 442 ++--
docs/reference/api/doxygen/source__map_8h.html | 2 +-
.../api/doxygen/source__map_8h__dep__incl.svg | 672 +++---
.../reference/api/doxygen/source__map_8h__incl.svg | 1462 ++++++------
.../api/doxygen/source__map_8h_source.html | 2 +-
.../api/doxygen/space__generator_8h__dep__incl.svg | 32 +-
.../api/doxygen/space__generator_8h__incl.svg | 604 ++---
docs/reference/api/doxygen/state_8h.html | 2 +-
docs/reference/api/doxygen/state_8h__dep__incl.svg | 144 +-
docs/reference/api/doxygen/state_8h__incl.svg | 1952 ++++++++--------
docs/reference/api/doxygen/stmt_8h.html | 2 +-
docs/reference/api/doxygen/stmt_8h__dep__incl.svg | 600 ++---
docs/reference/api/doxygen/stmt_8h__incl.svg | 1634 +++++++------
docs/reference/api/doxygen/stmt_8h_source.html | 2 +-
docs/reference/api/doxygen/stmt__functor_8h.html | 2 +-
.../api/doxygen/stmt__functor_8h__dep__incl.svg | 12 +-
.../api/doxygen/stmt__functor_8h__incl.svg | 1806 +++++++--------
docs/reference/api/doxygen/strided__slice_8h.html | 2 +-
.../api/doxygen/strided__slice_8h__dep__incl.svg | 72 +-
.../api/doxygen/strided__slice_8h__incl.svg | 1669 +++++++-------
docs/reference/api/doxygen/string_8h.html | 44 +-
.../reference/api/doxygen/string_8h__dep__incl.svg | 672 +++---
docs/reference/api/doxygen/string_8h__incl.svg | 90 +-
docs/reference/api/doxygen/string_8h_source.html | 66 +-
.../api/doxygen/structural__equal_8h.html | 2 +-
.../doxygen/structural__equal_8h__dep__incl.svg | 636 ++---
.../api/doxygen/structural__equal_8h__incl.svg | 460 ++--
.../reference/api/doxygen/structural__hash_8h.html | 2 +-
.../api/doxygen/structural__hash_8h__dep__incl.svg | 636 ++---
.../api/doxygen/structural__hash_8h__incl.svg | 488 ++--
docs/reference/api/doxygen/tag_8h.html | 2 +-
docs/reference/api/doxygen/tag_8h__incl.svg | 1908 ++++++++-------
docs/reference/api/doxygen/tag_8h_source.html | 2 +-
docs/reference/api/doxygen/target_8h.html | 2 +-
.../reference/api/doxygen/target_8h__dep__incl.svg | 508 ++--
docs/reference/api/doxygen/target_8h__incl.svg | 2343 +++++++++----------
docs/reference/api/doxygen/target_8h_source.html | 2 +-
docs/reference/api/doxygen/target__info_8h.html | 2 +-
.../api/doxygen/target__info_8h__incl.svg | 1506 ++++++------
docs/reference/api/doxygen/target__kind_8h.html | 2 +-
.../api/doxygen/target__kind_8h__dep__incl.svg | 508 ++--
.../api/doxygen/target__kind_8h__incl.svg | 1576 ++++++-------
.../api/doxygen/target__kind_8h_source.html | 2 +-
.../api/doxygen/task__scheduler_8h__incl.svg | 884 +++----
.../api/doxygen/te_2schedule_8h__dep__incl.svg | 648 +++---
.../api/doxygen/te_2schedule_8h__incl.svg | 464 ++--
docs/reference/api/doxygen/tensor_8h.html | 2 +-
.../reference/api/doxygen/tensor_8h__dep__incl.svg | 628 ++---
docs/reference/api/doxygen/tensor_8h__incl.svg | 1650 ++++++-------
docs/reference/api/doxygen/tensor_8h_source.html | 2 +-
docs/reference/api/doxygen/tensor__intrin_8h.html | 2 +-
.../api/doxygen/tensor__intrin_8h__dep__incl.svg | 640 +++---
.../api/doxygen/tensor__intrin_8h__incl.svg | 1657 +++++++------
.../api/doxygen/tensor__type_8h__dep__incl.svg | 248 +-
.../api/doxygen/tensor__type_8h__incl.svg | 1504 ++++++------
docs/reference/api/doxygen/tensor__utils_8h.html | 2 +-
.../api/doxygen/tensor__utils_8h__dep__incl.svg | 80 +-
.../api/doxygen/tensor__utils_8h__incl.svg | 1752 +++++++-------
.../api/doxygen/tir_2analysis_8h__dep__incl.svg | 176 +-
.../api/doxygen/tir_2analysis_8h__incl.svg | 702 +++---
docs/reference/api/doxygen/tir_2expr_8h.html | 2 +-
.../api/doxygen/tir_2expr_8h__dep__incl.svg | 608 ++---
docs/reference/api/doxygen/tir_2expr_8h__incl.svg | 1486 ++++++------
.../reference/api/doxygen/tir_2expr_8h_source.html | 2 +-
.../api/doxygen/tir_2expr__functor_8h.html | 2 +-
.../doxygen/tir_2expr__functor_8h__dep__incl.svg | 20 +-
.../api/doxygen/tir_2expr__functor_8h__incl.svg | 1618 +++++++------
docs/reference/api/doxygen/tir_2function_8h.html | 2 +-
.../api/doxygen/tir_2function_8h__dep__incl.svg | 536 ++---
.../api/doxygen/tir_2function_8h__incl.svg | 1911 ++++++++-------
.../api/doxygen/tir_2function_8h_source.html | 2 +-
docs/reference/api/doxygen/tir_2op_8h.html | 2 +-
.../api/doxygen/tir_2op_8h__dep__incl.svg | 652 +++---
docs/reference/api/doxygen/tir_2op_8h__incl.svg | 1904 +++++++--------
.../api/doxygen/tir_2op__attr__types_8h.html | 2 +-
.../doxygen/tir_2op__attr__types_8h__dep__incl.svg | 184 +-
.../api/doxygen/tir_2op__attr__types_8h__incl.svg | 1516 ++++++------
.../doxygen/tir_2op__attr__types_8h_source.html | 2 +-
.../tir_2schedule_2schedule_8h__dep__incl.svg | 136 +-
.../doxygen/tir_2schedule_2schedule_8h__incl.svg | 568 ++---
.../doxygen/tir_2schedule_2schedule_8h_source.html | 2 +-
docs/reference/api/doxygen/tir_2transform_8h.html | 2 +-
.../api/doxygen/tir_2transform_8h__incl.svg | 1720 +++++++-------
.../api/doxygen/tir_2usmp_2analysis_8h__incl.svg | 690 +++---
.../api/doxygen/tir_2usmp_2transform_8h.html | 2 +-
.../api/doxygen/tir_2usmp_2transform_8h__incl.svg | 1680 +++++++-------
.../reference/api/doxygen/tir_2usmp_2utils_8h.html | 2 +-
.../api/doxygen/tir_2usmp_2utils_8h__dep__incl.svg | 36 +-
.../api/doxygen/tir_2usmp_2utils_8h__incl.svg | 1762 +++++++-------
.../api/doxygen/tir_2usmp_2utils_8h_source.html | 2 +-
.../api/doxygen/topi_2nn_8h__dep__incl.svg | 12 +-
docs/reference/api/doxygen/topi_2nn_8h__incl.svg | 822 +++----
.../api/doxygen/topi_2transform_8h__dep__incl.svg | 64 +-
.../api/doxygen/topi_2transform_8h__incl.svg | 742 +++---
.../api/doxygen/topi_2transform_8h_source.html | 2 +-
docs/reference/api/doxygen/topi_2utils_8h.html | 2 +-
.../reference/api/doxygen/topi_2utils_8h__incl.svg | 1508 ++++++------
docs/reference/api/doxygen/trace_8h.html | 2 +-
docs/reference/api/doxygen/trace_8h__dep__incl.svg | 172 +-
docs/reference/api/doxygen/trace_8h__incl.svg | 1228 +++++-----
docs/reference/api/doxygen/traced__object_8h.html | 2 +-
.../api/doxygen/traced__object_8h__dep__incl.svg | 36 +-
.../api/doxygen/traced__object_8h__incl.svg | 1342 +++++------
.../api/doxygen/traced__object__functor_8h.html | 2 +-
.../traced__object__functor_8h__dep__incl.svg | 12 +-
.../doxygen/traced__object__functor_8h__incl.svg | 1470 ++++++------
.../doxygen/traced__object__functor_8h_source.html | 2 +-
.../api/doxygen/transform__step_8h__dep__incl.svg | 96 +-
.../api/doxygen/transform__step_8h__incl.svg | 638 ++---
.../api/doxygen/transform__step_8h_source.html | 2 +-
.../api/doxygen/tune__context_8h__dep__incl.svg | 24 +-
.../api/doxygen/tune__context_8h__incl.svg | 886 +++----
.../api/doxygen/type__functor_8h__incl.svg | 758 +++---
.../api/doxygen/type__relation_8h__dep__incl.svg | 552 ++---
.../api/doxygen/type__relation_8h__incl.svg | 604 ++---
docs/reference/api/doxygen/var_8h.html | 2 +-
docs/reference/api/doxygen/var_8h__dep__incl.svg | 644 +++---
docs/reference/api/doxygen/var_8h__incl.svg | 1520 ++++++------
docs/reference/api/doxygen/var_8h_source.html | 2 +-
docs/reference/api/doxygen/var__table_8h.html | 2 +-
.../api/doxygen/var__table_8h__dep__incl.svg | 12 +-
docs/reference/api/doxygen/var__table_8h__incl.svg | 1582 ++++++-------
.../api/doxygen/var__table_8h_source.html | 2 +-
docs/reference/api/doxygen/virtual__device_8h.html | 2 +-
.../api/doxygen/virtual__device_8h__dep__incl.svg | 256 +--
.../api/doxygen/virtual__device_8h__incl.svg | 1947 ++++++++--------
.../api/doxygen/virtual__device_8h_source.html | 4 +-
docs/reference/api/doxygen/vision_8h.html | 2 +-
docs/reference/api/doxygen/vision_8h__incl.svg | 1670 +++++++-------
docs/reference/api/doxygen/x86_2bnn_8h__incl.svg | 754 +++---
.../api/doxygen/x86_2default_8h__incl.svg | 758 +++---
.../api/doxygen/x86_2injective_8h__incl.svg | 754 +++---
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 | 4 +-
.../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 | 5 +-
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 | 26 +-
docs/tutorial/tensor_expr_get_started.html | 45 +-
656 files changed, 143186 insertions(+), 145689 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 0c2b74ce5..fbfb9b20b 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 9.820 seconds)
+ **Total running time of the script:** ( 1 minutes 7.078 seconds)
.. _sphx_glr_download_how_to_compile_models_from_darknet.py:
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 9b4220d62..7ba07ac83 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.zip0659f513-2db6-4c25-a8ef-a104b78991b5 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip...
+ Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zip4b8c443c-c2ca-4cbc-9cb2-6cc91ea8e8a8 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 b4156c79a..0ebd36f14 100644
--- a/docs/_sources/how_to/compile_models/from_oneflow.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_oneflow.rst.txt
@@ -113,7 +113,7 @@ 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]
15%|#5 | 6.33M/41.5M [00:00<00:00, 39.3MB/s]
24%|##4 | 10.1M/41.5M [00:00<00:01, 32.6MB/s]
35%|###4 | 14.3M/41.5M [00:00<00:01, 27.9MB/s]
41%|####1 | 17.0M/41.5M [00:00<00:01, 23.7MB/s]
54%|#####3 | 22.3M/41.5M [00:01<00:01, 14.1MB/s]
58%|#####8 | 24.1M/41.5M [00:01<00:01, 14.0MB/s]
77%|#######7 | 32.0M/41.5M [00:01<00:00, 24.4MB/s]
85%|########5 | 35.4M/41.5M [00:01<00:00, 25.1MB/s]
96%|#########6| 40.0M/41.5M [00:01<00:00, 29.4MB/s]
100%|##########| 41.5M/41.5M [00:01<00:00, 24.8MB/s]
+
0%| | 0.00/41.5M [00:00<?, ?B/s]
15%|#5 | 6.33M/41.5M [00:00<00:01, 25.1MB/s]
21%|##1 | 8.73M/41.5M [00:00<00:01, 18.6MB/s]
35%|###4 | 14.3M/41.5M [00:00<00:01, 27.3MB/s]
42%|####1 | 17.2M/41.5M [00:00<00:01, 24.2MB/s]
58%|#####7 | 24.0M/41.5M [00:00<00:00, 31.3MB/s]
77%|#######7 | 32.0M/41.5M [00:01<00:00, 39.4MB/s]
87%|########6 | 35.9M/41.5M [00:01<00:00, 29.9MB/s]
94%|#########4| 39.1M/41.5M [00:01<00:00, 24.7MB/s]
100%|##########| 41.5M/41.5M [00:01<00:00, 27.3MB/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 21dd2d67f..6d996e167 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]
2%|1 | 896k/44.7M [00:00<00:05, 9.17MB/s]
17%|#6 | 7.49M/44.7M [00:00<00:00, 44.5MB/s]
33%|###2 | 14.6M/44.7M [00:00<00:00, 58.3MB/s]
49%|####8 | 21.8M/44.7M [00:00<00:00, 64.7MB/s]
65%|######5 | 29.1M/44.7M [00:00<00:00, 68.8MB/s]
81%|######## | 36.1M/44.7M [00:00<00:00, 70.6MB/s]
97%|#########7| 43.4M/44.7M [00:00<00:00, 72.2MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 65.1MB/s]
+
0%| | 0.00/44.7M [00:00<?, ?B/s]
2%|2 | 968k/44.7M [00:00<00:04, 9.87MB/s]
17%|#7 | 7.73M/44.7M [00:00<00:00, 45.5MB/s]
30%|##9 | 13.2M/44.7M [00:00<00:00, 50.7MB/s]
43%|####3 | 19.3M/44.7M [00:00<00:00, 55.6MB/s]
55%|#####5 | 24.6M/44.7M [00:00<00:00, 54.3MB/s]
69%|######9 | 30.8M/44.7M [00:00<00:00, 58.0MB/s]
82%|########2 | 36.8M/44.7M [00:00<00:00, 56.7MB/s]
97%|#########6| 43.2M/44.7M [00:00<00:00, 59.3MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 54.7MB/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 49b0a5c93..f72828055 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 5.209 seconds)
+ **Total running time of the script:** ( 1 minutes 6.195 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 bf254159a..399cca16f 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
=================
-**05:18.343** total execution time for **how_to_compile_models** files:
+**05:17.453** total execution time for **how_to_compile_models** files:
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``) | 01:09.820 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``) | 01:07.078 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:05.209 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:06.195 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``) | 00:39.852 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``) | 00:40.806 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``) | 00:29.669 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``) | 00:29.287 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``) | 00:26.729 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``) | 00:26.858 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``) | 00:24.728 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``) | 00:26.161 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``) | 00:23.190 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``) | 00:22.911 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``) | 00:21.271 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``) | 00:20.628 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``) | 00:15.385 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``) | 00:14.803 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``) | 00:02.490 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``) | 00:02.726 | 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 56ae2f971..9d3a33d09 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)
- 15.9095 15.7277 16.6322 15.5471 0.3829
+ 15.6957 15.6786 15.8327 15.6209 0.0719
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 65b667a46..8e66bd994 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%|1 | 2.84M/170M [00:00<00:05, 29.8MB/s]
4%|4 | 7.62M/170M [00:00<00:04, 40.8MB/s]
8%|7 | 12.8M/170M [00:00<00:03, 46.6MB/s]
10%|# | 17.2M/170M [00:00<00:03, 43.8MB/s]
13%|#2 | 21.4M/170M [00:00<00:03, 39.5MB/s]
15%|#4 | 25.4M/170M [00:00<00:03, 40.0MB/s]
17%|#7 | 29.4M/170M [00:00<00:03, 40.4MB/s]
20%|#9 | 33.5M/170M [00:00<00:03, 41.3MB/s]
22%|##2 | 37.4M/170M [00:00<00:03, 40.6MB/s]
24%|##4 | 41.3M/170M [00:01<00:03, 39.9MB/s]
27%|##6 | 45.2M/170M [00:01<00:03, 37.2MB/s]
29%|##8 | 48.9M/170M [00:01<00:03, 37.5MB/s]
31%|###1 | 52.9M/170M [00:01<00:03, 38.6MB/s]
33%|###3 | 56.6M/170M [00:01<00:05, 21.7MB/s]
35%|###5 | 59.5M/170M [00:01<00:05, 21.3MB/s]
37%|###7 | 63.3M/170M [00:02<00:04, 24.8MB/s]
40%|#### | 68.0M/170M [00:02<00:03, 30.0MB/
s]
44%|####3 | 74.0M/170M [00:02<00:02, 37.9MB/s]
46%|####6 | 78.2M/170M [00:02<00:02, 38.5MB/s]
48%|####8 | 82.3M/170M [00:02<00:02, 38.5MB/s]
51%|#####1 | 86.8M/170M [00:02<00:02, 40.6MB/s]
54%|#####3 | 90.9M/170M [00:02<00:02, 38.4MB/s]
56%|#####5 | 94.7M/170M [00:02<00:02, 38.6MB/s]
58%|#####8 | 98.5M/170M [00:02<00:02, 36.6MB/s]
60%|###### | 102M/170M [00:03<00:01, 36.5MB/s]
62%|######2 | 106M/170M [00:03<00:01, 37.1MB/s]
65%|######5 | 111M/170M [00:03<00:01, 40.9MB/s]
68%|######8 | 116M/170M [00:03<00:01, 45.9MB/s]
71%|#######1 | 121M/170M [00:03<00:01, 44.1MB/s]
74%|#######3 | 125M/170M [00:03<00:01, 40.8MB/s]
76%|#######5 | 129M/170M [00:03<00:01, 39.5MB/s]
78%|#######8 | 133M/170M [00:03<00:01, 35.9MB/s]
80%|######## | 136M/170M [00:03<00:01, 35.1MB/s]
82%|########2 | 140M/170M [00:04<00:01, 27.9MB/s]
84%|########4 | 143M/170M [00:04<00:00, 30.1MB
/s]
86%|########6 | 146M/170M [00:04<00:00, 29.9MB/s]
88%|########7 | 149M/170M [00:04<00:00, 26.0MB/s]
90%|########9 | 152M/170M [00:04<00:00, 27.0MB/s]
91%|#########1| 155M/170M [00:04<00:00, 26.6MB/s]
93%|#########2| 158M/170M [00:04<00:00, 25.0MB/s]
95%|#########4| 161M/170M [00:04<00:00, 26.9MB/s]
97%|#########6| 165M/170M [00:05<00:00, 30.7MB/s]
99%|#########8| 168M/170M [00:05<00:00, 31.5MB/s]
100%|##########| 170M/170M [00:05<00:00, 34.3MB/s]
+
0%| | 0.00/170M [00:00<?, ?B/s]
1%| | 928k/170M [00:00<00:18, 9.48MB/s]
3%|2 | 5.09M/170M [00:00<00:05, 29.6MB/s]
7%|6 | 11.2M/170M [00:00<00:03, 45.5MB/s]
10%|9 | 16.2M/170M [00:00<00:03, 48.1MB/s]
13%|#2 | 21.7M/170M [00:00<00:03, 51.4MB/s]
16%|#5 | 26.6M/170M [00:00<00:03, 47.4MB/s]
19%|#8 | 31.7M/170M [00:00<00:02, 49.2MB/s]
21%|##1 | 36.5M/170M [00:01<00:04, 30.5MB/s]
24%|##3 | 40.2M/170M [00:01<00:04, 31.0MB/s]
27%|##6 | 45.0M/170M [00:01<00:03, 35.4MB/s]
30%|##9 | 50.8M/170M [00:01<00:03, 41.3MB/s]
33%|###2 | 55.2M/170M [00:01<00:03, 39.5MB/s]
35%|###4 | 59.4M/170M [00:01<00:02, 38.8MB/s]
37%|###7 | 63.3M/170M [00:01<00:02, 39.5MB/s]
41%|#### | 69.1M/170M [00:01<00:02, 45.1MB/s]
43%|####3 | 73.6M/170M [00:01<00:02, 44.6MB/s]
47%|####7 | 79.8M/170M [00:02<00:01, 50.2MB/s
]
50%|####9 | 84.8M/170M [00:02<00:01, 47.6MB/s]
53%|#####2 | 89.4M/170M [00:02<00:01, 47.3MB/s]
56%|#####5 | 94.5M/170M [00:02<00:01, 49.0MB/s]
58%|#####8 | 99.3M/170M [00:02<00:01, 46.4MB/s]
62%|######1 | 105M/170M [00:02<00:01, 50.2MB/s]
65%|######5 | 111M/170M [00:02<00:01, 53.7MB/s]
68%|######8 | 116M/170M [00:02<00:01, 52.3MB/s]
71%|#######1 | 121M/170M [00:02<00:00, 51.0MB/s]
75%|#######5 | 127M/170M [00:02<00:00, 55.0MB/s]
78%|#######8 | 133M/170M [00:03<00:00, 56.6MB/s]
82%|########1 | 139M/170M [00:03<00:00, 53.3MB/s]
85%|########4 | 144M/170M [00:03<00:00, 53.4MB/s]
88%|########7 | 149M/170M [00:03<00:00, 52.8MB/s]
91%|######### | 154M/170M [00:03<00:00, 51.3MB/s]
94%|#########3| 159M/170M [00:03<00:00, 50.3MB/s]
98%|#########7| 166M/170M [00:03<00:00, 56.6MB/s]
100%|##########| 170M/170M [00:03<00:00, 47.0MB/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').
@@ -292,7 +292,7 @@ Get boxes with score larger than 0.9
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 2 minutes 58.211 seconds)
+ **Total running time of the script:** ( 3 minutes 2.080 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 1f308677e..4f1f3b8c1 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]
20%|## | 2.72M/13.6M [00:00<00:00, 28.5MB/s]
45%|####5 | 6.15M/13.6M [00:00<00:00, 32.9MB/s]
69%|######8 | 9.29M/13.6M [00:00<00:00, 32.4MB/s]
93%|#########3| 12.6M/13.6M [00:00<00:00, 33.5MB/s]
100%|##########| 13.6M/13.6M [00:00<00:00, 32.9MB/s]
+
0%| | 0.00/13.6M [00:00<?, ?B/s]
7%|6 | 904k/13.6M [00:00<00:01, 9.25MB/s]
55%|#####5 | 7.47M/13.6M [00:00<00:00, 44.3MB/s]
100%|##########| 13.6M/13.6M [00:00<00:00, 49.5MB/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)
- 90.2987 90.2379 91.7357 90.0738 0.2322
+ 90.2603 90.1888 91.1384 90.0533 0.2049
@@ -461,7 +461,7 @@ TODO
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 8.947 seconds)
+ **Total running time of the script:** ( 1 minutes 10.686 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 af8804bb5..44562f5ce 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)
- 120.1967 120.1969 120.9517 119.5407 0.3156
+ 119.1746 119.1299 121.7804 118.4658 0.4477
@@ -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 51.089 seconds)
+ **Total running time of the script:** ( 1 minutes 51.991 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 bdad5daa5..465bdf1b0 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 50.354 seconds)
+ **Total running time of the script:** ( 1 minutes 41.133 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 fb4165828..7217ad709 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]
1%| | 1030/132723 [00:00<00:12, 10251.13KB/s]
6%|6 | 8176/132723 [00:00<00:02, 46182.28KB/s]
12%|#1 | 15534/132723 [00:00<00:01, 58676.08KB/s]
18%|#7 | 23262/132723 [00:00<00:01, 66011.25KB/s]
23%|##3 | 30920/132723 [00:00<00:01, 69817.40KB/s]
29%|##9 | 38542/132723 [00:00<00:01, 71991.91KB/s]
35%|###4 | 46180/132723 [00:00<00:01, 73424.71KB/s]
40%|#### | 53524/132723 [00:00<00:01, 49586.36KB/s]
46%|####6 | 61072/132723 [00:01<00:01, 55666.71KB/s]
52%|#####1 | 68723/132723 [00:01<00:01, 60903.36KB/s]
58%|#####7 | 76365/132723 [00:01<00:00, 65001.12KB/s]
63%|######2 | 83433/132723 [00:01<00:01, 29955.18KB/s]
68%|######8 | 90416/132723 [00:01<00:01, 35926.38KB/s]
73%|#######3 | 97075/132723 [00:02<00:00, 40777.62KB/s]
78%|#######7 | 103003/132723 [00:02<00:00, 38184.16KB/s]
83%|########3 |
110620/132723 [00:02<00:00, 45637.22KB/s]
88%|########7 | 116492/132723 [00:02<00:00, 36926.78KB/s]
94%|#########3| 124164/132723 [00:02<00:00, 44579.36KB/s]
99%|#########9| 131399/132723 [00:02<00:00, 50563.83KB/s]
100%|##########| 132723/132723 [00:02<00:00, 48284.75KB/s]
+
0%| | 0/132723 [00:00<?, ?KB/s]
2%|2 | 2701/132723 [00:00<00:04, 27008.42KB/s]
6%|5 | 7962/132723 [00:00<00:02, 41748.56KB/s]
10%|# | 13715/132723 [00:00<00:02, 48922.68KB/s]
16%|#6 | 21829/132723 [00:00<00:01, 61607.09KB/s]
23%|##2 | 29960/132723 [00:00<00:01, 68698.04KB/s]
29%|##8 | 38074/132723 [00:00<00:01, 72918.55KB/s]
35%|###4 | 46279/132723 [00:00<00:01, 75898.59KB/s]
41%|####1 | 54417/132723 [00:00<00:01, 77641.29KB/s]
47%|####7 | 62571/132723 [00:00<00:00, 78857.03KB/s]
53%|#####3 | 70686/132723 [00:01<00:00, 79563.09KB/s]
59%|#####9 | 78758/132723 [00:01<00:00, 79915.83KB/s]
65%|######5 | 86826/132723 [00:01<00:00, 80146.25KB/s]
71%|#######1 | 94864/132723 [00:01<00:00, 80216.15KB/s]
78%|#######7 | 102886/132723 [00:01<00:00, 67625.83KB/s]
84%|########3 | 111087/132723 [00:01<00:00, 71459.86KB/s]
89%|########9
| 118518/132723 [00:01<00:00, 61858.48KB/s]
95%|#########5| 126716/132723 [00:01<00:00, 66930.73KB/s]
100%|##########| 132723/132723 [00:01<00:00, 69919.40KB/s]
@@ -241,7 +241,7 @@ Display result
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 2 minutes 35.025 seconds)
+ **Total running time of the script:** ( 2 minutes 37.687 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 7ba7e8c66..b85e41ff0 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
=================
-**11:37.040** total execution time for **how_to_deploy_models** files:
+**11:39.319** 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``) | 02:58.211 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 03:02.080 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``) | 02:35.025 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``) | 02:37.687 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``) | 01:51.089 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``) | 01:51.991 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``) | 01:50.354 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``) | 01:41.133 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``) | 01:08.947 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``) | 01:10.686 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``) | 00:29.455 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``) | 00:30.953 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``) | 00:22.190 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``) | 00:22.579 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``) | 00:21.764 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``) | 00:22.204 | 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 74194f465..be9734d26 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.zipaad45c1c-7ae2-421e-9dd3-5b38617583f4 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
+ Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zip965811c5-01ad-4e01-a130-3ec1a6e56e45 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
@@ -590,7 +590,7 @@ Now, to actually convert the entire network, we have written `a pass in Relay <h
/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. "
- Check failed: (lower) is false: FloatImm lowering function for target llvm type 150 not found
+ Check failed: (lower) is false: Intrinsic lowering function for target llvm, intrinsic name tir.sqrt, type 150 not found
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 5c05e8e91..73cc9370b 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
=================
-**00:41.104** total execution time for **how_to_extend_tvm** files:
+**00:42.186** 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``) | 00:37.947 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_bring_your_own_datatypes.py` (``bring_your_own_datatypes.py``) | 00:38.864 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``) | 00:02.235 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``) | 00:02.382 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_infra.py` (``use_pass_infra.py``) | 00:00.915 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_infra.py` (``use_pass_infra.py``) | 00:00.933 | 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 |
+| :ref:`sphx_glr_how_to_extend_tvm_low_level_custom_pass.py` (``low_level_custom_pass.py``) | 00:00.008 | 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 a1e0732b5..7eacc197f 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: 6794us [6794us] (46.34%; 46.34%)
- FoldScaleAxis: 7869us [5us] (53.66%; 53.66%)
- FoldConstant: 7863us [1634us] (53.63%; 99.93%)
- InferType: 6230us [6230us] (42.49%; 79.22%)
+ InferType: 6850us [6850us] (46.71%; 46.71%)
+ FoldScaleAxis: 7816us [6us] (53.29%; 53.29%)
+ FoldConstant: 7810us [1580us] (53.26%; 99.93%)
+ InferType: 6230us [6230us] (42.48%; 79.77%)
@@ -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: 6309us [6309us] (44.78%; 44.78%)
- FoldScaleAxis: 7781us [5us] (55.22%; 55.22%)
- FoldConstant: 7777us [1595us] (55.19%; 99.94%)
- InferType: 6182us [6182us] (43.87%; 79.49%)
+ InferType: 6182us [6182us] (44.62%; 44.62%)
+ FoldScaleAxis: 7674us [4us] (55.38%; 55.38%)
+ FoldConstant: 7670us [1592us] (55.35%; 99.94%)
+ InferType: 6078us [6078us] (43.86%; 79.24%)
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 f14a5e799..c381e6a4a 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: 54.157457 ms
+ Convolution: 54.159903 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 dbfe321e6..10fc6b58b 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: 7.459653 ms
+ conv2d with tensor core: 7.746062 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 250a90bce..77e68ff5d 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: 0.018450
- Baseline: 3.395711
+ Numpy running time: 0.018982
+ Baseline: 3.249589
@@ -239,7 +239,7 @@ fill 32 * 32 * sizeof(float) which is 4KB in the cache whose total size is 32KB
.. code-block:: none
- Opt1: 0.290153
+ Opt1: 0.318530
@@ -342,7 +342,7 @@ In this tutorial, we chose to vectorize the inner loop row data since it is cach
.. code-block:: none
- Opt2: 0.325772
+ Opt2: 0.339569
@@ -438,7 +438,7 @@ the access pattern for A matrix is more cache friendly.
.. code-block:: none
- Opt3: 0.121021
+ Opt3: 0.116613
@@ -563,7 +563,7 @@ flattening.
.. code-block:: none
- Opt4: 0.110929
+ Opt4: 0.110777
@@ -685,7 +685,7 @@ write to C when all the block results are ready.
.. code-block:: none
- Opt5: 0.111273
+ Opt5: 0.112163
@@ -810,7 +810,7 @@ Futhermore, we can also utilize multi-core processors to do the thread-level par
.. code-block:: none
- Opt6: 0.144865
+ Opt6: 0.145942
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 07a6a9d40..9554d2ff0 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
=================
-**00:34.338** total execution time for **how_to_optimize_operators** files:
+**00:34.349** total execution time for **how_to_optimize_operators** files:
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``) | 00:31.944 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``) | 00:32.127 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.303 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.214 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``) | 00:01.092 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``) | 00:01.009 | 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 edb6c2d87..9b82c0ab4 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
=================
-**06:05.853** total execution time for **how_to_tune_with_autoscheduler** files:
+**06:13.103** 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``) | 03:20.355 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_conv2d_layer_cuda.py` (``tune_conv2d_layer_cuda.py``) | 03:22.148 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``) | 01:22.398 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``) | 01:23.959 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``) | 00:46.795 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``) | 00:47.825 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``) | 00:18.769 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``) | 00:21.450 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``) | 00:08.886 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``) | 00:08.892 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``) | 00:08.649 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``) | 00:08.829 | 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 f4c0f78a5..236aed387 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
@@ -240,562 +240,224 @@ cooperative fetching, unrolling and operator fusion.
compute: Buffer(compute_2: Pointer(float32), float32, [25088], [])}
buffer_map = {data_1: data, kernel_1: kernel, bias_1: bias, compute_1: compute}
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, [7]), storage_scope = local;
- allocate(pad_temp.shared: Pointer(shared float32), float32, [648]), storage_scope = shared;
- allocate(kernel.shared: Pointer(shared float32), float32, [2304]), storage_scope = shared;
- attr [IterVar(threadIdx.x: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224 {
- conv2d_nchw_1: Buffer(conv2d_nchw, float32, [7], [], scope="local", align=16)[0] = 0f32
+ attr [IterVar(blockIdx.x: int32, (nullptr), "ThreadIndex", "blockIdx.x")] "thread_extent" = 8;
+ allocate(conv2d_nchw: Pointer(local float32), float32, [28]), storage_scope = local;
+ allocate(pad_temp.shared: Pointer(shared float32), float32, [2016]), storage_scope = shared;
+ allocate(kernel.shared: Pointer(shared float32), float32, [6144]), storage_scope = shared;
+ attr [IterVar(threadIdx.x: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ conv2d_nchw_1: Buffer(conv2d_nchw, float32, [28], [], scope="local", align=64)[0] = 0f32
conv2d_nchw_1[1] = 0f32
conv2d_nchw_1[2] = 0f32
conv2d_nchw_1[3] = 0f32
conv2d_nchw_1[4] = 0f32
conv2d_nchw_1[5] = 0f32
conv2d_nchw_1[6] = 0f32
- for (rc.outer.outer: int32, 0, 64) {
- let cse_var_2: int32 = (rc.outer.outer*392)
- let cse_var_1: int32 = (rc.outer.outer*72)
- {
- attr [IterVar(threadIdx.x_1: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- pad_temp.shared_1: Buffer(pad_temp.shared, float32, [648], [], scope="shared")[threadIdx.x_1] = @tir.if_then_else(((((9 <= floormod(threadIdx.x_1, 81)) && (floormod(threadIdx.x_1, 81) < 72)) && (1 <= floormod(threadIdx.x_1, 9))) && (floormod(threadIdx.x_1, 9) < 8)), data[((((cse_var_2 + (floordiv(threadIdx.x_1, 81)*49)) + (floordiv(floormod(threadIdx.x_1, 81), 9)*7)) + floormod(threadIdx.x_1, 9)) - 8)], 0f32, dtype=float32)
- attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- pad_temp.shared_1[(threadIdx.x_1 + 224)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 62), 81)) && (floormod((threadIdx.x_1 + 62), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 8), 9))) && (floormod((threadIdx.x_1 + 8), 9) < 8)), data[((((cse_var_2 + (floordiv((threadIdx.x_1 + 224), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 62), 81), 9)*7)) + floormod((threadIdx.x_1 + 8), 9)) - 8)], 0f32, dtype=float32)
- attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- if @tir.likely((threadIdx.x_1 < 200), dtype=bool) {
- pad_temp.shared_1[(threadIdx.x_1 + 448)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 43), 81)) && (floormod((threadIdx.x_1 + 43), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 7), 9))) && (floormod((threadIdx.x_1 + 7), 9) < 8)), data[((((cse_var_2 + (floordiv((threadIdx.x_1 + 448), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 43), 81), 9)*7)) + floormod((threadIdx.x_1 + 7), 9)) - 8)], 0f32, dtype=float32)
+ conv2d_nchw_1[7] = 0f32
+ conv2d_nchw_1[8] = 0f32
+ conv2d_nchw_1[9] = 0f32
+ conv2d_nchw_1[10] = 0f32
+ conv2d_nchw_1[11] = 0f32
+ conv2d_nchw_1[12] = 0f32
+ conv2d_nchw_1[13] = 0f32
+ conv2d_nchw_1[14] = 0f32
+ conv2d_nchw_1[15] = 0f32
+ conv2d_nchw_1[16] = 0f32
+ conv2d_nchw_1[17] = 0f32
+ conv2d_nchw_1[18] = 0f32
+ conv2d_nchw_1[19] = 0f32
+ conv2d_nchw_1[20] = 0f32
+ conv2d_nchw_1[21] = 0f32
+ conv2d_nchw_1[22] = 0f32
+ conv2d_nchw_1[23] = 0f32
+ conv2d_nchw_1[24] = 0f32
+ conv2d_nchw_1[25] = 0f32
+ conv2d_nchw_1[26] = 0f32
+ conv2d_nchw_1[27] = 0f32
+ for (rc.outer.outer: int32, 0, 16) {
+ for (ry.outer.outer: int32, 0, 3) {
+ let cse_var_4: int32 = (rc.outer.outer*1568)
+ let cse_var_3: int32 = (ry.outer.outer*7)
+ 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" = 112;
+ pad_temp.shared_1: Buffer(pad_temp.shared, float32, [2016], [], scope="shared")[threadIdx.x_1] = @tir.if_then_else(((((1 <= (floordiv(floormod(threadIdx.x_1, 63), 9) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 9))) && (floormod(threadIdx.x_1, 9) < 8)), data[((((cse_var_4 + (floordiv(threadIdx.x_1, 9)*7)) + cse_var_3) + floormod(threadIdx.x_1, 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 112)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 49), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 49), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 4), 9))) && (floormod((threadIdx.x_1 + 4), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 112), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 4), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 224)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 35), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 35), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 8), 9))) && (floormod((threadIdx.x_1 + 8), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 224), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 8), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 336)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 21), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 21), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 3), 9))) && (floormod((threadIdx.x_1 + 3), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 336), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 3), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 448)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 7), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 7), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 7), 9))) && (floormod((threadIdx.x_1 + 7), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 448), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 7), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 560)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 56), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 56), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 2), 9))) && (floormod((threadIdx.x_1 + 2), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 560), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 2), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 672)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 42), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 42), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 6), 9))) && (floormod((threadIdx.x_1 + 6), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 672), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 6), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 784)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 28), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 28), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 1), 9))) && (floormod((threadIdx.x_1 + 1), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 784), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 1), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 896)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 14), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 14), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 5), 9))) && (floormod((threadIdx.x_1 + 5), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 896), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 5), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1008)] = @tir.if_then_else(((((1 <= (floordiv(floormod(threadIdx.x_1, 63), 9) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 9))) && (floormod(threadIdx.x_1, 9) < 8)), data[((((cse_var_4 + (floordiv(threadIdx.x_1, 9)*7)) + cse_var_3) + floormod(threadIdx.x_1, 9)) + 776)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1120)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 49), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 49), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 4), 9))) && (floormod((threadIdx.x_1 + 4), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1120), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 4), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1232)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 35), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 35), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 8), 9))) && (floormod((threadIdx.x_1 + 8), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1232), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 8), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1344)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 21), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 21), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 3), 9))) && (floormod((threadIdx.x_1 + 3), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1344), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 3), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1456)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 7), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 7), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 7), 9))) && (floormod((threadIdx.x_1 + 7), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1456), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 7), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1568)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 56), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 56), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 2), 9))) && (floormod((threadIdx.x_1 + 2), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1568), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 2), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1680)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 42), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 42), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 6), 9))) && (floormod((threadIdx.x_1 + 6), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1680), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 6), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1792)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 28), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 28), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 1), 9))) && (floormod((threadIdx.x_1 + 1), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1792), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 1), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1904)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 14), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 14), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 5), 9))) && (floormod((threadIdx.x_1 + 5), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1904), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 5), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_2: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1: Buffer(kernel.shared, float32, [6144], [], scope="shared")[(threadIdx.x_2*4)] = kernel[((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 1)] = kernel[((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(((floormod(threadIdx.x_2, 24)*4) + 1), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 2)] = kernel[((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(((floormod(threadIdx.x_2, 24)*4) + 2), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3)] = kernel[((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 448)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 112), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 449)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 112), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 450)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 112), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 451)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 112), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 448), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 896)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 224), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 32), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 897)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 224), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 11), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 898)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 224), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 34), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 899)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 224), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 896), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 1344)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 64512)]
+ kernel.shared_1[((threadIdx.x_2*4) + 1345)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 1), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3)) + 64512)]
+ kernel.shared_1[((threadIdx.x_2*4) + 1346)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 2), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3)) + 64512)]
+ kernel.shared_1[((threadIdx.x_2*4) + 1347)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 64512)]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 1792)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 448), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 1793)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 448), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 1794)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 448), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 1795)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 448), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 1792), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 2240)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 560), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 32), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 2241)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 560), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 11), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 2242)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 560), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 34), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 2243)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 560), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 2240), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 2688)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 129024)]
+ kernel.shared_1[((threadIdx.x_2*4) + 2689)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 1), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3)) + 129024)]
+ kernel.shared_1[((threadIdx.x_2*4) + 2690)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 2), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3)) + 129024)]
+ kernel.shared_1[((threadIdx.x_2*4) + 2691)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 129024)]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 3136)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 784), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3137)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 784), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3138)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 784), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3139)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 784), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 3136), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 3584)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 896), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 32), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3585)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 896), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 11), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3586)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 896), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 34), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3587)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 896), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 3584), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 4032)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 193536)]
+ kernel.shared_1[((threadIdx.x_2*4) + 4033)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 1), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3)) + 193536)]
+ kernel.shared_1[((threadIdx.x_2*4) + 4034)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 2), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3)) + 193536)]
+ kernel.shared_1[((threadIdx.x_2*4) + 4035)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 193536)]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 4480)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1120), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4481)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1120), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4482)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1120), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4483)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1120), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 4480), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 4928)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1232), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 32), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4929)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1232), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 11), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4930)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1232), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 34), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4931)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1232), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 4928), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 5376)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 258048)]
+ kernel.shared_1[((threadIdx.x_2*4) + 5377)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 1), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3)) + 258048)]
+ kernel.shared_1[((threadIdx.x_2*4) + 5378)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 2), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3)) + 258048)]
+ kernel.shared_1[((threadIdx.x_2*4) + 5379)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 258048)]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ if @tir.likely((threadIdx.x_2 < 80), dtype=bool) {
+ kernel.shared_1[((threadIdx.x_2*4) + 5824)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1456), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ if @tir.likely((threadIdx.x_2 < 80), dtype=bool) {
+ kernel.shared_1[((threadIdx.x_2*4) + 5825)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1456), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ if @tir.likely((threadIdx.x_2 < 80), dtype=bool) {
+ kernel.shared_1[((threadIdx.x_2*4) + 5826)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1456), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ }
+ if @tir.likely((threadIdx.x_2 < 80), dtype=bool) {
+ kernel.shared_1[((threadIdx.x_2*4) + 5827)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1456), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 5824), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ }
+ for (rc.outer.inner: int32, 0, 8) {
+ for (rx.outer.inner: int32, 0, 3) {
+ for (ff.outer.inner: int32, 0, 4) {
+ let cse_var_11: int32 = (ff.outer.inner*7)
+ let cse_var_10: int32 = (cse_var_11 + 6)
+ let cse_var_9: int32 = (cse_var_11 + 5)
+ let cse_var_8: int32 = (cse_var_11 + 4)
+ let cse_var_7: int32 = (cse_var_11 + 3)
+ let cse_var_6: int32 = (cse_var_11 + 2)
+ let cse_var_5: int32 = (cse_var_11 + 1)
+ {
+ conv2d_nchw_1[cse_var_11] = (conv2d_nchw_1[cse_var_11] + (pad_temp.shared_1[(((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_11] = (conv2d_nchw_1[cse_var_11] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 63)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_11] = (conv2d_nchw_1[cse_var_11] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 126)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_11] = (conv2d_nchw_1[cse_var_11] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 189)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_5] = (conv2d_nchw_1[cse_var_5] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 1)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_5] = (conv2d_nchw_1[cse_var_5] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 64)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_5] = (conv2d_nchw_1[cse_var_5] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 127)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_5] = (conv2d_nchw_1[cse_var_5] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 190)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_6] = (conv2d_nchw_1[cse_var_6] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 2)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_6] = (conv2d_nchw_1[cse_var_6] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 65)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_6] = (conv2d_nchw_1[cse_var_6] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 128)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_6] = (conv2d_nchw_1[cse_var_6] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 191)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_7] = (conv2d_nchw_1[cse_var_7] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 3)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_7] = (conv2d_nchw_1[cse_var_7] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 66)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_7] = (conv2d_nchw_1[cse_var_7] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 129)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_7] = (conv2d_nchw_1[cse_var_7] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 192)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_8] = (conv2d_nchw_1[cse_var_8] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 4)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_8] = (conv2d_nchw_1[cse_var_8] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 67)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_8] = (conv2d_nchw_1[cse_var_8] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 130)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_8] = (conv2d_nchw_1[cse_var_8] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 193)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_9] = (conv2d_nchw_1[cse_var_9] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 5)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_9] = (conv2d_nchw_1[cse_var_9] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 68)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_9] = (conv2d_nchw_1[cse_var_9] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 131)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_9] = (conv2d_nchw_1[cse_var_9] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 194)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_10] = (conv2d_nchw_1[cse_var_10] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 6)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_10] = (conv2d_nchw_1[cse_var_10] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 69)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_10] = (conv2d_nchw_1[cse_var_10] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 132)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_10] = (conv2d_nchw_1[cse_var_10] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 195)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ }
+ }
+ }
+ }
}
- attr [IterVar(threadIdx.x_2: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1: Buffer(kernel.shared, float32, [2304], [], scope="shared")[threadIdx.x_2] = kernel[((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 72)*4608)) + cse_var_1) + floormod(threadIdx.x_2, 72))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 224)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 224), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 8), 72), 3)*3)) + floormod((threadIdx.x_2 + 2), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 448)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 448), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 16), 72), 3)*3)) + floormod((threadIdx.x_2 + 1), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 672)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 672), 72)*4608)) + cse_var_1) + (floormod((floordiv(threadIdx.x_2, 3) + 8), 24)*3)) + floormod(threadIdx.x_2, 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 896)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 896), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 32), 72), 3)*3)) + floormod((threadIdx.x_2 + 2), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 1120)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 1120), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 40), 72), 3)*3)) + floormod((threadIdx.x_2 + 1), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 1344)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 1344), 72)*4608)) + cse_var_1) + (floormod((floordiv(threadIdx.x_2, 3) + 16), 24)*3)) + floormod(threadIdx.x_2, 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 1568)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 1568), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 56), 72), 3)*3)) + floormod((threadIdx.x_2 + 2), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 1792)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 1792), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 64), 72), 3)*3)) + floormod((threadIdx.x_2 + 1), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 2016)] = kernel[(((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 72)*4608)) + cse_var_1) + floormod(threadIdx.x_2, 72)) + 129024)]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- if @tir.likely((threadIdx.x_2 < 64), dtype=bool) {
- kernel.shared_1[(threadIdx.x_2 + 2240)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 2240), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 8), 72), 3)*3)) + floormod((threadIdx.x_2 + 2), 3))]
- }
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 7)*9)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 1)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 2)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 3)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 4)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 5)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 6)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 1)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 2)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 3)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 4)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 5)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 6)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 7)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 2)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 3)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 4)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 5)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 6)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 7)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 8)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 81)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 82)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 83)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 84)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 85)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 86)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 87)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 82)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 83)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 84)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 85)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 86)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 87)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 88)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 83)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 84)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 85)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 86)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 87)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 88)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 89)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 9)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 10)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 11)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 12)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 13)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 14)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 15)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 10)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 11)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 12)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 13)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 14)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 15)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 16)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 11)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 12)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 13)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 14)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 15)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 16)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 17)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 90)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 91)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 92)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 93)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 94)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 95)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 96)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 91)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 92)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 93)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 94)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 95)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 96)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 97)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 92)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 93)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 94)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 95)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 96)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 97)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 18)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 19)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 20)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 21)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 22)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 23)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 24)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 19)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 20)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 21)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 22)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 23)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 24)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 25)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 20)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 21)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 22)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 23)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 24)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 25)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 26)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 99)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 100)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 101)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 102)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 103)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 104)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 105)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 100)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 101)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 102)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 103)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 104)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 105)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 106)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 101)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 102)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 103)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 104)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 105)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 106)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 107)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 162)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 163)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 164)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 165)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 166)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 167)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 168)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 163)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 164)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 165)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 166)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 167)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 168)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 169)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 164)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 165)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 166)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 167)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 168)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 169)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 170)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 243)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 244)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 246)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 247)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 248)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 249)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 244)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 246)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 247)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 248)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 249)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 250)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 246)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 247)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 248)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 249)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 250)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 251)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 171)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 172)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 173)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 174)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 175)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 176)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 177)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 172)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 173)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 174)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 175)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 176)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 177)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 178)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 173)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 174)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 175)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 176)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 177)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 178)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 179)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 252)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 253)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 254)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 255)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 256)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 257)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 258)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 253)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 254)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 255)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 256)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 257)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 258)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 259)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 254)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 255)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 256)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 257)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 258)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 259)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 260)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 180)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 181)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 182)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 183)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 184)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 185)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 186)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 181)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 182)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 183)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 184)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 185)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 186)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 187)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 182)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 183)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 184)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 185)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 186)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 187)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 188)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 261)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 262)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 263)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 264)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 265)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 266)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 267)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 262)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 263)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 264)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 265)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 266)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 267)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 268)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 263)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 264)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 265)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 266)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 267)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 268)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 269)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 324)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 325)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 326)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 327)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 328)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 329)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 330)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 325)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 326)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 327)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 328)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 329)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 330)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 331)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 326)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 327)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 328)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 329)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 330)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 331)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 332)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 405)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 406)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 407)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 408)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 409)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 410)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 411)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 406)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 407)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 408)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 409)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 410)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 411)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 412)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 407)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 408)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 409)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 410)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 411)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 412)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 413)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 333)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 334)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 335)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 336)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 337)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 338)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 339)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 334)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 335)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 336)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 337)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 338)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 339)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 340)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 335)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 336)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 337)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 338)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 339)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 340)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 341)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 414)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 415)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 416)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 417)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 418)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 419)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 420)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 415)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 416)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 417)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 418)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 419)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 420)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 421)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 416)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 417)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 418)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 419)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 420)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 421)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 422)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 342)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 344)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 345)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 346)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 347)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 348)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 344)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 345)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 346)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 347)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 348)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 349)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 344)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 345)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 346)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 347)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 348)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 349)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 350)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 423)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 424)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 425)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 426)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 427)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 428)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 429)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 424)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 425)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 426)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 427)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 428)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 429)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 430)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 425)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 426)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 427)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 428)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 429)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 430)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 431)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 486)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 487)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 488)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 489)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 491)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 492)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 487)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 488)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 489)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 491)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 492)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 493)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 488)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 489)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 491)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 492)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 493)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 494)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 567)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 568)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 569)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 570)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 571)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 572)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 573)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 568)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 569)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 570)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 571)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 572)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 573)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 574)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 569)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 570)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 571)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 572)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 573)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 574)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 575)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 495)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 496)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 497)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 498)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 499)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 500)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 501)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 496)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 497)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 498)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 499)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 500)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 501)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 502)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 497)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 498)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 499)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 500)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 501)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 502)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 503)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 576)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 577)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 578)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 579)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 580)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 581)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 582)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 577)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 578)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 579)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 580)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 581)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 582)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 583)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 578)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 579)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 580)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 581)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 582)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 583)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 584)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 504)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 505)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 506)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 507)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 508)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 509)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 510)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 505)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 506)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 507)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 508)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 509)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 510)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 511)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 506)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 507)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 508)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 509)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 510)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 511)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 512)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 585)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 586)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 587)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 589)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 590)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 591)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 586)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 587)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 589)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 590)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 591)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 592)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 587)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 589)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 590)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 591)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 592)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 593)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
}
}
- for (i3.inner: int32, 0, 7) {
- compute[(((blockIdx.x*1568) + (threadIdx.x*7)) + i3.inner)] = max((conv2d_nchw_1[i3.inner] + bias[((blockIdx.x*32) + floordiv(threadIdx.x, 7))]), 0f32)
+ for (i1.inner: int32, 0, 4) {
+ for (i3.inner: int32, 0, 7) {
+ compute[(((((blockIdx.x*3136) + (floordiv(threadIdx.x, 7)*196)) + (i1.inner*49)) + (floormod(threadIdx.x, 7)*7)) + i3.inner)] = max((conv2d_nchw_1[((i1.inner*7) + i3.inner)] + bias[(((blockIdx.x*64) + (floordiv(threadIdx.x, 7)*4)) + i1.inner)]), 0f32)
+ }
}
}
}
@@ -850,7 +512,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 0.248 ms
+ Execution time of this operator: 0.420 ms
@@ -899,29 +561,29 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
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=1)
- conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=1)
- conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=32)
+ conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=4)
+ conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=16)
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=1)
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)
conv2d_nchw_yy_o_o_o_o, conv2d_nchw_yy_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_o_o_i, factor=1)
- conv2d_nchw_xx_o_i, conv2d_nchw_xx_i = s[conv2d_nchw].split(conv2d_nchw_xx, factor=7)
- 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_i, conv2d_nchw_xx_i = s[conv2d_nchw].split(conv2d_nchw_xx, factor=1)
+ conv2d_nchw_xx_o_o_i, conv2d_nchw_xx_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_i, factor=7)
conv2d_nchw_xx_o_o_o_i, conv2d_nchw_xx_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_i, factor=1)
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=2)
- conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=4)
+ conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=4)
+ conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=8)
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=3)
- conv2d_nchw_rx_o_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=3)
- conv2d_nchw_rx_o_o, conv2d_nchw_rx_o_i = s[conv2d_nchw].split(conv2d_nchw_rx_o_i, 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=1)
+ conv2d_nchw_rx_o_o, conv2d_nchw_rx_o_i = s[conv2d_nchw].split(conv2d_nchw_rx_o_i, factor=3)
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=1)
- compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=32)
+ compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=4)
+ compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=16)
compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=1)
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)
@@ -945,16 +607,16 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
compute_i0_o_i_i1_o_i_fused_i2_o_i_fused_i3_o_i_fused = s[compute].fuse(compute_i0_o_i, compute_i1_o_i, compute_i2_o_i, compute_i3_o_i)
s[compute].bind(compute_i0_o_i_i1_o_i_fused_i2_o_i_fused_i3_o_i_fused, te.thread_axis("threadIdx.x"))
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[kernel_shared].fuse(kernel_shared_ax0, kernel_shared_ax1, kernel_shared_ax2, kernel_shared_ax3)
- kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
+ kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=4)
s[kernel_shared].vectorize(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
- kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=224)
+ kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=112)
s[kernel_shared].bind(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[pad_temp_shared].fuse(pad_temp_shared_ax0, pad_temp_shared_ax1, pad_temp_shared_ax2, pad_temp_shared_ax3)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
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=224)
+ 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=112)
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", 512)
+ s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 64)
s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
@@ -972,10 +634,10 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
#define int64_t long long
#define uint64_t unsigned long long
#endif
- extern "C" __global__ void __launch_bounds__(224) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
- float conv2d_nchw[7];
- __shared__ float pad_temp_shared[648];
- __shared__ float kernel_shared[2304];
+ extern "C" __global__ void __launch_bounds__(112) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
+ float conv2d_nchw[28];
+ __shared__ float pad_temp_shared[2016];
+ __shared__ float kernel_shared[6144];
conv2d_nchw[0] = 0.000000e+00f;
conv2d_nchw[1] = 0.000000e+00f;
conv2d_nchw[2] = 0.000000e+00f;
@@ -983,534 +645,153 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
conv2d_nchw[4] = 0.000000e+00f;
conv2d_nchw[5] = 0.000000e+00f;
conv2d_nchw[6] = 0.000000e+00f;
- for (int rc_outer_outer = 0; rc_outer_outer < 64; ++rc_outer_outer) {
- __syncthreads();
- pad_temp_shared[((int)threadIdx.x)] = (((((9 <= (((int)threadIdx.x) % 81)) && ((((int)threadIdx.x) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + ((((int)threadIdx.x) / 81) * 49)) + (((((int)threadIdx.x) % 81) / 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((9 <= ((((int)threadIdx.x) + 62) % 81)) && (((((int)threadIdx.x) + 62) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 224) / 81) * 49)) + ((((((int)threadIdx.x) + 62) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- if (((int)threadIdx.x) < 200) {
- pad_temp_shared[(((int)threadIdx.x) + 448)] = (((((9 <= ((((int)threadIdx.x) + 43) % 81)) && (((((int)threadIdx.x) + 43) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 448) / 81) * 49)) + ((((((int)threadIdx.x) + 43) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- }
- kernel_shared[((int)threadIdx.x)] = kernel[((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72))];
- kernel_shared[(((int)threadIdx.x) + 224)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 224) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 448)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 448) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 672)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 672) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) / 3) + 8) % 24) * 3)) + (((int)threadIdx.x) % 3))];
- kernel_shared[(((int)threadIdx.x) + 896)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 896) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1120)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 1120) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 1344) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) / 3) + 16) % 24) * 3)) + (((int)threadIdx.x) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1568)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 1568) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1792)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 1792) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2016)] = kernel[(((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 129024)];
- if (((int)threadIdx.x) < 64) {
- kernel_shared[(((int)threadIdx.x) + 2240)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 2240) / 72) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) + 8) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ conv2d_nchw[7] = 0.000000e+00f;
+ conv2d_nchw[8] = 0.000000e+00f;
+ conv2d_nchw[9] = 0.000000e+00f;
+ conv2d_nchw[10] = 0.000000e+00f;
+ conv2d_nchw[11] = 0.000000e+00f;
+ conv2d_nchw[12] = 0.000000e+00f;
+ conv2d_nchw[13] = 0.000000e+00f;
+ conv2d_nchw[14] = 0.000000e+00f;
+ conv2d_nchw[15] = 0.000000e+00f;
+ conv2d_nchw[16] = 0.000000e+00f;
+ conv2d_nchw[17] = 0.000000e+00f;
+ conv2d_nchw[18] = 0.000000e+00f;
+ conv2d_nchw[19] = 0.000000e+00f;
+ conv2d_nchw[20] = 0.000000e+00f;
+ conv2d_nchw[21] = 0.000000e+00f;
+ conv2d_nchw[22] = 0.000000e+00f;
+ conv2d_nchw[23] = 0.000000e+00f;
+ conv2d_nchw[24] = 0.000000e+00f;
+ conv2d_nchw[25] = 0.000000e+00f;
+ conv2d_nchw[26] = 0.000000e+00f;
+ conv2d_nchw[27] = 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();
+ pad_temp_shared[((int)threadIdx.x)] = (((((1 <= (((((int)threadIdx.x) % 63) / 9) + ry_outer_outer)) && ((((((int)threadIdx.x) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + ((((int)threadIdx.x) / 9) * 7)) + (ry_outer_outer * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 112)] = (((((1 <= ((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 112) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((1 <= ((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 224) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 336)] = (((((1 <= ((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 336) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 448)] = (((((1 <= ((((((int)threadIdx.x) + 7) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 7) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 448) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 560)] = (((((1 <= ((((((int)threadIdx.x) + 56) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 56) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 560) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 672)] = (((((1 <= ((((((int)threadIdx.x) + 42) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 42) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 672) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = (((((1 <= ((((((int)threadIdx.x) + 28) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 28) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 784) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 896)] = (((((1 <= ((((((int)threadIdx.x) + 14) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 14) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 896) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1008)] = (((((1 <= (((((int)threadIdx.x) % 63) / 9) + ry_outer_outer)) && ((((((int)threadIdx.x) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + ((((int)threadIdx.x) / 9) * 7)) + (ry_outer_outer * 7)) + (((int)threadIdx.x) % 9)) + 776)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1120)] = (((((1 <= ((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1120) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1232)] = (((((1 <= ((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1232) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1344)] = (((((1 <= ((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1344) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1456)] = (((((1 <= ((((((int)threadIdx.x) + 7) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 7) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1456) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1568)] = (((((1 <= ((((((int)threadIdx.x) + 56) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 56) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1568) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1680)] = (((((1 <= ((((((int)threadIdx.x) + 42) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 42) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1680) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1792)] = (((((1 <= ((((((int)threadIdx.x) + 28) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 28) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1792) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1904)] = (((((1 <= ((((((int)threadIdx.x) + 14) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 14) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1904) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
+ kernel_shared[(((int)threadIdx.x) * 4)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 448)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 449)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 450)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 451)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 448) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 896)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 32) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 897)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 11) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 898)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 34) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 899)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 896) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1344)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 64512)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1345)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3)) + 64512)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1346)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3)) + 64512)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1347)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 64512)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1792)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1793)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1794)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1795)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 1792) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2240)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 32) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2241)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 11) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2242)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 34) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2243)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 2240) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2688)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 129024)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2689)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3)) + 129024)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2690)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3)) + 129024)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2691)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 129024)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3136)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3137)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3138)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3139)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 3136) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3584)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 32) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3585)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 11) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3586)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 34) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3587)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 3584) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4032)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 193536)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4033)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3)) + 193536)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4034)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3)) + 193536)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4035)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 193536)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4480)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4481)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4482)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4483)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 4480) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4928)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 32) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4929)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 11) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4930)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 34) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4931)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 4928) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 5376)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 258048)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 5377)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3)) + 258048)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 5378)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3)) + 258048)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 5379)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 258048)];
+ if (((int)threadIdx.x) < 80) {
+ kernel_shared[((((int)threadIdx.x) * 4) + 5824)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ }
+ if (((int)threadIdx.x) < 80) {
+ kernel_shared[((((int)threadIdx.x) * 4) + 5825)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ }
+ if (((int)threadIdx.x) < 80) {
+ kernel_shared[((((int)threadIdx.x) * 4) + 5826)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ }
+ if (((int)threadIdx.x) < 80) {
+ kernel_shared[((((int)threadIdx.x) * 4) + 5827)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 5824) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ }
+ __syncthreads();
+ for (int rc_outer_inner = 0; rc_outer_inner < 8; ++rc_outer_inner) {
+ for (int rx_outer_inner = 0; rx_outer_inner < 3; ++rx_outer_inner) {
+ for (int ff_outer_inner = 0; ff_outer_inner < 4; ++ff_outer_inner) {
+ conv2d_nchw[(ff_outer_inner * 7)] = (conv2d_nchw[(ff_outer_inner * 7)] + (pad_temp_shared[(((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[(ff_outer_inner * 7)] = (conv2d_nchw[(ff_outer_inner * 7)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 63)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[(ff_outer_inner * 7)] = (conv2d_nchw[(ff_outer_inner * 7)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 126)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[(ff_outer_inner * 7)] = (conv2d_nchw[(ff_outer_inner * 7)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 189)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 1)] = (conv2d_nchw[((ff_outer_inner * 7) + 1)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 1)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 1)] = (conv2d_nchw[((ff_outer_inner * 7) + 1)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 64)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 1)] = (conv2d_nchw[((ff_outer_inner * 7) + 1)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 127)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 1)] = (conv2d_nchw[((ff_outer_inner * 7) + 1)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 190)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 2)] = (conv2d_nchw[((ff_outer_inner * 7) + 2)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 2)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 2)] = (conv2d_nchw[((ff_outer_inner * 7) + 2)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 65)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 2)] = (conv2d_nchw[((ff_outer_inner * 7) + 2)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 128)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 2)] = (conv2d_nchw[((ff_outer_inner * 7) + 2)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 191)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 3)] = (conv2d_nchw[((ff_outer_inner * 7) + 3)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 3)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 3)] = (conv2d_nchw[((ff_outer_inner * 7) + 3)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 66)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 3)] = (conv2d_nchw[((ff_outer_inner * 7) + 3)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 129)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 3)] = (conv2d_nchw[((ff_outer_inner * 7) + 3)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 192)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 4)] = (conv2d_nchw[((ff_outer_inner * 7) + 4)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 4)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 4)] = (conv2d_nchw[((ff_outer_inner * 7) + 4)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 67)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 4)] = (conv2d_nchw[((ff_outer_inner * 7) + 4)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 130)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 4)] = (conv2d_nchw[((ff_outer_inner * 7) + 4)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 193)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 5)] = (conv2d_nchw[((ff_outer_inner * 7) + 5)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 5)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 5)] = (conv2d_nchw[((ff_outer_inner * 7) + 5)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 68)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 5)] = (conv2d_nchw[((ff_outer_inner * 7) + 5)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 131)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 5)] = (conv2d_nchw[((ff_outer_inner * 7) + 5)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 194)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 6)] = (conv2d_nchw[((ff_outer_inner * 7) + 6)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 6)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 6)] = (conv2d_nchw[((ff_outer_inner * 7) + 6)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 69)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 6)] = (conv2d_nchw[((ff_outer_inner * 7) + 6)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 132)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 6)] = (conv2d_nchw[((ff_outer_inner * 7) + 6)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 195)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ }
+ }
+ }
}
- __syncthreads();
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 81)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 82)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 82)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 88)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 88)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 89)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 9)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 10)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 11)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 12)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 13)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 14)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 15)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 10)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 11)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 12)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 13)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 14)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 15)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 16)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 11)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 12)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 13)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 14)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 15)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 16)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 17)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 90)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 91)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 92)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 93)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 94)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 95)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 96)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 91)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 92)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 93)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 94)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 95)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 96)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 97)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 92)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 93)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 94)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 95)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 96)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 97)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 98)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 18)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 19)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 20)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 21)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 22)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 23)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 24)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 19)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 20)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 21)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 22)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 23)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 24)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 25)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 20)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 21)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 22)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 23)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 24)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 25)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 26)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 99)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 100)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 101)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 102)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 103)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 104)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 105)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 100)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 101)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 102)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 103)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 104)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 105)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 106)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 101)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 102)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 103)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 104)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 105)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 106)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 107)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 162)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 163)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 163)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 169)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 169)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 170)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 243)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 244)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 244)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 250)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 250)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 251)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 171)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 172)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 173)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 174)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 175)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 176)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 177)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 172)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 173)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 174)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 175)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 176)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 177)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 178)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 173)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 174)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 175)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 176)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 177)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 178)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 179)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 180)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 181)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 182)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 183)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 184)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 185)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 186)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 181)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 182)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 183)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 184)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 185)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 186)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 187)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 182)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 183)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 184)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 185)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 186)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 187)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 188)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 261)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 262)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 263)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 264)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 265)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 266)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 267)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 262)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 263)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 264)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 265)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 266)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 267)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 268)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 263)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 264)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 265)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 266)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 267)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 268)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 269)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 324)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 325)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 325)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 331)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 331)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 332)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 405)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 406)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 406)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 412)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 412)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 413)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 333)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 334)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 335)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 336)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 337)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 338)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 339)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 334)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 335)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 336)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 337)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 338)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 339)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 340)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 335)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 336)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 337)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 338)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 339)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 340)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 341)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 414)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 415)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 416)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 417)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 418)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 419)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 420)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 415)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 416)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 417)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 418)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 419)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 420)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 421)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 416)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 417)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 418)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 419)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 420)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 421)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 422)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 342)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 343)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 344)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 345)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 346)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 347)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 348)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 343)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 344)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 345)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 346)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 347)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 348)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 349)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 344)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 345)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 346)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 347)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 348)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 349)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 350)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 423)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 424)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 425)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 426)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 427)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 428)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 429)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 424)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 425)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 426)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 427)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 428)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 429)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 430)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 425)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 426)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 427)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 428)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 429)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 430)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 431)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 486)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 487)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 488)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 489)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 490)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 491)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 492)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 487)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 488)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 489)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 490)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 491)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 492)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 493)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 488)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 489)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 490)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 491)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 492)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 493)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 494)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 567)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 568)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 569)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 570)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 571)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 572)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 573)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 568)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 569)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 570)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 571)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 572)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 573)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 574)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 569)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 570)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 571)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 572)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 573)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 574)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 575)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 495)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 496)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 497)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 498)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 499)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 500)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 501)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 496)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 497)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 498)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 499)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 500)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 501)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 502)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 497)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 498)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 499)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 500)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 501)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 502)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 503)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 576)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 577)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 578)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 579)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 580)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 581)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 582)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 577)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 578)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 579)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 580)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 581)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 582)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 583)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 578)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 579)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 580)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 581)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 582)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 583)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 584)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 504)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 505)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 506)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 507)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 508)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 509)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 510)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 505)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 506)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 507)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 508)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 509)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 510)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 511)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 506)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 507)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 508)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 509)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 510)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 511)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 512)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 585)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 586)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 587)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 588)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 589)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 590)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 591)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 586)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 587)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 588)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 589)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 590)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 591)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 592)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 587)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 588)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 589)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 590)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 591)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 592)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 593)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
}
- for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
- compute[(((((int)blockIdx.x) * 1568) + (((int)threadIdx.x) * 7)) + i3_inner)] = max((conv2d_nchw[i3_inner] + bias[((((int)blockIdx.x) * 32) + (((int)threadIdx.x) / 7))]), 0.000000e+00f);
+ for (int i1_inner = 0; i1_inner < 4; ++i1_inner) {
+ for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
+ compute[(((((((int)blockIdx.x) * 3136) + ((((int)threadIdx.x) / 7) * 196)) + (i1_inner * 49)) + ((((int)threadIdx.x) % 7) * 7)) + i3_inner)] = max((conv2d_nchw[((i1_inner * 7) + i3_inner)] + bias[(((((int)blockIdx.x) * 64) + ((((int)threadIdx.x) / 7) * 4)) + i1_inner)]), 0.000000e+00f);
+ }
}
}
@@ -1572,7 +853,7 @@ 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 20.355 seconds)
+ **Total running time of the script:** ( 3 minutes 22.148 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 929998645..2420531ea 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)
- 9.7893 9.8049 9.8140 9.7491 0.0287
+ 9.9589 9.9571 9.9964 9.9231 0.0299
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 749b3fb38..75c69b47e 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)
- 754.1263 754.0259 754.4682 753.8849 0.2485
+ 761.6540 760.7114 764.0088 760.2418 1.6761
@@ -694,7 +694,7 @@ Other Tips
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 22.398 seconds)
+ **Total running time of the script:** ( 1 minutes 23.959 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 47d379c6c..289ceb1b7 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
@@ -397,15 +397,15 @@ 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 = {compute_1: compute_3: Buffer(compute_2, float32, [128, 512], []), placeholder_6: placeholder_15: Buffer(placeholder_11, float32, [4916, 16, 1], []), placeholder_7: placeholder_16: Buffer(placeholder_12, int32, [4916], []), placeholder_5: placeholder_17: Buffer(placeholder_10, float32, [128, 256], []), placeholder_9: placeholder_18: Buffer(placeholder_14, float32, [128, 512], []), placeholder_8: placeholder_19: Buffer(placeholder_13, int32, [33], [])} {
- for (i0.outer.i1.outer.fused: int32, 0, 16) "parallel" {
- allocate(compute_4: Pointer(global float32), float32, [4096]), storage_scope = global {
- for (i.outer.inner: int32, 0, 16) {
+ preflattened_buffer_map = {placeholder_7: placeholder_15: Buffer(placeholder_12, int32, [4916], []), placeholder_8: placeholder_16: Buffer(placeholder_13, int32, [33], []), placeholder_6: placeholder_17: Buffer(placeholder_11, float32, [4916, 16, 1], []), placeholder_5: placeholder_18: Buffer(placeholder_10, float32, [128, 256], []), compute_1: compute_3: Buffer(compute_2, float32, [128, 512], []), placeholder_9: placeholder_19: Buffer(placeholder_14, 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 (i.outer.inner: int32, 0, 2) {
for (nb_j.inner: int32, 0, 2) {
for (i.inner.init: int32, 0, 8) {
let cse_var_1: int32 = (((i.outer.inner*256) + (i.inner.init*32)) + (nb_j.inner*16))
{
- compute_5: Buffer(compute_4, float32, [4096], [])[cse_var_1] = 0f32
+ 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
@@ -423,51 +423,51 @@ layout transformation, parallelization, vectorization, unrolling, and operator f
compute_5[(cse_var_1 + 15)] = 0f32
}
}
- for (elem_idx: int32, 0, let cse_var_2: int32 = ((i0.outer.i1.outer.fused*2) + nb_j.inner) in (placeholder_3[(cse_var_2 + 1)] - placeholder_3[cse_var_2])) {
+ 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, 8) {
let cse_var_21: int32 = (elem_idx*16)
- let cse_var_20: int32 = ((i0.outer.i1.outer.fused*2) + nb_j.inner)
- let cse_var_19: int32 = ((i.outer.inner*2048) + (i.inner*256))
- let cse_var_18: int32 = (((i.outer.inner*256) + (i.inner*32)) + (nb_j.inner*16))
- let cse_var_17: int32 = (cse_var_18 + 9)
- let cse_var_16: int32 = (cse_var_18 + 8)
- let cse_var_15: int32 = (cse_var_18 + 7)
- let cse_var_14: int32 = (cse_var_18 + 6)
- let cse_var_13: int32 = (cse_var_18 + 5)
- let cse_var_12: int32 = (cse_var_18 + 4)
- let cse_var_11: int32 = (cse_var_18 + 3)
- let cse_var_10: int32 = (cse_var_18 + 2)
- let cse_var_9: int32 = (cse_var_18 + 15)
- let cse_var_8: int32 = (cse_var_18 + 14)
- let cse_var_7: int32 = (cse_var_18 + 13)
- let cse_var_6: int32 = (cse_var_18 + 12)
- let cse_var_5: int32 = (cse_var_18 + 11)
- let cse_var_4: int32 = (cse_var_18 + 10)
- let cse_var_3: int32 = (cse_var_18 + 1)
+ let cse_var_20: int32 = ((floormod(i0.outer.i1.outer.fused, 16)*2) + nb_j.inner)
+ let cse_var_19: int32 = (((i.outer.inner*256) + (i.inner*32)) + (nb_j.inner*16))
+ let cse_var_18: int32 = (((floordiv(i0.outer.i1.outer.fused, 16)*4096) + (i.outer.inner*2048)) + (i.inner*256))
+ let cse_var_17: int32 = (cse_var_19 + 9)
+ let cse_var_16: int32 = (cse_var_19 + 8)
+ let cse_var_15: int32 = (cse_var_19 + 7)
+ let cse_var_14: int32 = (cse_var_19 + 6)
+ let cse_var_13: int32 = (cse_var_19 + 5)
+ let cse_var_12: int32 = (cse_var_19 + 4)
+ let cse_var_11: int32 = (cse_var_19 + 3)
+ let cse_var_10: int32 = (cse_var_19 + 2)
+ let cse_var_9: int32 = (cse_var_19 + 15)
+ let cse_var_8: int32 = (cse_var_19 + 14)
+ let cse_var_7: int32 = (cse_var_19 + 13)
+ let cse_var_6: int32 = (cse_var_19 + 12)
+ let cse_var_5: int32 = (cse_var_19 + 11)
+ let cse_var_4: int32 = (cse_var_19 + 10)
+ let cse_var_3: int32 = (cse_var_19 + 1)
{
- compute_5[cse_var_18] = (compute_5[cse_var_18] + (placeholder_1[((placeholder_3[cse_var_20]*16) + cse_var_21)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_3] = (compute_5[cse_var_3] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 1)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_10] = (compute_5[cse_var_10] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 2)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_11] = (compute_5[cse_var_11] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 3)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_12] = (compute_5[cse_var_12] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 4)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_13] = (compute_5[cse_var_13] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 5)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_14] = (compute_5[cse_var_14] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 6)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_15] = (compute_5[cse_var_15] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 7)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_16] = (compute_5[cse_var_16] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 8)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_17] = (compute_5[cse_var_17] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 9)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_4] = (compute_5[cse_var_4] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 10)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_5] = (compute_5[cse_var_5] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 11)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_6] = (compute_5[cse_var_6] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 12)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_7] = (compute_5[cse_var_7] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 13)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_8] = (compute_5[cse_var_8] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 14)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_9] = (compute_5[cse_var_9] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 15)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_19] = (compute_5[cse_var_19] + (placeholder_1[((placeholder_3[cse_var_20]*16) + cse_var_21)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_3] = (compute_5[cse_var_3] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 1)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_10] = (compute_5[cse_var_10] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 2)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_11] = (compute_5[cse_var_11] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 3)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_12] = (compute_5[cse_var_12] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 4)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_13] = (compute_5[cse_var_13] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 5)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_14] = (compute_5[cse_var_14] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 6)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_15] = (compute_5[cse_var_15] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 7)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_16] = (compute_5[cse_var_16] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 8)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_17] = (compute_5[cse_var_17] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 9)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_4] = (compute_5[cse_var_4] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 10)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_5] = (compute_5[cse_var_5] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 11)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_6] = (compute_5[cse_var_6] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 12)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_7] = (compute_5[cse_var_7] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 13)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_8] = (compute_5[cse_var_8] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 14)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_9] = (compute_5[cse_var_9] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 15)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
}
}
}
}
}
- for (i0.inner: int32, 0, 128) {
- let cse_var_22: int32 = ((i0.inner*512) + (i0.outer.i1.outer.fused*32))
+ 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))
}
}
@@ -524,7 +524,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 1.846 ms
+ Execution time of this operator: 1.852 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 a45ccc050..283aba782 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
@@ -5,12 +5,12 @@
Computation times
=================
-**00:46.493** total execution time for **how_to_tune_with_autotvm** files:
+**00:45.864** 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``) | 00:46.463 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_conv2d_cuda.py` (``tune_conv2d_cuda.py``) | 00:45.828 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_x86.py` (``tune_relay_x86.py``) | 00:00.016 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_x86.py` (``tune_relay_x86.py``) | 00:00.021 | 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 |
+--------------------------------------------------------------------------------------------------+-----------+--------+
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 eb556721c..ae3d084f4 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: 218.25/218.25 result: MeasureResult(costs=(0.0010607314827586207,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.160534620285034, timestamp=1660697029.041432) [('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: 0.00/218.25 result: Traceback (most recent call last):
+ No: 9 GFLOPS: 176.20/176.20 result: MeasureResult(costs=(0.0013138815888888889,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.0211830139160156, timestamp=1660714966.3254552) [('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: 0.00/176.20 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: 260.45/260.45 result: MeasureResult(costs=(0.0008888644861878454,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7346055507659912, timestamp=1660697029.9757693) [('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: 0.00/260.45 result: Traceback (most recent call last):
+ No: 11 GFLOPS: 260.82/260.82 result: MeasureResult(costs=(0.0008875853812154695,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7145953178405762, timestamp=1660714967.240541) [('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: 0.00/260.82 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: 0.00/260.45 result: Traceback (most recent call last):
+ No: 13 GFLOPS: 0.00/260.82 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: 0.00/260.45 result: Traceback (most recent call last):
+ No: 14 GFLOPS: 0.00/260.82 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: 5.29/260.45 result: MeasureResult(costs=(0.043759062,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.805462121963501, timestamp=1660697034.50367) [('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: 3.34/260.45 result: MeasureResult(costs=(0.06928359925,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.539597749710083, timestamp=1660697035.7375107) [('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: 0.00/260.45 result: Traceback (most recent call last):
+ No: 15 GFLOPS: 5.34/260.82 result: MeasureResult(costs=(0.04338321325,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.8048241138458252, timestamp=1660714971.793172) [('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: 3.35/260.82 result: MeasureResult(costs=(0.069062067,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.55722713470459, timestamp=1660714973.0338068) [('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: 0.00/260.82 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: 28.06/260.45 result: MeasureResult(costs=(0.008249963714285715,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.280989170074463, timestamp=1660697046.7702246) [('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: 0.00/260.45 result: Traceback (most recent call last):
+ No: 18 GFLOPS: 28.06/260.82 result: MeasureResult(costs=(0.008250961,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.285449504852295, timestamp=1660714984.027661) [('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: 0.00/260.82 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: 0.00/260.45 result: Traceback (most recent call last):
+ No: 20 GFLOPS: 0.00/260.82 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: 0.001275
+ Time cost of this operator: 0.001243
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 8973b5fc5..3c1865766 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 313.4 98.744 (1, 2, 10, 10, 3) 2 1 [313.4]
- tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.018 0.951 (1, 6, 10, 10) 1 1 [3.018]
- tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.97 0.306 (1, 1, 10, 10, 3) 1 1 [0.97]
- Total_time - 317.388 - - - - -
+ tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 315.6 98.728 (1, 2, 10, 10, 3) 2 1 [315.6]
+ tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.094 0.968 (1, 6, 10, 10) 1 1 [3.094]
+ tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.971 0.304 (1, 1, 10, 10, 3) 1 1 [0.971]
+ Total_time - 319.665 - - - - -
@@ -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 223.2 98.704 (1, 1, 10, 10, 6) 2 1 [223.2]
- tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.971 0.871 (1, 6, 10, 10) 1 1 [1.971]
- tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.96 0.425 (1, 1, 10, 10, 3) 1 1 [0.96]
- Total_time - 226.131 - - - - -
+ tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 81.625 96.775 (1, 6, 10, 10, 1) 2 1 [81.625]
+ tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.767 2.095 (1, 6, 10, 10) 1 1 [1.767]
+ tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.953 1.13 (1, 1, 10, 10, 3) 1 1 [0.953]
+ Total_time - 84.345 - - - - -
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 4d364a9db..5645fb23e 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/tmp2y4e1w0m/images/random'
+ '/tmp/tmpptt6e863/images/random'
@@ -325,8 +325,8 @@ objects to other stuff? We can display some examples from our datasets using ``m
.. code-block:: none
- /tmp/tmp2y4e1w0m/images/target contains 8144 images
- /tmp/tmp2y4e1w0m/images/random contains 5000 images
+ /tmp/tmpptt6e863/images/target contains 8144 images
+ /tmp/tmpptt6e863/images/random contains 5000 images
@@ -501,13 +501,13 @@ the time on our validation set).
.. code-block:: none
Epoch 1/3
- 328/328 - 55s - loss: 0.2112 - accuracy: 0.9246 - val_loss: 0.1566 - val_accuracy: 0.9524
+ 328/328 - 55s - loss: 0.2233 - accuracy: 0.9226 - val_loss: 0.1452 - val_accuracy: 0.9554
Epoch 2/3
- 328/328 - 53s - loss: 0.0963 - accuracy: 0.9641 - val_loss: 0.1274 - val_accuracy: 0.9622
+ 328/328 - 52s - loss: 0.0975 - accuracy: 0.9641 - val_loss: 0.1087 - val_accuracy: 0.9626
Epoch 3/3
- 328/328 - 52s - loss: 0.0642 - accuracy: 0.9767 - val_loss: 0.1251 - val_accuracy: 0.9611
+ 328/328 - 52s - loss: 0.0657 - accuracy: 0.9744 - val_loss: 0.0945 - val_accuracy: 0.9653
- <keras.callbacks.History object at 0x7f69a974ef50>
+ <keras.callbacks.History object at 0x7fc4005cac50>
@@ -864,7 +864,7 @@ 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:** ( 5 minutes 25.927 seconds)
+ **Total running time of the script:** ( 5 minutes 15.500 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 d741b9e59..b6244c9ec 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
@@ -5,20 +5,20 @@
Computation times
=================
-**06:19.787** total execution time for **how_to_work_with_microtvm** files:
+**06:08.879** total execution time for **how_to_work_with_microtvm** files:
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``) | 05:25.927 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``) | 05:15.500 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``) | 00:42.430 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``) | 00:42.359 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``) | 00:08.102 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``) | 00:07.572 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``) | 00:03.326 | 0.0 MB |
-+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_ethosu.py` (``micro_ethosu.py``) | 00:00.001 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``) | 00:03.446 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_microtvm_micro_reference_vm.py` (``micro_reference_vm.py``) | 00:00.001 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_ethosu.py` (``micro_ethosu.py``) | 00:00.001 | 0.0 MB |
++---------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tvmc.py` (``micro_tvmc.py``) | 00:00.000 | 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 ea2c9f14b..4d382072b 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
=================
-**00:42.409** total execution time for **how_to_work_with_relay** files:
+**00:42.763** 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``) | 00:30.599 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_pipeline_executor.py` (``using_pipeline_executor.py``) | 00:31.189 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``) | 00:10.171 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``) | 00:10.045 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``) | 00:01.631 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``) | 00:01.522 | 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 e73e4be68..e0795d187 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 0x7f69307e8a70>
+ <function my_cuda_math_rule at 0x7fc3980a5440>
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 7111f0969..9a23a8ef5 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
@@ -5,22 +5,22 @@
Computation times
=================
-**00:04.372** total execution time for **how_to_work_with_schedules** files:
+**00:03.998** total execution time for **how_to_work_with_schedules** files:
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``) | 00:02.018 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``) | 00:01.854 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``) | 00:01.062 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``) | 00:00.955 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``) | 00:00.558 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``) | 00:00.508 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``) | 00:00.551 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``) | 00:00.492 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_schedules_extern_op.py` (``extern_op.py``) | 00:00.101 | 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_tedd.py` (``tedd.py``) | 00:00.026 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tedd.py` (``tedd.py``) | 00:00.029 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tuple_inputs.py` (``tuple_inputs.py``) | 00:00.015 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tuple_inputs.py` (``tuple_inputs.py``) | 00:00.016 | 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 97a2cf4c6..d070391fb 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/tmpd4a8juer/input0.cc'\nsource_filename = \"/tmp/tmpd4a8juer/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/tmpqz7f6hvr/input0.cc'\nsource_filename = \"/tmp/tmpqz7f6hvr/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 86ab45ff9..4619a9037 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
=================
-**00:20.931** total execution time for **topic_vta_tutorials_autotvm** files:
+**00:21.953** total execution time for **topic_vta_tutorials_autotvm** files:
+---------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:20.924 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:21.946 | 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 e774039fd..3cb849ba2 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 22.58s!
+ resnet18_v1 inference graph built in 23.66s!
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 f0f4bd956..774feb643 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:411: 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 16.39s!
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 93385a9ba..222098cc3 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
=================
-**01:31.719** total execution time for **topic_vta_tutorials_frontend** files:
+**01:33.083** total execution time for **topic_vta_tutorials_frontend** files:
+------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``) | 00:48.727 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``) | 00:49.098 | 0.0 MB |
+------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:42.992 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:43.985 | 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 b346ccfb2..e9f3b1097 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
=================
-**00:03.383** total execution time for **topic_vta_tutorials_optimize** files:
+**00:03.240** total execution time for **topic_vta_tutorials_optimize** files:
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``) | 00:02.952 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``) | 00:02.861 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.431 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.379 | 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 25aa6af31..b32261d59 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
=================
-**00:00.791** total execution time for **topic_vta_tutorials** files:
+**00:00.683** total execution time for **topic_vta_tutorials** files:
+---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.422 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.368 | 0.0 MB |
+---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.369 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.315 | 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 ed6bf05e9..25c32b4a6 100644
--- a/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
+++ b/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
@@ -205,13 +205,6 @@ trials, we can load the best schedule from the log file and apply it.
-.. rst-class:: sphx-glr-script-out
-
- .. code-block:: none
-
- .T
-
-
@@ -335,7 +328,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 94.295 ms
+ Execution time of this operator: 93.235 ms
diff --git a/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt b/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt
index ea37dbd35..f4cf3e61e 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: 10.52/10.52 result: MeasureResult(costs=(0.025511030400000002,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5435776710510254, timestamp=1660695784.9513922) [('tile_y', [-1, 1]), ('tile_x', [-1, 256])],None,80
- No: 2 GFLOPS: 2.94/10.52 result: MeasureResult(costs=(0.0912931846,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6098172664642334, timestamp=1660695787.1091762) [('tile_y', [-1, 4]), ('tile_x', [-1, 8])],None,32
- No: 3 GFLOPS: 11.86/11.86 result: MeasureResult(costs=(0.022624653,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5812883377075195, timestamp=1660695787.6740446) [('tile_y', [-1, 64]), ('tile_x', [-1, 32])],None,56
- No: 4 GFLOPS: 1.73/11.86 result: MeasureResult(costs=(0.1553442216,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.6116950511932373, timestamp=1660695790.8507726) [('tile_y', [-1, 1]), ('tile_x', [-1, 4])],None,20
- No: 5 GFLOPS: 3.70/11.86 result: MeasureResult(costs=(0.07259004,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2964963912963867, timestamp=1660695792.2815092) [('tile_y', [-1, 256]), ('tile_x', [-1, 16])],None,48
- No: 6 GFLOPS: 1.81/11.86 result: MeasureResult(costs=(0.148696068,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.5013206005096436, timestamp=1660695795.3546283) [('tile_y', [-1, 512]), ('tile_x', [-1, 4])],None,29
- No: 7 GFLOPS: 0.87/11.86 result: MeasureResult(costs=(0.3076598544,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.04739236831665, timestamp=1660695800.4520345) [('tile_y', [-1, 512]), ('tile_x', [-1, 2])],None,19
- No: 8 GFLOPS: 10.70/11.86 result: MeasureResult(costs=(0.0250991456,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5425457954406738, timestamp=1660695801.0157876) [('tile_y', [-1, 4]), ('tile_x', [-1, 64])],None,62
- No: 9 GFLOPS: 1.91/11.86 result: MeasureResult(costs=(0.14062305819999998,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.347400665283203, timestamp=1660695803.481642) [('tile_y', [-1, 2]), ('tile_x', [-1, 2])],None,11
- No: 10 GFLOPS: 2.79/11.86 result: MeasureResult(costs=(0.0961699172,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6440229415893555, timestamp=1660695805.1829948) [('tile_y', [-1, 4]), ('tile_x', [-1, 4])],None,22
+ No: 1 GFLOPS: 10.61/10.61 result: MeasureResult(costs=(0.025309961000000002,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5380620956420898, timestamp=1660713726.9657187) [('tile_y', [-1, 1]), ('tile_x', [-1, 256])],None,80
+ No: 2 GFLOPS: 2.95/10.61 result: MeasureResult(costs=(0.091052301,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6047918796539307, timestamp=1660713728.5838535) [('tile_y', [-1, 4]), ('tile_x', [-1, 8])],None,32
+ No: 3 GFLOPS: 11.88/11.88 result: MeasureResult(costs=(0.022603546600000003,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5476012229919434, timestamp=1660713729.6313217) [('tile_y', [-1, 64]), ('tile_x', [-1, 32])],None,56
+ No: 4 GFLOPS: 1.86/11.88 result: MeasureResult(costs=(0.1443650432,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.4363057613372803, timestamp=1660713732.6374052) [('tile_y', [-1, 1]), ('tile_x', [-1, 4])],None,20
+ No: 5 GFLOPS: 3.67/11.88 result: MeasureResult(costs=(0.07323907040000001,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.3236827850341797, timestamp=1660713734.0952196) [('tile_y', [-1, 256]), ('tile_x', [-1, 16])],None,48
+ No: 6 GFLOPS: 1.79/11.88 result: MeasureResult(costs=(0.1503416152,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.5826635360717773, timestamp=1660713736.7138216) [('tile_y', [-1, 512]), ('tile_x', [-1, 4])],None,29
+ No: 7 GFLOPS: 0.87/11.88 result: MeasureResult(costs=(0.3085500146,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.06335973739624, timestamp=1660713742.3658824) [('tile_y', [-1, 512]), ('tile_x', [-1, 2])],None,19
+ No: 8 GFLOPS: 10.40/11.88 result: MeasureResult(costs=(0.025818131799999998,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5601685047149658, timestamp=1660713742.942604) [('tile_y', [-1, 4]), ('tile_x', [-1, 64])],None,62
+ No: 9 GFLOPS: 1.80/11.88 result: MeasureResult(costs=(0.1491520388,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.494178295135498, timestamp=1660713745.5578163) [('tile_y', [-1, 2]), ('tile_x', [-1, 2])],None,11
+ No: 10 GFLOPS: 2.79/11.88 result: MeasureResult(costs=(0.096178762,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6495332717895508, timestamp=1660713747.2624712) [('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 543dfadf5..50f332132 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': 493.82212321001134, 'median': 493.62453624999034, 'std': 2.9124872138303655}
+ {'mean': 495.14578981999875, 'median': 494.7857773499891, 'std': 2.9163719179460474}
@@ -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: 17.43/ 17.43 GFLOPS | Progress: (4/20) | 7.01 s
[Task 1/25] Current/Best: 6.14/ 17.43 GFLOPS | Progress: (8/20) | 9.44 s
[Task 1/25] Current/Best: 11.51/ 22.68 GFLOPS | Progress: (12/20) | 11.87 s
[Task 1/25] Current/Best: 16.72/ 22.78 GFLOPS | Progress: (16/20) | 13.57 s
[Task 1/25] Current/Best: 11.60/ 23.78 GFLOPS | Progress: (20/20) | 15.33 s Done.
-
[Task 2/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 2/25] Current/Best: 12.29/ 12.86 GFLOPS | Progress: (4/20) | 3.70 s
[Task 2/25] Current/Best: 14.18/ 18.04 GFLOPS | Progress: (8/20) | 5.01 s
[Task 2/25] Current/Best: 20.92/ 20.92 GFLOPS | Progress: (12/20) | 6.39 s
[Task 2/25] Current/Best: 12.45/ 20.92 GFLOPS | Progress: (16/20) | 7.66 s
[Task 2/25] Current/Best: 19.07/ 20.92 GFLOPS | Progress: (20/20) | 9.28 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/ 10.56 GFLOPS | Progress: (4/20) | 5.94 s
[Task 3/25] Current/Best: 15.53/ 16.86 GFLOPS | Progress: (8/20) | 7.87 s
[Task 3/25] Current/Best: 14.76/ 16.86 GFLOPS | Progress: (12/20) | 9.61 s
[Task 3/25] Current/Best: 7.18/ 23.78 GFLOPS | Progress: (16/20) | 11.62 s
[Task 3/25] Current/Best: 12.66/ 23.78 GFLOPS | Progress: (20/20) | 16.15 s Done.
-
[Task 4/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 4/25] Current/Best: 9.55/ 20.38 GFLOPS | Progress: (4/20) | 2.42 s
[Task 4/25] Current/Best: 6.82/ 20.38 GFLOPS | Progress: (8/20) | 6.74 s
[Task 4/25] Current/Best: 22.36/ 22.36 GFLOPS | Progress: (12/20) | 11.14 s
[Task 4/25] Current/Best: 17.42/ 22.36 GFLOPS | Progress: (16/20) | 13.38 s
[Task 4/25] Current/Best: 13.50/ 22.36 GFLOPS | Progress: (20/20) | 15.37 s Done.
-
[Task 5/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 5/25] Current/Best: 9.81/ 10.47 GFLOPS | Progress: (4/20) | 2.62 s
[Task 5/25] Current/Best: 11.88/ 11.88 GFLOPS | Progress: (8/20) | 4.70 s
[Task 5/25] Current/Best: 11.72/ 18.00 GFLOPS | Progress: (12/20) | 7.62 s
[Task 5/25] Current/Best: 11.88/ 22.65 GFLOPS | Progress: (16/20) | 9.05 s
[Task 5/25] Current/Best: 12.07/ 22.65 GFLOPS | Progress: (20/20) | 10.92 s Done.
-
[Task 6/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 6/25] Current/Best: 12.34/ 20.75 GFLOPS | Progress: (4/20) | 4.01 s
[Task 6/25] Current/Best: 18.95/ 20.75 GFLOPS | Progress: (8/20) | 5.79 s
[Task 6/25] Current/Best: 13.30/ 20.75 GFLOPS | Progress: (12/20) | 7.72 s
[Task 6/25] Current/Best: 20.07/ 20.75 GFLOPS | Progress: (16/20) | 9.96 s
[Task 6/25] Current/Best: 3.69/ 20.75 GFLOPS | Progress: (20/20) | 12.51 s Done.
-
[Task 7/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 7/25] Current/Best: 11.20/ 12.85 GFLOPS | Progress: (4/20) | 3.67 s
[Task 7/25] Current/Best: 20.24/ 21.00 GFLOPS | Progress: (8/20) | 5.20 s
[Task 7/25] Current/Best: 15.92/ 21.00 GFLOPS | Progress: (12/20) | 7.17 s
[Task 7/25] Current/Best: 12.25/ 21.00 GFLOPS | Progress: (16/20) | 9.22 s
[Task 7/25] Current/Best: 6.33/ 21.78 GFLOPS | Progress: (20/20) | 11.68 s Done.
-
[Task 8/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 8/25] Current/Best: 9.75/ 14.63 GFLOPS | Progress: (4/20) | 2.92 s
[Task 8/25] Current/Best: 9.89/ 14.63 GFLOPS | Progress: (8/20) | 7.60 s
[Task 8/25] Current/Best: 12.60/ 14.63 GFLOPS | Progress: (12/20) | 13.60 s
[Task 8/25] Current/Best: 18.83/ 18.83 GFLOPS | Progress: (16/20) | 15.69 s
[Task 8/25] Current/Best: 19.77/ 19.77 GFLOPS | Progress: (20/20) | 22.10 s Done.
-
[Task 9/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 9/25] Current/Best: 14.40/ 15.68 GFLOPS | Progress: (4/20) | 12.02 s
[Task 9/25] Current/Best: 23.45/ 23.45 GFLOPS | Progress: (8/20) | 13.88 s
[Task 9/25] Current/Best: 8.26/ 23.45 GFLOPS | Progress: (12/20) | 16.29 s
[Task 9/25] Current/Best: 18.02/ 23.45 GFLOPS | Progress: (16/20) | 18.93 s
[Task 9/25] Current/Best: 9.23/ 23.45 GFLOPS | Progress: (20/20) | 26.41 s
[Task 10/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 10/25] Current/Best: 18.14/ 18.14 GFLOPS | Progress: (4/20) | 2.62 s
[Task 10/25] Current/Best: 15.39/ 18.14 GFLOPS | Progress: (8/20) | 4.18 s
[Task 10/25] Current/Best: 12.79/ 18.74 GFLOPS | Progress: (12/20) | 5.71 s
[Task 10/25] Current/Best: 19.10/ 19.89 GFLOPS | Progress: (16/20) | 6.82 s
[Task 10/25] Current/Best: 8.71/ 19.89 GFLOPS | Progress: (20/20
) | 8.35 s Done.
-
[Task 11/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 11/25] Current/Best: 11.98/ 18.06 GFLOPS | Progress: (4/20) | 3.28 s
[Task 11/25] Current/Best: 16.98/ 18.06 GFLOPS | Progress: (8/20) | 6.02 s
[Task 11/25] Current/Best: 18.28/ 18.28 GFLOPS | Progress: (12/20) | 8.02 s
[Task 11/25] Current/Best: 13.42/ 21.17 GFLOPS | Progress: (16/20) | 10.78 s
[Task 11/25] Current/Best: 19.49/ 21.53 GFLOPS | Progress: (20/20) | 12.79 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/ 18.03 GFLOPS | Progress: (4/20) | 5.32 s
[Task 12/25] Current/Best: 5.20/ 18.03 GFLOPS | Progress: (8/20) | 9.01 s
[Task 12/25] Current/Best: 18.93/ 18.93 GFLOPS | Progress: (12/20) | 11.01 s
[Task 12/25] Current/Best: 15.43/ 18.93 GFLOPS | Progress: (16/20) | 13.75 s
[Task 12/25] Current/Best: 15.18/ 18.93 GFLOPS | Progress: (20/20) | 15.70 s Done.
-
[Task 13/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 13/25] Current/Best: 8.71/ 17.30 GFLOPS | Progress: (4/20) | 3.69 s
[Task 13/25] Current/Best: 15.68/ 20.99 GFLOPS | Progress: (8/20) | 6.12 s
[Task 13/25] Current/Best: 19.64/ 21.62 GFLOPS | Progress: (12/20) | 9.00 s
[Task 13/25] Current/Best: 12.24/ 21.62 GFLOPS | Progress: (16/20) | 12.41 s
[Task 13/25] Current/Best: 18.63/ 21.62 GFLOPS | Progress: (20/20) | 14.67 s Done.
-
[Task 14/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 14/25] Current/Best: 13.64/ 13.64 GFLOPS | Progress: (4/20) | 3.36 s
[Task 14/25] Current/Best: 6.02/ 13.64 GFLOPS | Progress: (8/20) | 5.54 s
[Task 14/25] Current/Best: 20.50/ 20.50 GFLOPS | Progress: (12/20) | 8.09 s
[Task 14/25] Current/Best: 16.67/ 20.50 GFLOPS | Progress: (16/20) | 9.75 s Done.
-
[Task 14/25] Current/Best: 16.97/ 20.50 GFLOPS | Progress: (20/20) | 11.50 s
[Task 15/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 15/25] Current/Best: 16.16/ 17.66 GFLOPS | Progress: (4/20) | 2.74 s
[Task 15/25] Current/Best: 14.42/ 17.97 GFLOPS | Progress: (8/20) | 4.04 s
[Task 15/25] Current/Best: 10.39/ 22.37 GFLOPS | Progress: (12/20) | 6.05 s
[Task 15/25] Current/Best: 20.40/ 22.37 GFLOPS | Progress: (16/20) | 9.20 s
[Task 15/25] Current/Best: 9.43/ 22.37 GFLOPS | Progress: (20/20) | 10.23 s
[Task 16/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 16/25] Current/Best: 20.94/ 20.94 GFLOPS | Progress: (4/20) | 2.95 s
[Task 16/25] Current/Best: 3.03/ 20.94 GFLOPS | Progress: (8/20) | 4.56 s
[Task 16/25] Current/Best: 19.53/ 20.94 GFLOPS | Progress: (12/20) | 5.79 s
[Task 16/25] Current/Best: 17.47/ 20.94 GFLOPS | Progress: (16/20) |
7.13 s
[Task 16/25] Current/Best: 10.13/ 22.27 GFLOPS | Progress: (20/20) | 9.17 s Done.
-
[Task 17/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 17/25] Current/Best: 12.84/ 18.81 GFLOPS | Progress: (4/20) | 4.72 s
[Task 17/25] Current/Best: 13.75/ 23.35 GFLOPS | Progress: (8/20) | 7.58 s
[Task 17/25] Current/Best: 17.26/ 23.35 GFLOPS | Progress: (12/20) | 9.67 s
[Task 17/25] Current/Best: 16.52/ 23.35 GFLOPS | Progress: (16/20) | 11.79 s
[Task 17/25] Current/Best: 10.05/ 23.35 GFLOPS | Progress: (20/20) | 13.92 s Done.
-
[Task 18/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 18/25] Current/Best: 11.31/ 17.94 GFLOPS | Progress: (4/20) | 3.69 s
[Task 18/25] Current/Best: 10.55/ 19.21 GFLOPS | Progress: (8/20) | 7.15 s
[Task 18/25] Current/Best: 19.27/ 19.27 GFLOPS | Progress: (12/20) | 9.09 s
[Task 18/25] Current/Best: 10.25/ 19.27 GFLOPS | Progress: (16/20) | 12.61 s
[Task 18/25] Current/Best: 20.70/ 20.70 GFLOPS | Progress: (20/20) | 14.14 s Done.
-
[Task 19/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 19/25] Current/Best: 7.20/ 20.32 GFLOPS | Progress: (4/20) | 6.04 s
[Task 19/25] Current/Best: 2.60/ 20.32 GFLOPS | Progress: (8/20) | 9.32 s
[Task 19/25] Current/Best: 19.53/ 21.59 GFLOPS | Progress: (12/20) | 12.07 s
[Task 19/25] Current/Best: 15.43/ 21.59 GFLOPS | Progress: (16/20) | 14.88 s
[Task 19/25] Current/Best: 2.70/ 23.68 GFLOPS | Progress: (20/20) | 17.70 s Done.
-
[Task 20/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 20/25] Current/Best: 8.99/ 15.25 GFLOPS | Progress: (4/20) | 3.37 s Done.
+
[Task 1/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 1/25] Current/Best: 17.57/ 17.57 GFLOPS | Progress: (4/20) | 6.50 s
[Task 1/25] Current/Best: 6.15/ 17.57 GFLOPS | Progress: (8/20) | 9.41 s
[Task 1/25] Current/Best: 11.50/ 22.65 GFLOPS | Progress: (12/20) | 11.87 s
[Task 1/25] Current/Best: 16.62/ 22.69 GFLOPS | Progress: (16/20) | 13.58 s
[Task 1/25] Current/Best: 11.54/ 23.83 GFLOPS | Progress: (20/20) | 15.34 s Done.
+
[Task 2/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 2/25] Current/Best: 12.33/ 13.29 GFLOPS | Progress: (4/20) | 3.90 s
[Task 2/25] Current/Best: 14.25/ 18.45 GFLOPS | Progress: (8/20) | 5.18 s
[Task 2/25] Current/Best: 20.97/ 20.97 GFLOPS | Progress: (12/20) | 6.55 s
[Task 2/25] Current/Best: 12.50/ 20.97 GFLOPS | Progress: (16/20) | 7.84 s
[Task 2/25] Current/Best: 19.93/ 20.97 GFLOPS | Progress: (20/20) | 9.46 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/ 10.51 GFLOPS | Progress: (4/20) | 5.96 s
[Task 3/25] Current/Best: 15.47/ 16.89 GFLOPS | Progress: (8/20) | 7.88 s
[Task 3/25] Current/Best: 14.89/ 16.89 GFLOPS | Progress: (12/20) | 9.63 s
[Task 3/25] Current/Best: 7.17/ 23.79 GFLOPS | Progress: (16/20) | 11.54 s
[Task 3/25] Current/Best: 11.46/ 23.79 GFLOPS | Progress: (20/20) | 16.10 s Done.
+
[Task 4/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 4/25] Current/Best: 9.56/ 20.55 GFLOPS | Progress: (4/20) | 2.38 s
[Task 4/25] Current/Best: 6.85/ 20.55 GFLOPS | Progress: (8/20) | 6.71 s
[Task 4/25] Current/Best: 21.16/ 21.16 GFLOPS | Progress: (12/20) | 11.28 s
[Task 4/25] Current/Best: 17.43/ 21.47 GFLOPS | Progress: (16/20) | 13.52 s
[Task 4/25] Current/Best: 13.54/ 21.47 GFLOPS | Progress: (20/20) | 15.52 s Done.
+
[Task 5/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 5/25] Current/Best: 9.70/ 10.50 GFLOPS | Progress: (4/20) | 2.66 s
[Task 5/25] Current/Best: 11.79/ 12.93 GFLOPS | Progress: (8/20) | 4.72 s
[Task 5/25] Current/Best: 9.54/ 18.08 GFLOPS | Progress: (12/20) | 7.70 s
[Task 5/25] Current/Best: 11.97/ 22.51 GFLOPS | Progress: (16/20) | 9.12 s
[Task 5/25] Current/Best: 10.72/ 22.51 GFLOPS | Progress: (20/20) | 11.03 s Done.
+
[Task 6/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 6/25] Current/Best: 12.27/ 20.72 GFLOPS | Progress: (4/20) | 4.10 s
[Task 6/25] Current/Best: 19.06/ 20.72 GFLOPS | Progress: (8/20) | 5.85 s
[Task 6/25] Current/Best: 12.87/ 20.72 GFLOPS | Progress: (12/20) | 7.77 s
[Task 6/25] Current/Best: 20.07/ 20.72 GFLOPS | Progress: (16/20) | 10.06 s
[Task 6/25] Current/Best: 3.76/ 20.72 GFLOPS | Progress: (20/20) | 12.62 s Done.
+
[Task 7/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 7/25] Current/Best: 10.92/ 12.92 GFLOPS | Progress: (4/20) | 3.62 s
[Task 7/25] Current/Best: 20.28/ 21.20 GFLOPS | Progress: (8/20) | 5.14 s
[Task 7/25] Current/Best: 15.78/ 21.20 GFLOPS | Progress: (12/20) | 7.09 s
[Task 7/25] Current/Best: 12.24/ 21.20 GFLOPS | Progress: (16/20) | 9.16 s
[Task 7/25] Current/Best: 6.37/ 21.61 GFLOPS | Progress: (20/20) | 11.64 s Done.
+
[Task 8/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 8/25] Current/Best: 9.95/ 14.03 GFLOPS | Progress: (4/20) | 2.93 s
[Task 8/25] Current/Best: 9.56/ 14.03 GFLOPS | Progress: (8/20) | 7.72 s
[Task 8/25] Current/Best: 12.68/ 14.03 GFLOPS | Progress: (12/20) | 13.94 s
[Task 8/25] Current/Best: 17.43/ 17.43 GFLOPS | Progress: (16/20) | 16.02 s
[Task 8/25] Current/Best: 20.08/ 20.08 GFLOPS | Progress: (20/20) | 22.51 s Done.
+
[Task 9/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 9/25] Current/Best: 14.35/ 15.87 GFLOPS | Progress: (4/20) | 11.94 s
[Task 9/25] Current/Best: 23.43/ 23.43 GFLOPS | Progress: (8/20) | 13.67 s
[Task 9/25] Current/Best: 8.24/ 23.43 GFLOPS | Progress: (12/20) | 16.08 s
[Task 9/25] Current/Best: 18.00/ 23.43 GFLOPS | Progress: (16/20) | 18.71 s
[Task 9/25] Current/Best: 9.17/ 23.43 GFLOPS | Progress: (20/20) | 26.33 s
[Task 10/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 10/25] Current/Best: 18.23/ 18.23 GFLOPS | Progress: (4/20) | 2.55 s
[Task 10/25] Current/Best: 15.47/ 18.23 GFLOPS | Progress: (8/20) | 4.12 s
[Task 10/25] Current/Best: 12.89/ 19.04 GFLOPS | Progress: (12/20) | 5.67 s
[Task 10/25] Current/Best: 19.15/ 20.46 GFLOPS | Progress: (16/20) | 6.79 s
[Task 10/25] Current/Best: 8.94/ 20.46 GFLOPS | Progress: (20/20
) | 8.33 s Done.
+
[Task 11/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 11/25] Current/Best: 12.22/ 18.09 GFLOPS | Progress: (4/20) | 3.39 s
[Task 11/25] Current/Best: 16.94/ 18.09 GFLOPS | Progress: (8/20) | 6.11 s
[Task 11/25] Current/Best: 18.15/ 18.15 GFLOPS | Progress: (12/20) | 8.12 s
[Task 11/25] Current/Best: 11.80/ 21.24 GFLOPS | Progress: (16/20) | 10.98 s
[Task 11/25] Current/Best: 19.47/ 21.49 GFLOPS | Progress: (20/20) | 13.06 s Done.
+
[Task 12/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 12/25] Current/Best: 7.73/ 18.12 GFLOPS | Progress: (4/20) | 5.39 s
[Task 12/25] Current/Best: 5.29/ 18.12 GFLOPS | Progress: (8/20) | 9.08 s
[Task 12/25] Current/Best: 18.51/ 18.90 GFLOPS | Progress: (12/20) | 11.05 s
[Task 12/25] Current/Best: 15.30/ 18.90 GFLOPS | Progress: (16/20) | 13.82 s
[Task 12/25] Current/Best: 15.23/ 19.29 GFLOPS | Progress: (20/20) | 15.76 s Done.
+
[Task 13/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 13/25] Current/Best: 8.76/ 17.31 GFLOPS | Progress: (4/20) | 3.69 s
[Task 13/25] Current/Best: 16.17/ 21.13 GFLOPS | Progress: (8/20) | 6.17 s
[Task 13/25] Current/Best: 19.54/ 21.75 GFLOPS | Progress: (12/20) | 9.07 s
[Task 13/25] Current/Best: 12.26/ 21.75 GFLOPS | Progress: (16/20) | 12.51 s
[Task 13/25] Current/Best: 18.65/ 21.75 GFLOPS | Progress: (20/20) | 14.78 s Done.
+
[Task 14/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 14/25] Current/Best: 13.49/ 13.49 GFLOPS | Progress: (4/20) | 3.33 s
[Task 14/25] Current/Best: 6.10/ 13.49 GFLOPS | Progress: (8/20) | 5.50 s
[Task 14/25] Current/Best: 20.69/ 20.69 GFLOPS | Progress: (12/20) | 8.06 s
[Task 14/25] Current/Best: 17.05/ 20.69 GFLOPS | Progress: (16/20) | 9.76 s Done.
+
[Task 14/25] Current/Best: 17.33/ 20.69 GFLOPS | Progress: (20/20) | 11.58 s
[Task 15/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 15/25] Current/Best: 16.08/ 17.50 GFLOPS | Progress: (4/20) | 2.85 s
[Task 15/25] Current/Best: 14.40/ 18.13 GFLOPS | Progress: (8/20) | 4.20 s
[Task 15/25] Current/Best: 10.32/ 22.08 GFLOPS | Progress: (12/20) | 6.24 s
[Task 15/25] Current/Best: 20.41/ 22.08 GFLOPS | Progress: (16/20) | 9.58 s
[Task 15/25] Current/Best: 9.65/ 22.08 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: 20.60/ 20.60 GFLOPS | Progress: (4/20) | 2.95 s
[Task 16/25] Current/Best: 3.00/ 20.60 GFLOPS | Progress: (8/20) | 4.57 s
[Task 16/25] Current/Best: 19.23/ 20.60 GFLOPS | Progress: (12/20) | 5.78 s
[Task 16/25] Current/Best: 17.55/ 20.60 GFLOPS | Progress: (16/20) |
7.15 s
[Task 16/25] Current/Best: 10.03/ 22.25 GFLOPS | Progress: (20/20) | 9.20 s Done.
+
[Task 17/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 17/25] Current/Best: 13.03/ 18.79 GFLOPS | Progress: (4/20) | 4.84 s
[Task 17/25] Current/Best: 14.48/ 23.41 GFLOPS | Progress: (8/20) | 7.59 s
[Task 17/25] Current/Best: 16.97/ 23.41 GFLOPS | Progress: (12/20) | 9.66 s
[Task 17/25] Current/Best: 16.54/ 23.41 GFLOPS | Progress: (16/20) | 11.76 s
[Task 17/25] Current/Best: 10.03/ 23.41 GFLOPS | Progress: (20/20) | 13.87 s Done.
+
[Task 18/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 18/25] Current/Best: 11.02/ 17.11 GFLOPS | Progress: (4/20) | 3.73 s
[Task 18/25] Current/Best: 10.55/ 19.16 GFLOPS | Progress: (8/20) | 7.25 s
[Task 18/25] Current/Best: 19.45/ 19.45 GFLOPS | Progress: (12/20) | 9.18 s
[Task 18/25] Current/Best: 10.16/ 19.45 GFLOPS | Progress: (16/20) | 12.78 s
[Task 18/25] Current/Best: 20.58/ 20.58 GFLOPS | Progress: (20/20) | 14.31 s Done.
+
[Task 19/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 19/25] Current/Best: 7.19/ 20.32 GFLOPS | Progress: (4/20) | 6.13 s
[Task 19/25] Current/Best: 2.60/ 20.32 GFLOPS | Progress: (8/20) | 9.41 s
[Task 19/25] Current/Best: 19.02/ 21.77 GFLOPS | Progress: (12/20) | 12.14 s
[Task 19/25] Current/Best: 14.56/ 21.77 GFLOPS | Progress: (16/20) | 14.92 s
[Task 19/25] Current/Best: 2.69/ 23.61 GFLOPS | Progress: (20/20) | 17.69 s Done.
+
[Task 20/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 20/25] Current/Best: 9.41/ 15.35 GFLOPS | Progress: (4/20) | 3.35 s Done.
Done.
-
[Task 20/25] Current/Best: 9.69/ 15.25 GFLOPS | Progress: (8/20) | 6.81 s
[Task 20/25] Current/Best: 2.32/ 16.55 GFLOPS | Progress: (12/20) | 10.68 s
[Task 20/25] Current/Best: 12.42/ 16.55 GFLOPS | Progress: (16/20) | 14.38 s
[Task 20/25] Current/Best: 12.02/ 22.09 GFLOPS | Progress: (20/20) | 16.50 s
[Task 21/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 21/25] Current/Best: 6.42/ 17.71 GFLOPS | Progress: (4/20) | 3.26 s
[Task 21/25] Current/Best: 14.67/ 17.71 GFLOPS | Progress: (8/20) | 4.79 s
[Task 21/25] Current/Best: 1.61/ 17.71 GFLOPS | Progress: (12/20) | 6.94 s
[Task 21/25] Current/Best: 18.10/ 18.10 GFLOPS | Progress: (16/20) | 10.37 s
[Task 21/25] Current/Best: 4.46/ 18.10 GFLOPS | Progress: (20/20) | 17.44 s
[Task 22/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 22/25] Current/Best: 2.70/ 17.02 GFLOPS | Progress: (4/20
) | 2.72 s
[Task 22/25] Current/Best: 8.62/ 22.00 GFLOPS | Progress: (8/20) | 4.64 s
[Task 22/25] Current/Best: 20.05/ 22.00 GFLOPS | Progress: (12/20) | 6.91 s
[Task 22/25] Current/Best: 15.56/ 22.00 GFLOPS | Progress: (16/20) | 8.96 s
[Task 22/25] Current/Best: 14.02/ 22.00 GFLOPS | Progress: (20/20) | 10.69 s Done.
-
[Task 23/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 23/25] Current/Best: 17.63/ 20.59 GFLOPS | Progress: (4/20) | 3.27 s
[Task 23/25] Current/Best: 15.57/ 20.59 GFLOPS | Progress: (8/20) | 6.62 s
[Task 23/25] Current/Best: 20.88/ 21.59 GFLOPS | Progress: (12/20) | 8.44 s
[Task 23/25] Current/Best: 6.42/ 21.59 GFLOPS | Progress: (16/20) | 15.37 s
[Task 23/25] Current/Best: 7.80/ 21.59 GFLOPS | Progress: (20/20) | 19.61 s Done.
-
[Task 24/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 24/25] Current/Best: 8.59/ 8.59 GFLOPS | Progress: (4/20) | 11.82 s
[Task 24/25] Current/Best: 2.15/ 8.59 GFLOPS | Progress: (8/20) | 22.84 s
[Task 24/25] Current/Best: 4.42/ 8.59 GFLOPS | Progress: (12/20) | 34.40 s Done.
-
[Task 24/25] Current/Best: 6.16/ 8.87 GFLOPS | Progress: (16/20) | 39.71 s
[Task 24/25] Current/Best: 3.39/ 8.87 GFLOPS | Progress: (20/20) | 45.50 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/ 2.85 GFLOPS | Progress: (4/20) | 11.65 s
[Task 25/25] Current/Best: 6.04/ 8.01 GFLOPS | Progress: (8/20) | 22.91 s
[Task 25/25] Current/Best: 5.86/ 8.01 GFLOPS | Progress: (12/20) | 34.40 s
[Task 25/25] Current/Best: 5.77/ 8.77 GFLOPS | Progress: (16/20) | 36.22 s
[Task 25/25] Current/Best: 2.87/ 8.97 GFLOPS | Progress: (20/20) | 46.94 s
+
[Task 20/25] Current/Best: 9.84/ 15.35 GFLOPS | Progress: (8/20) | 6.78 s
[Task 20/25] Current/Best: 2.33/ 15.51 GFLOPS | Progress: (12/20) | 10.74 s
[Task 20/25] Current/Best: 12.41/ 15.51 GFLOPS | Progress: (16/20) | 14.47 s
[Task 20/25] Current/Best: 11.36/ 22.20 GFLOPS | Progress: (20/20) | 16.57 s
[Task 21/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 21/25] Current/Best: 6.40/ 17.78 GFLOPS | Progress: (4/20) | 3.30 s
[Task 21/25] Current/Best: 14.65/ 17.78 GFLOPS | Progress: (8/20) | 4.86 s
[Task 21/25] Current/Best: 1.61/ 17.78 GFLOPS | Progress: (12/20) | 7.02 s
[Task 21/25] Current/Best: 17.92/ 17.92 GFLOPS | Progress: (16/20) | 10.48 s
[Task 21/25] Current/Best: 4.47/ 17.92 GFLOPS | Progress: (20/20) | 17.44 s
[Task 22/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 22/25] Current/Best: 2.70/ 17.03 GFLOPS | Progress: (4/20
) | 2.68 s
[Task 22/25] Current/Best: 9.15/ 21.20 GFLOPS | Progress: (8/20) | 4.58 s
[Task 22/25] Current/Best: 20.12/ 21.20 GFLOPS | Progress: (12/20) | 6.89 s
[Task 22/25] Current/Best: 15.30/ 21.20 GFLOPS | Progress: (16/20) | 8.92 s
[Task 22/25] Current/Best: 14.74/ 21.20 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: 17.61/ 20.10 GFLOPS | Progress: (4/20) | 3.30 s
[Task 23/25] Current/Best: 15.07/ 20.10 GFLOPS | Progress: (8/20) | 6.56 s
[Task 23/25] Current/Best: 16.40/ 20.10 GFLOPS | Progress: (12/20) | 8.66 s
[Task 23/25] Current/Best: 5.10/ 20.10 GFLOPS | Progress: (16/20) | 16.35 s
[Task 23/25] Current/Best: 7.42/ 20.10 GFLOPS | Progress: (20/20) | 20.75 s Done.
+
[Task 24/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 24/25] Current/Best: 8.55/ 8.55 GFLOPS | Progress: (4/20) | 11.84 s
[Task 24/25] Current/Best: 2.13/ 8.55 GFLOPS | Progress: (8/20) | 22.97 s
[Task 24/25] Current/Best: 4.76/ 8.55 GFLOPS | Progress: (12/20) | 34.50 s Done.
+
[Task 24/25] Current/Best: 6.01/ 8.74 GFLOPS | Progress: (16/20) | 39.93 s
[Task 24/25] Current/Best: 3.45/ 8.82 GFLOPS | Progress: (20/20) | 45.83 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/ 2.80 GFLOPS | Progress: (4/20) | 11.60 s
[Task 25/25] Current/Best: 5.44/ 7.80 GFLOPS | Progress: (8/20) | 22.91 s
[Task 25/25] Current/Best: 5.77/ 7.80 GFLOPS | Progress: (12/20) | 34.28 s
[Task 25/25] Current/Best: 5.87/ 9.10 GFLOPS | Progress: (16/20) | 36.16 s
[Task 25/25] Current/Best: 2.86/ 9.10 GFLOPS | Progress: (20/20) | 46.87 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=0.621103
+ class='n02123159 tiger cat' with probability=0.356379
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': 412.3627621700143, 'median': 412.2131008499764, 'std': 1.20978274671407}
- unoptimized: {'mean': 493.82212321001134, 'median': 493.62453624999034, 'std': 2.9124872138303655}
+ optimized: {'mean': 416.093456170006, 'median': 416.4144474499835, 'std': 2.0205042965037308}
+ unoptimized: {'mean': 495.14578981999875, 'median': 494.7857773499891, 'std': 2.9163719179460474}
@@ -772,7 +772,7 @@ profiling/benchmarking.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 10 minutes 16.142 seconds)
+ **Total running time of the script:** ( 10 minutes 22.993 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 642992665..f3149da93 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
- 1.254e-07 secs/op
+ 1.223e-07 secs/op
diff --git a/docs/_sources/tutorial/intro_topi.rst.txt b/docs/_sources/tutorial/intro_topi.rst.txt
index eca08ca65..90022552f 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, 0xc510c20)), stage(b, placeholder(b, 0x57ef8b0)), 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, 0x27e0a490)), stage(b, placeholder(b, 0x249fde60)), 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(mi [...]
diff --git a/docs/_sources/tutorial/sg_execution_times.rst.txt b/docs/_sources/tutorial/sg_execution_times.rst.txt
index 7b648ef83..b4592b42a 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
=================
-**13:11.372** total execution time for **tutorial** files:
+**13:10.523** total execution time for **tutorial** files:
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``) | 10:16.142 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``) | 10:22.993 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``) | 01:00.721 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``) | 00:59.113 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 00:58.197 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 00:51.901 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``) | 00:30.915 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``) | 00:31.400 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``) | 00:23.733 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``) | 00:23.745 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``) | 00:00.793 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``) | 00:00.722 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``) | 00:00.708 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``) | 00:00.498 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.154 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.141 | 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``) | 00:00.002 | 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 |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tvmc_python.py` (``tvmc_python.py``) | 00:00.001 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_install.py` (``install.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 b0a31e58e..9c4881534 100644
--- a/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
+++ b/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
@@ -302,7 +302,7 @@ helper function to run a profile of the TVM generated code.
.. code-block:: none
Numpy running time: 0.000008
- naive: 0.000020
+ naive: 0.000007
@@ -403,7 +403,7 @@ compile and run this new schedule with the parallel operation applied:
/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. "
- parallel: 0.000012
+ parallel: 0.000006
@@ -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: 0.000028
+ 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 8.041399996727706e-06 1.0
- naive 1.9528699999999998e-05 2.4285199104567403
- parallel 1.21433e-05 1.5100977447884054
- vector 2.80581e-05 3.4892058610960417
+ numpy 8.241269997597555e-06 1.0
+ naive 6.6443e-06 0.8062228275419815
+ parallel 6.0232e-06 0.7308582295878968
+ vector 2.4576399999999997e-05 2.982113194588256
@@ -936,7 +936,7 @@ matrix multiplication.
.. code-block:: none
- Numpy running time: 0.019501
+ Numpy running time: 0.018991
@@ -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: 3.393298
+ none: 3.255485
@@ -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: 0.294168
+ blocking: 0.306095
@@ -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: 0.331265
+ vectorization: 0.335504
@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: 0.116738
+ loop permutation: 0.127335
@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: 0.110640
+ array packing: 0.112244
@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: 0.110902
+ block caching: 0.113245
@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: 0.145311
+ parallelization: 0.144645
@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 3.3932977262999997 1.0
- blocking 0.29416804129999996 0.08669090219229196
- vectorization 0.3312650438 0.09762333591671202
- loop permutation 0.1167384342 0.034402650051956925
- array packing 0.11063964600000001 0.032605345868262435
- block caching 0.11090168900000001 0.03268256956660432
- parallelization 0.1453112336 0.042823013281078987
+ none 3.2554853655000002 1.0
+ blocking 0.30609518609999997 0.09402443928756157
+ vectorization 0.335503931 0.10305803692300455
+ loop permutation 0.1273347151 0.03911389571872428
+ array packing 0.1122438013 0.03447836150317352
+ block caching 0.11324539559999999 0.0347860250886451
+ parallelization 0.1446452098 0.0444312271628917
@@ -1686,11 +1686,6 @@ operations with tunable parameters that allows you to automatically optimize
the computation for specific platforms.
-.. rst-class:: sphx-glr-timing
-
- **Total running time of the script:** ( 1 minutes 0.721 seconds)
-
-
.. _sphx_glr_download_tutorial_tensor_expr_get_started.py:
.. only:: html
diff --git a/docs/commit_hash b/docs/commit_hash
index 7f4c5de59..044b01517 100644
--- a/docs/commit_hash
+++ b/docs/commit_hash
@@ -1 +1 @@
-a1ddfb592fd8c369e6b7331f4819adabcd26648b
+d2f9f254d275df256dbcbc5a9f8b3a07cee1d81f
diff --git a/docs/how_to/compile_models/from_darknet.html b/docs/how_to/compile_models/from_darknet.html
index 39afafb64..e9d75541f 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 9.820 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 7.078 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_mxnet.html b/docs/how_to/compile_models/from_mxnet.html
index e62763a6b..52f20a240 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.zip0659f513-2db6-4c25-a8ef-a104b78991b5 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.zip4b8c443c-c2ca-4cbc-9cb2-6cc91ea8e8a8 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 ff68cf32c..3497a0098 100644
--- a/docs/how_to/compile_models/from_oneflow.html
+++ b/docs/how_to/compile_models/from_oneflow.html
@@ -432,16 +432,15 @@ python3 -m pip install -f https://release.oneflow.info <span class="nv">oneflow<
<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]
- 15%|#5 | 6.33M/41.5M [00:00<00:00, 39.3MB/s]
- 24%|##4 | 10.1M/41.5M [00:00<00:01, 32.6MB/s]
- 35%|###4 | 14.3M/41.5M [00:00<00:01, 27.9MB/s]
- 41%|####1 | 17.0M/41.5M [00:00<00:01, 23.7MB/s]
- 54%|#####3 | 22.3M/41.5M [00:01<00:01, 14.1MB/s]
- 58%|#####8 | 24.1M/41.5M [00:01<00:01, 14.0MB/s]
- 77%|#######7 | 32.0M/41.5M [00:01<00:00, 24.4MB/s]
- 85%|########5 | 35.4M/41.5M [00:01<00:00, 25.1MB/s]
- 96%|#########6| 40.0M/41.5M [00:01<00:00, 29.4MB/s]
-100%|##########| 41.5M/41.5M [00:01<00:00, 24.8MB/s]
+ 15%|#5 | 6.33M/41.5M [00:00<00:01, 25.1MB/s]
+ 21%|##1 | 8.73M/41.5M [00:00<00:01, 18.6MB/s]
+ 35%|###4 | 14.3M/41.5M [00:00<00:01, 27.3MB/s]
+ 42%|####1 | 17.2M/41.5M [00:00<00:01, 24.2MB/s]
+ 58%|#####7 | 24.0M/41.5M [00:00<00:00, 31.3MB/s]
+ 77%|#######7 | 32.0M/41.5M [00:01<00:00, 39.4MB/s]
+ 87%|########6 | 35.9M/41.5M [00:01<00:00, 29.9MB/s]
+ 94%|#########4| 39.1M/41.5M [00:01<00:00, 24.7MB/s]
+100%|##########| 41.5M/41.5M [00:01<00:00, 27.3MB/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 1787c414a..04e3126cb 100644
--- a/docs/how_to/compile_models/from_pytorch.html
+++ b/docs/how_to/compile_models/from_pytorch.html
@@ -414,14 +414,15 @@ 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]
- 2%|1 | 896k/44.7M [00:00<00:05, 9.17MB/s]
- 17%|#6 | 7.49M/44.7M [00:00<00:00, 44.5MB/s]
- 33%|###2 | 14.6M/44.7M [00:00<00:00, 58.3MB/s]
- 49%|####8 | 21.8M/44.7M [00:00<00:00, 64.7MB/s]
- 65%|######5 | 29.1M/44.7M [00:00<00:00, 68.8MB/s]
- 81%|######## | 36.1M/44.7M [00:00<00:00, 70.6MB/s]
- 97%|#########7| 43.4M/44.7M [00:00<00:00, 72.2MB/s]
-100%|##########| 44.7M/44.7M [00:00<00:00, 65.1MB/s]
+ 2%|2 | 968k/44.7M [00:00<00:04, 9.87MB/s]
+ 17%|#7 | 7.73M/44.7M [00:00<00:00, 45.5MB/s]
+ 30%|##9 | 13.2M/44.7M [00:00<00:00, 50.7MB/s]
+ 43%|####3 | 19.3M/44.7M [00:00<00:00, 55.6MB/s]
+ 55%|#####5 | 24.6M/44.7M [00:00<00:00, 54.3MB/s]
+ 69%|######9 | 30.8M/44.7M [00:00<00:00, 58.0MB/s]
+ 82%|########2 | 36.8M/44.7M [00:00<00:00, 56.7MB/s]
+ 97%|#########6| 43.2M/44.7M [00:00<00:00, 59.3MB/s]
+100%|##########| 44.7M/44.7M [00:00<00:00, 54.7MB/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 b0663a021..c27c736e2 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 5.209 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 6.195 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 747c8c4ef..dd3ea53be 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>05:18.343</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
+<p><strong>05:17.453</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_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>01:09.820</p></td>
+<td><p>01:07.078</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><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>01:05.209</p></td>
+<td><p>01:06.195</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>00:39.852</p></td>
+<td><p>00:40.806</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>00:29.669</p></td>
+<td><p>00:29.287</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>00:26.729</p></td>
+<td><p>00:26.858</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>00:24.728</p></td>
+<td><p>00:26.161</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>00:23.190</p></td>
+<td><p>00:22.911</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>00:21.271</p></td>
+<td><p>00:20.628</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>00:15.385</p></td>
+<td><p>00:14.803</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>00:02.490</p></td>
+<td><p>00:02.726</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 2bc7f42a9..df75179ec 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)
- 15.9095 15.7277 16.6322 15.5471 0.3829
+ 15.6957 15.6786 15.8327 15.6209 0.0719
</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 592a4af4f..d4bdf9a6e 100644
--- a/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
+++ b/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
@@ -436,50 +436,40 @@ 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%|1 | 2.84M/170M [00:00<00:05, 29.8MB/s]
- 4%|4 | 7.62M/170M [00:00<00:04, 40.8MB/s]
- 8%|7 | 12.8M/170M [00:00<00:03, 46.6MB/s]
- 10%|# | 17.2M/170M [00:00<00:03, 43.8MB/s]
- 13%|#2 | 21.4M/170M [00:00<00:03, 39.5MB/s]
- 15%|#4 | 25.4M/170M [00:00<00:03, 40.0MB/s]
- 17%|#7 | 29.4M/170M [00:00<00:03, 40.4MB/s]
- 20%|#9 | 33.5M/170M [00:00<00:03, 41.3MB/s]
- 22%|##2 | 37.4M/170M [00:00<00:03, 40.6MB/s]
- 24%|##4 | 41.3M/170M [00:01<00:03, 39.9MB/s]
- 27%|##6 | 45.2M/170M [00:01<00:03, 37.2MB/s]
- 29%|##8 | 48.9M/170M [00:01<00:03, 37.5MB/s]
- 31%|###1 | 52.9M/170M [00:01<00:03, 38.6MB/s]
- 33%|###3 | 56.6M/170M [00:01<00:05, 21.7MB/s]
- 35%|###5 | 59.5M/170M [00:01<00:05, 21.3MB/s]
- 37%|###7 | 63.3M/170M [00:02<00:04, 24.8MB/s]
- 40%|#### | 68.0M/170M [00:02<00:03, 30.0MB/s]
- 44%|####3 | 74.0M/170M [00:02<00:02, 37.9MB/s]
- 46%|####6 | 78.2M/170M [00:02<00:02, 38.5MB/s]
- 48%|####8 | 82.3M/170M [00:02<00:02, 38.5MB/s]
- 51%|#####1 | 86.8M/170M [00:02<00:02, 40.6MB/s]
- 54%|#####3 | 90.9M/170M [00:02<00:02, 38.4MB/s]
- 56%|#####5 | 94.7M/170M [00:02<00:02, 38.6MB/s]
- 58%|#####8 | 98.5M/170M [00:02<00:02, 36.6MB/s]
- 60%|###### | 102M/170M [00:03<00:01, 36.5MB/s]
- 62%|######2 | 106M/170M [00:03<00:01, 37.1MB/s]
- 65%|######5 | 111M/170M [00:03<00:01, 40.9MB/s]
- 68%|######8 | 116M/170M [00:03<00:01, 45.9MB/s]
- 71%|#######1 | 121M/170M [00:03<00:01, 44.1MB/s]
- 74%|#######3 | 125M/170M [00:03<00:01, 40.8MB/s]
- 76%|#######5 | 129M/170M [00:03<00:01, 39.5MB/s]
- 78%|#######8 | 133M/170M [00:03<00:01, 35.9MB/s]
- 80%|######## | 136M/170M [00:03<00:01, 35.1MB/s]
- 82%|########2 | 140M/170M [00:04<00:01, 27.9MB/s]
- 84%|########4 | 143M/170M [00:04<00:00, 30.1MB/s]
- 86%|########6 | 146M/170M [00:04<00:00, 29.9MB/s]
- 88%|########7 | 149M/170M [00:04<00:00, 26.0MB/s]
- 90%|########9 | 152M/170M [00:04<00:00, 27.0MB/s]
- 91%|#########1| 155M/170M [00:04<00:00, 26.6MB/s]
- 93%|#########2| 158M/170M [00:04<00:00, 25.0MB/s]
- 95%|#########4| 161M/170M [00:04<00:00, 26.9MB/s]
- 97%|#########6| 165M/170M [00:05<00:00, 30.7MB/s]
- 99%|#########8| 168M/170M [00:05<00:00, 31.5MB/s]
-100%|##########| 170M/170M [00:05<00:00, 34.3MB/s]
+ 1%| | 928k/170M [00:00<00:18, 9.48MB/s]
+ 3%|2 | 5.09M/170M [00:00<00:05, 29.6MB/s]
+ 7%|6 | 11.2M/170M [00:00<00:03, 45.5MB/s]
+ 10%|9 | 16.2M/170M [00:00<00:03, 48.1MB/s]
+ 13%|#2 | 21.7M/170M [00:00<00:03, 51.4MB/s]
+ 16%|#5 | 26.6M/170M [00:00<00:03, 47.4MB/s]
+ 19%|#8 | 31.7M/170M [00:00<00:02, 49.2MB/s]
+ 21%|##1 | 36.5M/170M [00:01<00:04, 30.5MB/s]
+ 24%|##3 | 40.2M/170M [00:01<00:04, 31.0MB/s]
+ 27%|##6 | 45.0M/170M [00:01<00:03, 35.4MB/s]
+ 30%|##9 | 50.8M/170M [00:01<00:03, 41.3MB/s]
+ 33%|###2 | 55.2M/170M [00:01<00:03, 39.5MB/s]
+ 35%|###4 | 59.4M/170M [00:01<00:02, 38.8MB/s]
+ 37%|###7 | 63.3M/170M [00:01<00:02, 39.5MB/s]
+ 41%|#### | 69.1M/170M [00:01<00:02, 45.1MB/s]
+ 43%|####3 | 73.6M/170M [00:01<00:02, 44.6MB/s]
+ 47%|####7 | 79.8M/170M [00:02<00:01, 50.2MB/s]
+ 50%|####9 | 84.8M/170M [00:02<00:01, 47.6MB/s]
+ 53%|#####2 | 89.4M/170M [00:02<00:01, 47.3MB/s]
+ 56%|#####5 | 94.5M/170M [00:02<00:01, 49.0MB/s]
+ 58%|#####8 | 99.3M/170M [00:02<00:01, 46.4MB/s]
+ 62%|######1 | 105M/170M [00:02<00:01, 50.2MB/s]
+ 65%|######5 | 111M/170M [00:02<00:01, 53.7MB/s]
+ 68%|######8 | 116M/170M [00:02<00:01, 52.3MB/s]
+ 71%|#######1 | 121M/170M [00:02<00:00, 51.0MB/s]
+ 75%|#######5 | 127M/170M [00:02<00:00, 55.0MB/s]
+ 78%|#######8 | 133M/170M [00:03<00:00, 56.6MB/s]
+ 82%|########1 | 139M/170M [00:03<00:00, 53.3MB/s]
+ 85%|########4 | 144M/170M [00:03<00:00, 53.4MB/s]
+ 88%|########7 | 149M/170M [00:03<00:00, 52.8MB/s]
+ 91%|######### | 154M/170M [00:03<00:00, 51.3MB/s]
+ 94%|#########3| 159M/170M [00:03<00:00, 50.3MB/s]
+ 98%|#########7| 166M/170M [00:03<00:00, 56.6MB/s]
+100%|##########| 170M/170M [00:03<00:00, 47.0MB/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').
@@ -574,7 +564,7 @@ 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 58.211 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes 2.080 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 daa3ad2cd..f9d4a78ea 100644
--- a/docs/how_to/deploy_models/deploy_prequantized.html
+++ b/docs/how_to/deploy_models/deploy_prequantized.html
@@ -480,11 +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]
- 20%|## | 2.72M/13.6M [00:00<00:00, 28.5MB/s]
- 45%|####5 | 6.15M/13.6M [00:00<00:00, 32.9MB/s]
- 69%|######8 | 9.29M/13.6M [00:00<00:00, 32.4MB/s]
- 93%|#########3| 12.6M/13.6M [00:00<00:00, 33.5MB/s]
-100%|##########| 13.6M/13.6M [00:00<00:00, 32.9MB/s]
+ 7%|6 | 904k/13.6M [00:00<00:01, 9.25MB/s]
+ 55%|#####5 | 7.47M/13.6M [00:00<00:00, 44.3MB/s]
+100%|##########| 13.6M/13.6M [00:00<00:00, 49.5MB/s]
</pre></div>
</div>
</div>
@@ -573,7 +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)
- 90.2987 90.2379 91.7357 90.0738 0.2322
+ 90.2603 90.1888 91.1384 90.0533 0.2049
</pre></div>
</div>
<div class="admonition note">
@@ -612,7 +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 8.947 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 10.686 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 fe2e2df2c..92c148087 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)
- 120.1967 120.1969 120.9517 119.5407 0.3156
+ 119.1746 119.1299 121.7804 118.4658 0.4477
</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 51.089 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 51.991 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 7204d3fee..041e5fe35 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 50.354 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 41.133 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 e7191b4e9..b4bef9812 100644
--- a/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
+++ b/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
@@ -441,26 +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]
- 1%| | 1030/132723 [00:00<00:12, 10251.13KB/s]
- 6%|6 | 8176/132723 [00:00<00:02, 46182.28KB/s]
- 12%|#1 | 15534/132723 [00:00<00:01, 58676.08KB/s]
- 18%|#7 | 23262/132723 [00:00<00:01, 66011.25KB/s]
- 23%|##3 | 30920/132723 [00:00<00:01, 69817.40KB/s]
- 29%|##9 | 38542/132723 [00:00<00:01, 71991.91KB/s]
- 35%|###4 | 46180/132723 [00:00<00:01, 73424.71KB/s]
- 40%|#### | 53524/132723 [00:00<00:01, 49586.36KB/s]
- 46%|####6 | 61072/132723 [00:01<00:01, 55666.71KB/s]
- 52%|#####1 | 68723/132723 [00:01<00:01, 60903.36KB/s]
- 58%|#####7 | 76365/132723 [00:01<00:00, 65001.12KB/s]
- 63%|######2 | 83433/132723 [00:01<00:01, 29955.18KB/s]
- 68%|######8 | 90416/132723 [00:01<00:01, 35926.38KB/s]
- 73%|#######3 | 97075/132723 [00:02<00:00, 40777.62KB/s]
- 78%|#######7 | 103003/132723 [00:02<00:00, 38184.16KB/s]
- 83%|########3 | 110620/132723 [00:02<00:00, 45637.22KB/s]
- 88%|########7 | 116492/132723 [00:02<00:00, 36926.78KB/s]
- 94%|#########3| 124164/132723 [00:02<00:00, 44579.36KB/s]
- 99%|#########9| 131399/132723 [00:02<00:00, 50563.83KB/s]
-100%|##########| 132723/132723 [00:02<00:00, 48284.75KB/s]
+ 2%|2 | 2701/132723 [00:00<00:04, 27008.42KB/s]
+ 6%|5 | 7962/132723 [00:00<00:02, 41748.56KB/s]
+ 10%|# | 13715/132723 [00:00<00:02, 48922.68KB/s]
+ 16%|#6 | 21829/132723 [00:00<00:01, 61607.09KB/s]
+ 23%|##2 | 29960/132723 [00:00<00:01, 68698.04KB/s]
+ 29%|##8 | 38074/132723 [00:00<00:01, 72918.55KB/s]
+ 35%|###4 | 46279/132723 [00:00<00:01, 75898.59KB/s]
+ 41%|####1 | 54417/132723 [00:00<00:01, 77641.29KB/s]
+ 47%|####7 | 62571/132723 [00:00<00:00, 78857.03KB/s]
+ 53%|#####3 | 70686/132723 [00:01<00:00, 79563.09KB/s]
+ 59%|#####9 | 78758/132723 [00:01<00:00, 79915.83KB/s]
+ 65%|######5 | 86826/132723 [00:01<00:00, 80146.25KB/s]
+ 71%|#######1 | 94864/132723 [00:01<00:00, 80216.15KB/s]
+ 78%|#######7 | 102886/132723 [00:01<00:00, 67625.83KB/s]
+ 84%|########3 | 111087/132723 [00:01<00:00, 71459.86KB/s]
+ 89%|########9 | 118518/132723 [00:01<00:00, 61858.48KB/s]
+ 95%|#########5| 126716/132723 [00:01<00:00, 66930.73KB/s]
+100%|##########| 132723/132723 [00:01<00:00, 69919.40KB/s]
</pre></div>
</div>
<p>Create TVM runtime and do inference
@@ -503,7 +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 35.025 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 37.687 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 4988ccb82..33e39b4f7 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>11:37.040</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
+<p><strong>11:39.319</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>02:58.211</p></td>
+<td><p>03:02.080</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>02:35.025</p></td>
+<td><p>02:37.687</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>01:51.089</p></td>
+<td><p>01:51.991</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>01:50.354</p></td>
+<td><p>01:41.133</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>01:08.947</p></td>
+<td><p>01:10.686</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>00:29.455</p></td>
+<td><p>00:30.953</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>00:22.190</p></td>
+<td><p>00:22.579</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>00:21.764</p></td>
+<td><p>00:22.204</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 3c43b7bcf..2fb371678 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.zipaad45c1c-7ae2-421e-9dd3-5b38617583f4 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.zip965811c5-01ad-4e01-a130-3ec1a6e56e45 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>
@@ -676,7 +676,7 @@ In this alpha state of the Bring Your Own Datatypes framework, we have not imple
</div>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>/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. "
- Check failed: (lower) is false: FloatImm lowering function for target llvm type 150 not found
+ Check failed: (lower) is false: Intrinsic lowering function for target llvm, intrinsic name tir.sqrt, type 150 not found
</pre></div>
</div>
<p>When we attempt to run the model, we get a familiar error telling us that more functions need to be registered for myfloat.</p>
diff --git a/docs/how_to/extend_tvm/sg_execution_times.html b/docs/how_to/extend_tvm/sg_execution_times.html
index 0e2506255..31c89152f 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>00:41.104</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
+<p><strong>00:42.186</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>00:37.947</p></td>
+<td><p>00:38.864</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>00:02.235</p></td>
+<td><p>00:02.382</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>00:00.915</p></td>
+<td><p>00:00.933</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>00:00.007</p></td>
+<td><p>00:00.008</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 389d0bfa3..5613754f8 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: 6794us [6794us] (46.34%; 46.34%)
-FoldScaleAxis: 7869us [5us] (53.66%; 53.66%)
- FoldConstant: 7863us [1634us] (53.63%; 99.93%)
- InferType: 6230us [6230us] (42.49%; 79.22%)
+InferType: 6850us [6850us] (46.71%; 46.71%)
+FoldScaleAxis: 7816us [6us] (53.29%; 53.29%)
+ FoldConstant: 7810us [1580us] (53.26%; 99.93%)
+ InferType: 6230us [6230us] (42.48%; 79.77%)
</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: 6309us [6309us] (44.78%; 44.78%)
-FoldScaleAxis: 7781us [5us] (55.22%; 55.22%)
- FoldConstant: 7777us [1595us] (55.19%; 99.94%)
- InferType: 6182us [6182us] (43.87%; 79.49%)
+InferType: 6182us [6182us] (44.62%; 44.62%)
+FoldScaleAxis: 7674us [4us] (55.38%; 55.38%)
+ FoldConstant: 7670us [1592us] (55.35%; 99.94%)
+ InferType: 6078us [6078us] (43.86%; 79.24%)
</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 805c9ef5f..9d9d4e626 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: 54.157457 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Convolution: 54.159903 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 5e60e3e7e..aa6d7d029 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: 7.459653 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>conv2d with tensor core: 7.746062 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 bcee5e2b4..eb063a02a 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: 0.018450
-Baseline: 3.395711
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Numpy running time: 0.018982
+Baseline: 3.249589
</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: 0.290153
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt1: 0.318530
</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: 0.325772
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt2: 0.339569
</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: 0.121021
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt3: 0.116613
</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: 0.110929
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt4: 0.110777
</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: 0.111273
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt5: 0.112163
</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: 0.144865
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt6: 0.145942
</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 9b7f1b16a..b9deca63c 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>00:34.338</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
+<p><strong>00:34.349</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>00:31.944</p></td>
+<td><p>00:32.127</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>00:01.303</p></td>
+<td><p>00:01.214</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>00:01.092</p></td>
+<td><p>00:01.009</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 b1575dcbe..2777f213a 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>06:05.853</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
+<p><strong>06:13.103</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>03:20.355</p></td>
+<td><p>03:22.148</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>01:22.398</p></td>
+<td><p>01:23.959</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>00:46.795</p></td>
+<td><p>00:47.825</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>00:18.769</p></td>
+<td><p>00:21.450</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>00:08.886</p></td>
+<tr class="row-odd"><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>00:08.892</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>00:08.649</p></td>
+<tr class="row-even"><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>00:08.829</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 544b111d9..2ead19871 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
@@ -491,562 +491,224 @@ cooperative fetching, unrolling and operator fusion.</p>
compute: Buffer(compute_2: Pointer(float32), float32, [25088], [])}
buffer_map = {data_1: data, kernel_1: kernel, bias_1: bias, compute_1: compute}
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, [7]), storage_scope = local;
- allocate(pad_temp.shared: Pointer(shared float32), float32, [648]), storage_scope = shared;
- allocate(kernel.shared: Pointer(shared float32), float32, [2304]), storage_scope = shared;
- attr [IterVar(threadIdx.x: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224 {
- conv2d_nchw_1: Buffer(conv2d_nchw, float32, [7], [], scope="local", align=16)[0] = 0f32
+ attr [IterVar(blockIdx.x: int32, (nullptr), "ThreadIndex", "blockIdx.x")] "thread_extent" = 8;
+ allocate(conv2d_nchw: Pointer(local float32), float32, [28]), storage_scope = local;
+ allocate(pad_temp.shared: Pointer(shared float32), float32, [2016]), storage_scope = shared;
+ allocate(kernel.shared: Pointer(shared float32), float32, [6144]), storage_scope = shared;
+ attr [IterVar(threadIdx.x: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ conv2d_nchw_1: Buffer(conv2d_nchw, float32, [28], [], scope="local", align=64)[0] = 0f32
conv2d_nchw_1[1] = 0f32
conv2d_nchw_1[2] = 0f32
conv2d_nchw_1[3] = 0f32
conv2d_nchw_1[4] = 0f32
conv2d_nchw_1[5] = 0f32
conv2d_nchw_1[6] = 0f32
- for (rc.outer.outer: int32, 0, 64) {
- let cse_var_2: int32 = (rc.outer.outer*392)
- let cse_var_1: int32 = (rc.outer.outer*72)
- {
- attr [IterVar(threadIdx.x_1: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- pad_temp.shared_1: Buffer(pad_temp.shared, float32, [648], [], scope="shared")[threadIdx.x_1] = @tir.if_then_else(((((9 <= floormod(threadIdx.x_1, 81)) && (floormod(threadIdx.x_1, 81) < 72)) && (1 <= floormod(threadIdx.x_1, 9))) && (floormod(threadIdx.x_1, 9) < 8)), data[((((cse_var_2 + (floordiv(threadIdx.x_1, 81)*49)) + (floordiv(floormod(threadIdx.x_1, 81), 9)*7)) + floormod(threadIdx.x_1, 9)) - 8)], 0f32, dtype=float32)
- attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- pad_temp.shared_1[(threadIdx.x_1 + 224)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 62), 81)) && (floormod((threadIdx.x_1 + 62), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 8), 9))) && (floormod((threadIdx.x_1 + 8), 9) < 8)), data[((((cse_var_2 + (floordiv((threadIdx.x_1 + 224), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 62), 81), 9)*7)) + floormod((threadIdx.x_1 + 8), 9)) - 8)], 0f32, dtype=float32)
- attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- if @tir.likely((threadIdx.x_1 < 200), dtype=bool) {
- pad_temp.shared_1[(threadIdx.x_1 + 448)] = @tir.if_then_else(((((9 <= floormod((threadIdx.x_1 + 43), 81)) && (floormod((threadIdx.x_1 + 43), 81) < 72)) && (1 <= floormod((threadIdx.x_1 + 7), 9))) && (floormod((threadIdx.x_1 + 7), 9) < 8)), data[((((cse_var_2 + (floordiv((threadIdx.x_1 + 448), 81)*49)) + (floordiv(floormod((threadIdx.x_1 + 43), 81), 9)*7)) + floormod((threadIdx.x_1 + 7), 9)) - 8)], 0f32, dtype=float32)
+ conv2d_nchw_1[7] = 0f32
+ conv2d_nchw_1[8] = 0f32
+ conv2d_nchw_1[9] = 0f32
+ conv2d_nchw_1[10] = 0f32
+ conv2d_nchw_1[11] = 0f32
+ conv2d_nchw_1[12] = 0f32
+ conv2d_nchw_1[13] = 0f32
+ conv2d_nchw_1[14] = 0f32
+ conv2d_nchw_1[15] = 0f32
+ conv2d_nchw_1[16] = 0f32
+ conv2d_nchw_1[17] = 0f32
+ conv2d_nchw_1[18] = 0f32
+ conv2d_nchw_1[19] = 0f32
+ conv2d_nchw_1[20] = 0f32
+ conv2d_nchw_1[21] = 0f32
+ conv2d_nchw_1[22] = 0f32
+ conv2d_nchw_1[23] = 0f32
+ conv2d_nchw_1[24] = 0f32
+ conv2d_nchw_1[25] = 0f32
+ conv2d_nchw_1[26] = 0f32
+ conv2d_nchw_1[27] = 0f32
+ for (rc.outer.outer: int32, 0, 16) {
+ for (ry.outer.outer: int32, 0, 3) {
+ let cse_var_4: int32 = (rc.outer.outer*1568)
+ let cse_var_3: int32 = (ry.outer.outer*7)
+ 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" = 112;
+ pad_temp.shared_1: Buffer(pad_temp.shared, float32, [2016], [], scope="shared")[threadIdx.x_1] = @tir.if_then_else(((((1 <= (floordiv(floormod(threadIdx.x_1, 63), 9) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 9))) && (floormod(threadIdx.x_1, 9) < 8)), data[((((cse_var_4 + (floordiv(threadIdx.x_1, 9)*7)) + cse_var_3) + floormod(threadIdx.x_1, 9)) - 8)], 0f [...]
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 112)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 49), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 49), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 4), 9))) && (floormod((threadIdx.x_1 + 4), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 112), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 4), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 224)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 35), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 35), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 8), 9))) && (floormod((threadIdx.x_1 + 8), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 224), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 8), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 336)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 21), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 21), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 3), 9))) && (floormod((threadIdx.x_1 + 3), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 336), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 3), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 448)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 7), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 7), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 7), 9))) && (floormod((threadIdx.x_1 + 7), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 448), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 7), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 560)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 56), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 56), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 2), 9))) && (floormod((threadIdx.x_1 + 2), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 560), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 2), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 672)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 42), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 42), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 6), 9))) && (floormod((threadIdx.x_1 + 6), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 672), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 6), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 784)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 28), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 28), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 1), 9))) && (floormod((threadIdx.x_1 + 1), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 784), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 1), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 896)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 14), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 14), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 5), 9))) && (floormod((threadIdx.x_1 + 5), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 896), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 5), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1008)] = @tir.if_then_else(((((1 <= (floordiv(floormod(threadIdx.x_1, 63), 9) + ry.outer.outer)) && ((floordiv(floormod(threadIdx.x_1, 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod(threadIdx.x_1, 9))) && (floormod(threadIdx.x_1, 9) < 8)), data[((((cse_var_4 + (floordiv(threadIdx.x_1, 9)*7)) + cse_var_3) + floormod(threadIdx.x_1, 9)) + 776)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1120)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 49), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 49), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 4), 9))) && (floormod((threadIdx.x_1 + 4), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1120), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 4), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1232)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 35), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 35), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 8), 9))) && (floormod((threadIdx.x_1 + 8), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1232), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 8), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1344)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 21), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 21), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 3), 9))) && (floormod((threadIdx.x_1 + 3), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1344), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 3), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1456)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 7), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 7), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 7), 9))) && (floormod((threadIdx.x_1 + 7), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1456), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 7), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1568)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 56), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 56), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 2), 9))) && (floormod((threadIdx.x_1 + 2), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1568), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 2), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1680)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 42), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 42), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 6), 9))) && (floormod((threadIdx.x_1 + 6), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1680), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 6), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1792)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 28), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 28), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 1), 9))) && (floormod((threadIdx.x_1 + 1), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1792), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 1), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_1, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112;
+ pad_temp.shared_1[(threadIdx.x_1 + 1904)] = @tir.if_then_else(((((1 <= (floordiv(floormod((threadIdx.x_1 + 14), 63), 9) + ry.outer.outer)) && ((floordiv(floormod((threadIdx.x_1 + 14), 63), 9) + ry.outer.outer) < 8)) && (1 <= floormod((threadIdx.x_1 + 5), 9))) && (floormod((threadIdx.x_1 + 5), 9) < 8)), data[((((cse_var_4 + (floordiv((threadIdx.x_1 + 1904), 9)*7)) + cse_var_3) + floormod((threadIdx.x_1 + 5), 9)) - 8)], 0f32, dtype=float32)
+ attr [IterVar(threadIdx.x_2: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1: Buffer(kernel.shared, float32, [6144], [], scope="shared")[(threadIdx.x_2*4)] = kernel[((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 1)] = kernel[((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(((floormod(threadIdx.x_2, 24)*4) + 1), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 2)] = kernel[((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(((floormod(threadIdx.x_2, 24)*4) + 2), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3)] = kernel[((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 448)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 112), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 449)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 112), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 450)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 112), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 451)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 112), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 448), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 896)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 224), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 32), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 897)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 224), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 11), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 898)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 224), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 34), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 899)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 224), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 896), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 1344)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 64512)]
+ kernel.shared_1[((threadIdx.x_2*4) + 1345)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 1), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3)) + 64512)]
+ kernel.shared_1[((threadIdx.x_2*4) + 1346)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 2), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3)) + 64512)]
+ kernel.shared_1[((threadIdx.x_2*4) + 1347)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 64512)]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 1792)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 448), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 1793)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 448), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 1794)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 448), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 1795)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 448), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 1792), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 2240)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 560), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 32), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 2241)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 560), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 11), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 2242)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 560), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 34), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 2243)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 560), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 2240), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 2688)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 129024)]
+ kernel.shared_1[((threadIdx.x_2*4) + 2689)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 1), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3)) + 129024)]
+ kernel.shared_1[((threadIdx.x_2*4) + 2690)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 2), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3)) + 129024)]
+ kernel.shared_1[((threadIdx.x_2*4) + 2691)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 129024)]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 3136)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 784), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3137)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 784), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3138)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 784), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3139)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 784), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 3136), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 3584)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 896), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 32), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3585)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 896), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 11), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3586)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 896), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 34), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 3587)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 896), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 3584), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 4032)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 193536)]
+ kernel.shared_1[((threadIdx.x_2*4) + 4033)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 1), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3)) + 193536)]
+ kernel.shared_1[((threadIdx.x_2*4) + 4034)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 2), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3)) + 193536)]
+ kernel.shared_1[((threadIdx.x_2*4) + 4035)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 193536)]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 4480)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1120), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4481)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1120), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4482)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1120), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4483)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1120), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 4480), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 4928)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1232), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 32), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4929)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1232), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 11), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4930)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1232), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 34), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ kernel.shared_1[((threadIdx.x_2*4) + 4931)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1232), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 4928), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ kernel.shared_1[((threadIdx.x_2*4) + 5376)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv((floormod(threadIdx.x_2, 24)*4), 3)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 258048)]
+ kernel.shared_1[((threadIdx.x_2*4) + 5377)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 1), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3)) + 258048)]
+ kernel.shared_1[((threadIdx.x_2*4) + 5378)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 2), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3)) + 258048)]
+ kernel.shared_1[((threadIdx.x_2*4) + 5379)] = kernel[(((((((blockIdx.x*294912) + (floordiv(threadIdx.x_2, 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 1), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3)) + 258048)]
+ }
+ attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 112 {
+ if @tir.likely((threadIdx.x_2 < 80), dtype=bool) {
+ kernel.shared_1[((threadIdx.x_2*4) + 5824)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1456), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 64), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ if @tir.likely((threadIdx.x_2 < 80), dtype=bool) {
+ kernel.shared_1[((threadIdx.x_2*4) + 5825)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1456), 24)*4608)) + cse_var_2) + (floordiv(floormod(((threadIdx.x_2*4) + 65), 96), 3)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 2), 3))]
+ }
+ if @tir.likely((threadIdx.x_2 < 80), dtype=bool) {
+ kernel.shared_1[((threadIdx.x_2*4) + 5826)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1456), 24)*4608)) + cse_var_2) + (floormod((floordiv((threadIdx.x_2*4), 3) + 22), 32)*9)) + cse_var_1) + floormod(threadIdx.x_2, 3))]
+ }
+ if @tir.likely((threadIdx.x_2 < 80), dtype=bool) {
+ kernel.shared_1[((threadIdx.x_2*4) + 5827)] = kernel[((((((blockIdx.x*294912) + (floordiv((threadIdx.x_2 + 1456), 24)*4608)) + cse_var_2) + (floormod((floordiv(((threadIdx.x_2*4) + 5824), 3) + 1), 32)*9)) + cse_var_1) + floormod((threadIdx.x_2 + 1), 3))]
+ }
+ }
+ for (rc.outer.inner: int32, 0, 8) {
+ for (rx.outer.inner: int32, 0, 3) {
+ for (ff.outer.inner: int32, 0, 4) {
+ let cse_var_11: int32 = (ff.outer.inner*7)
+ let cse_var_10: int32 = (cse_var_11 + 6)
+ let cse_var_9: int32 = (cse_var_11 + 5)
+ let cse_var_8: int32 = (cse_var_11 + 4)
+ let cse_var_7: int32 = (cse_var_11 + 3)
+ let cse_var_6: int32 = (cse_var_11 + 2)
+ let cse_var_5: int32 = (cse_var_11 + 1)
+ {
+ conv2d_nchw_1[cse_var_11] = (conv2d_nchw_1[cse_var_11] + (pad_temp.shared_1[(((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_11] = (conv2d_nchw_1[cse_var_11] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 63)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_11] = (conv2d_nchw_1[cse_var_11] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 126)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_11] = (conv2d_nchw_1[cse_var_11] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 189)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_5] = (conv2d_nchw_1[cse_var_5] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 1)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_5] = (conv2d_nchw_1[cse_var_5] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 64)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_5] = (conv2d_nchw_1[cse_var_5] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 127)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_5] = (conv2d_nchw_1[cse_var_5] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 190)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_6] = (conv2d_nchw_1[cse_var_6] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 2)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_6] = (conv2d_nchw_1[cse_var_6] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 65)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_6] = (conv2d_nchw_1[cse_var_6] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 128)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_6] = (conv2d_nchw_1[cse_var_6] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 191)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_7] = (conv2d_nchw_1[cse_var_7] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 3)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_7] = (conv2d_nchw_1[cse_var_7] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 66)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_7] = (conv2d_nchw_1[cse_var_7] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 129)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_7] = (conv2d_nchw_1[cse_var_7] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 192)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_8] = (conv2d_nchw_1[cse_var_8] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 4)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_8] = (conv2d_nchw_1[cse_var_8] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 67)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_8] = (conv2d_nchw_1[cse_var_8] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 130)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_8] = (conv2d_nchw_1[cse_var_8] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 193)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_9] = (conv2d_nchw_1[cse_var_9] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 5)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_9] = (conv2d_nchw_1[cse_var_9] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 68)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_9] = (conv2d_nchw_1[cse_var_9] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 131)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_9] = (conv2d_nchw_1[cse_var_9] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 194)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ conv2d_nchw_1[cse_var_10] = (conv2d_nchw_1[cse_var_10] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 6)]*kernel.shared_1[((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner)]))
+ conv2d_nchw_1[cse_var_10] = (conv2d_nchw_1[cse_var_10] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 69)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 3)]))
+ conv2d_nchw_1[cse_var_10] = (conv2d_nchw_1[cse_var_10] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 132)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 6)]))
+ conv2d_nchw_1[cse_var_10] = (conv2d_nchw_1[cse_var_10] + (pad_temp.shared_1[((((rc.outer.inner*252) + (floormod(threadIdx.x, 7)*9)) + rx.outer.inner) + 195)]*kernel.shared_1[(((((floordiv(threadIdx.x, 7)*384) + (ff.outer.inner*96)) + (rc.outer.inner*12)) + rx.outer.inner) + 9)]))
+ }
+ }
+ }
+ }
}
- attr [IterVar(threadIdx.x_2: int32, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1: Buffer(kernel.shared, float32, [2304], [], scope="shared")[threadIdx.x_2] = kernel[((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 72)*4608)) + cse_var_1) + floormod(threadIdx.x_2, 72))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 224)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 224), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 8), 72), 3)*3)) + floormod((threadIdx.x_2 + 2), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 448)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 448), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 16), 72), 3)*3)) + floormod((threadIdx.x_2 + 1), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 672)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 672), 72)*4608)) + cse_var_1) + (floormod((floordiv(threadIdx.x_2, 3) + 8), 24)*3)) + floormod(threadIdx.x_2, 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 896)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 896), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 32), 72), 3)*3)) + floormod((threadIdx.x_2 + 2), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 1120)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 1120), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 40), 72), 3)*3)) + floormod((threadIdx.x_2 + 1), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 1344)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 1344), 72)*4608)) + cse_var_1) + (floormod((floordiv(threadIdx.x_2, 3) + 16), 24)*3)) + floormod(threadIdx.x_2, 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 1568)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 1568), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 56), 72), 3)*3)) + floormod((threadIdx.x_2 + 2), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 1792)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 1792), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 64), 72), 3)*3)) + floormod((threadIdx.x_2 + 1), 3))]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- kernel.shared_1[(threadIdx.x_2 + 2016)] = kernel[(((((blockIdx.x*147456) + (floordiv(threadIdx.x_2, 72)*4608)) + cse_var_1) + floormod(threadIdx.x_2, 72)) + 129024)]
- attr [IterVar(threadIdx.x_2, (nullptr), "ThreadIndex", "threadIdx.x")] "thread_extent" = 224;
- if @tir.likely((threadIdx.x_2 < 64), dtype=bool) {
- kernel.shared_1[(threadIdx.x_2 + 2240)] = kernel[(((((blockIdx.x*147456) + (floordiv((threadIdx.x_2 + 2240), 72)*4608)) + cse_var_1) + (floordiv(floormod((threadIdx.x_2 + 8), 72), 3)*3)) + floormod((threadIdx.x_2 + 2), 3))]
- }
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[(floormod(threadIdx.x, 7)*9)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 1)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 2)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 3)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 4)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 5)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 6)]*kernel.shared_1[(floordiv(threadIdx.x, 7)*72)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 1)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 2)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 3)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 4)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 5)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 6)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 7)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 1)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 2)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 3)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 4)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 5)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 6)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 7)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 8)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 2)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 81)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 82)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 83)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 84)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 85)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 86)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 87)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 9)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 82)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 83)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 84)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 85)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 86)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 87)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 88)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 10)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 83)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 84)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 85)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 86)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 87)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 88)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 89)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 11)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 9)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 10)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 11)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 12)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 13)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 14)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 15)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 3)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 10)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 11)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 12)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 13)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 14)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 15)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 16)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 4)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 11)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 12)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 13)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 14)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 15)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 16)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 17)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 5)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 90)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 91)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 92)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 93)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 94)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 95)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 96)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 12)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 91)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 92)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 93)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 94)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 95)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 96)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 97)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 13)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 92)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 93)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 94)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 95)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 96)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 97)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 98)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 14)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 18)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 19)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 20)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 21)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 22)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 23)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 24)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 6)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 19)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 20)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 21)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 22)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 23)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 24)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 25)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 7)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 20)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 21)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 22)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 23)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 24)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 25)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 26)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 8)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 99)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 100)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 101)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 102)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 103)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 104)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 105)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 15)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 100)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 101)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 102)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 103)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 104)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 105)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 106)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 16)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 101)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 102)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 103)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 104)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 105)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 106)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 107)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 17)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 162)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 163)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 164)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 165)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 166)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 167)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 168)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 18)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 163)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 164)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 165)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 166)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 167)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 168)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 169)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 19)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 164)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 165)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 166)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 167)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 168)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 169)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 170)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 20)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 243)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 244)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 246)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 247)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 248)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 249)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 27)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 244)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 246)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 247)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 248)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 249)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 250)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 28)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 245)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 246)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 247)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 248)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 249)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 250)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 251)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 29)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 171)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 172)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 173)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 174)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 175)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 176)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 177)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 21)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 172)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 173)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 174)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 175)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 176)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 177)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 178)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 22)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 173)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 174)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 175)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 176)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 177)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 178)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 179)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 23)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 252)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 253)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 254)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 255)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 256)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 257)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 258)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 30)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 253)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 254)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 255)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 256)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 257)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 258)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 259)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 31)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 254)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 255)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 256)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 257)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 258)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 259)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 260)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 32)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 180)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 181)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 182)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 183)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 184)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 185)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 186)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 24)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 181)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 182)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 183)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 184)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 185)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 186)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 187)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 25)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 182)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 183)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 184)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 185)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 186)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 187)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 188)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 26)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 261)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 262)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 263)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 264)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 265)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 266)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 267)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 33)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 262)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 263)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 264)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 265)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 266)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 267)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 268)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 34)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 263)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 264)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 265)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 266)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 267)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 268)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 269)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 35)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 324)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 325)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 326)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 327)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 328)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 329)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 330)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 36)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 325)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 326)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 327)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 328)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 329)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 330)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 331)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 37)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 326)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 327)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 328)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 329)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 330)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 331)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 332)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 38)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 405)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 406)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 407)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 408)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 409)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 410)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 411)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 45)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 406)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 407)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 408)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 409)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 410)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 411)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 412)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 46)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 407)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 408)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 409)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 410)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 411)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 412)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 413)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 47)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 333)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 334)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 335)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 336)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 337)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 338)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 339)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 39)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 334)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 335)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 336)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 337)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 338)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 339)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 340)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 40)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 335)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 336)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 337)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 338)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 339)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 340)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 341)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 41)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 414)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 415)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 416)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 417)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 418)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 419)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 420)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 48)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 415)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 416)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 417)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 418)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 419)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 420)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 421)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 49)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 416)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 417)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 418)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 419)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 420)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 421)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 422)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 50)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 342)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 344)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 345)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 346)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 347)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 348)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 42)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 343)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 344)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 345)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 346)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 347)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 348)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 349)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 43)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 344)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 345)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 346)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 347)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 348)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 349)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 350)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 44)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 423)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 424)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 425)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 426)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 427)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 428)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 429)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 51)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 424)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 425)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 426)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 427)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 428)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 429)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 430)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 52)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 425)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 426)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 427)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 428)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 429)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 430)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 431)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 53)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 486)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 487)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 488)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 489)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 491)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 492)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 54)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 487)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 488)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 489)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 491)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 492)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 493)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 55)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 488)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 489)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 490)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 491)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 492)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 493)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 494)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 56)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 567)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 568)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 569)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 570)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 571)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 572)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 573)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 63)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 568)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 569)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 570)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 571)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 572)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 573)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 574)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 64)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 569)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 570)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 571)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 572)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 573)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 574)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 575)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 65)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 495)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 496)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 497)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 498)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 499)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 500)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 501)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 57)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 496)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 497)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 498)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 499)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 500)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 501)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 502)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 58)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 497)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 498)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 499)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 500)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 501)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 502)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 503)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 59)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 576)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 577)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 578)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 579)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 580)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 581)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 582)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 66)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 577)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 578)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 579)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 580)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 581)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 582)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 583)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 67)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 578)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 579)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 580)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 581)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 582)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 583)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 584)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 68)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 504)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 505)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 506)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 507)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 508)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 509)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 510)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 60)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 505)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 506)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 507)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 508)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 509)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 510)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 511)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 61)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 506)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 507)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 508)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 509)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 510)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 511)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 512)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 62)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 585)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 586)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 587)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 589)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 590)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 591)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 69)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 586)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 587)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 589)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 590)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 591)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 592)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 70)]))
- conv2d_nchw_1[0] = (conv2d_nchw_1[0] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 587)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[1] = (conv2d_nchw_1[1] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 588)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[2] = (conv2d_nchw_1[2] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 589)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[3] = (conv2d_nchw_1[3] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 590)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[4] = (conv2d_nchw_1[4] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 591)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[5] = (conv2d_nchw_1[5] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 592)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
- conv2d_nchw_1[6] = (conv2d_nchw_1[6] + (pad_temp.shared_1[((floormod(threadIdx.x, 7)*9) + 593)]*kernel.shared_1[((floordiv(threadIdx.x, 7)*72) + 71)]))
}
}
- for (i3.inner: int32, 0, 7) {
- compute[(((blockIdx.x*1568) + (threadIdx.x*7)) + i3.inner)] = max((conv2d_nchw_1[i3.inner] + bias[((blockIdx.x*32) + floordiv(threadIdx.x, 7))]), 0f32)
+ for (i1.inner: int32, 0, 4) {
+ for (i3.inner: int32, 0, 7) {
+ compute[(((((blockIdx.x*3136) + (floordiv(threadIdx.x, 7)*196)) + (i1.inner*49)) + (floormod(threadIdx.x, 7)*7)) + i3.inner)] = max((conv2d_nchw_1[((i1.inner*7) + i3.inner)] + bias[(((blockIdx.x*64) + (floordiv(threadIdx.x, 7)*4)) + i1.inner)]), 0f32)
+ }
}
}
}
@@ -1083,7 +745,7 @@ 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: 0.248 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 0.420 ms
</pre></div>
</div>
</div>
@@ -1113,29 +775,29 @@ conv2d_nchw_nn_o_o_i, conv2d_nchw_nn_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o
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=1)
-conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=1)
-conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=32)
+conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=4)
+conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=16)
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=1)
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)
conv2d_nchw_yy_o_o_o_o, conv2d_nchw_yy_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_o_o_i, factor=1)
-conv2d_nchw_xx_o_i, conv2d_nchw_xx_i = s[conv2d_nchw].split(conv2d_nchw_xx, factor=7)
-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_i, conv2d_nchw_xx_i = s[conv2d_nchw].split(conv2d_nchw_xx, factor=1)
+conv2d_nchw_xx_o_o_i, conv2d_nchw_xx_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_i, factor=7)
conv2d_nchw_xx_o_o_o_i, conv2d_nchw_xx_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_i, factor=1)
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=2)
-conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=4)
+conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=4)
+conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=8)
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=3)
-conv2d_nchw_rx_o_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=3)
-conv2d_nchw_rx_o_o, conv2d_nchw_rx_o_i = s[conv2d_nchw].split(conv2d_nchw_rx_o_i, 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=1)
+conv2d_nchw_rx_o_o, conv2d_nchw_rx_o_i = s[conv2d_nchw].split(conv2d_nchw_rx_o_i, factor=3)
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=1)
-compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=32)
+compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=4)
+compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=16)
compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=1)
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)
@@ -1159,16 +821,16 @@ s[compute].bind(compute_i0_o_o_i_i1_o_o_i_fused_i2_o_o_i_fused_i3_o_o_i_fused, t
compute_i0_o_i_i1_o_i_fused_i2_o_i_fused_i3_o_i_fused = s[compute].fuse(compute_i0_o_i, compute_i1_o_i, compute_i2_o_i, compute_i3_o_i)
s[compute].bind(compute_i0_o_i_i1_o_i_fused_i2_o_i_fused_i3_o_i_fused, te.thread_axis("threadIdx.x"))
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[kernel_shared].fuse(kernel_shared_ax0, kernel_shared_ax1, kernel_shared_ax2, kernel_shared_ax3)
-kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
+kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=4)
s[kernel_shared].vectorize(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
-kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=224)
+kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=112)
s[kernel_shared].bind(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[pad_temp_shared].fuse(pad_temp_shared_ax0, pad_temp_shared_ax1, pad_temp_shared_ax2, pad_temp_shared_ax3)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
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=224)
+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=112)
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", 512)
+s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 64)
s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
@@ -1186,10 +848,10 @@ CUDA source code:
#define int64_t long long
#define uint64_t unsigned long long
#endif
-extern "C" __global__ void __launch_bounds__(224) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
- float conv2d_nchw[7];
- __shared__ float pad_temp_shared[648];
- __shared__ float kernel_shared[2304];
+extern "C" __global__ void __launch_bounds__(112) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
+ float conv2d_nchw[28];
+ __shared__ float pad_temp_shared[2016];
+ __shared__ float kernel_shared[6144];
conv2d_nchw[0] = 0.000000e+00f;
conv2d_nchw[1] = 0.000000e+00f;
conv2d_nchw[2] = 0.000000e+00f;
@@ -1197,534 +859,153 @@ extern "C" __global__ void __launch_bounds__(224) default_function_ker
conv2d_nchw[4] = 0.000000e+00f;
conv2d_nchw[5] = 0.000000e+00f;
conv2d_nchw[6] = 0.000000e+00f;
- for (int rc_outer_outer = 0; rc_outer_outer < 64; ++rc_outer_outer) {
- __syncthreads();
- pad_temp_shared[((int)threadIdx.x)] = (((((9 <= (((int)threadIdx.x) % 81)) && ((((int)threadIdx.x) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + ((((int)threadIdx.x) / 81) * 49)) + (((((int)threadIdx.x) % 81) / 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((9 <= ((((int)threadIdx.x) + 62) % 81)) && (((((int)threadIdx.x) + 62) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 224) / 81) * 49)) + ((((((int)threadIdx.x) + 62) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- if (((int)threadIdx.x) < 200) {
- pad_temp_shared[(((int)threadIdx.x) + 448)] = (((((9 <= ((((int)threadIdx.x) + 43) % 81)) && (((((int)threadIdx.x) + 43) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 448) / 81) * 49)) + ((((((int)threadIdx.x) + 43) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- }
- kernel_shared[((int)threadIdx.x)] = kernel[((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72))];
- kernel_shared[(((int)threadIdx.x) + 224)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 224) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 448)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 448) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 672)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 672) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) / 3) + 8) % 24) * 3)) + (((int)threadIdx.x) % 3))];
- kernel_shared[(((int)threadIdx.x) + 896)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 896) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1120)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 1120) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 1344) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) / 3) + 16) % 24) * 3)) + (((int)threadIdx.x) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1568)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 1568) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1792)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 1792) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2016)] = kernel[(((((((int)blockIdx.x) * 147456) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 129024)];
- if (((int)threadIdx.x) < 64) {
- kernel_shared[(((int)threadIdx.x) + 2240)] = kernel[(((((((int)blockIdx.x) * 147456) + (((((int)threadIdx.x) + 2240) / 72) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) + 8) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ conv2d_nchw[7] = 0.000000e+00f;
+ conv2d_nchw[8] = 0.000000e+00f;
+ conv2d_nchw[9] = 0.000000e+00f;
+ conv2d_nchw[10] = 0.000000e+00f;
+ conv2d_nchw[11] = 0.000000e+00f;
+ conv2d_nchw[12] = 0.000000e+00f;
+ conv2d_nchw[13] = 0.000000e+00f;
+ conv2d_nchw[14] = 0.000000e+00f;
+ conv2d_nchw[15] = 0.000000e+00f;
+ conv2d_nchw[16] = 0.000000e+00f;
+ conv2d_nchw[17] = 0.000000e+00f;
+ conv2d_nchw[18] = 0.000000e+00f;
+ conv2d_nchw[19] = 0.000000e+00f;
+ conv2d_nchw[20] = 0.000000e+00f;
+ conv2d_nchw[21] = 0.000000e+00f;
+ conv2d_nchw[22] = 0.000000e+00f;
+ conv2d_nchw[23] = 0.000000e+00f;
+ conv2d_nchw[24] = 0.000000e+00f;
+ conv2d_nchw[25] = 0.000000e+00f;
+ conv2d_nchw[26] = 0.000000e+00f;
+ conv2d_nchw[27] = 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();
+ pad_temp_shared[((int)threadIdx.x)] = (((((1 <= (((((int)threadIdx.x) % 63) / 9) + ry_outer_outer)) && ((((((int)threadIdx.x) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + ((((int)threadIdx.x) / 9) * 7)) + (ry_outer_outer * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 112)] = (((((1 <= ((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 112) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((1 <= ((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 224) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 336)] = (((((1 <= ((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 336) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 448)] = (((((1 <= ((((((int)threadIdx.x) + 7) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 7) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 448) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 560)] = (((((1 <= ((((((int)threadIdx.x) + 56) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 56) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 560) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 672)] = (((((1 <= ((((((int)threadIdx.x) + 42) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 42) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 672) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = (((((1 <= ((((((int)threadIdx.x) + 28) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 28) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 784) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 896)] = (((((1 <= ((((((int)threadIdx.x) + 14) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 14) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 896) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1008)] = (((((1 <= (((((int)threadIdx.x) % 63) / 9) + ry_outer_outer)) && ((((((int)threadIdx.x) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + ((((int)threadIdx.x) / 9) * 7)) + (ry_outer_outer * 7)) + (((int)threadIdx.x) % 9)) + 776)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1120)] = (((((1 <= ((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1120) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1232)] = (((((1 <= ((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1232) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1344)] = (((((1 <= ((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1344) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1456)] = (((((1 <= ((((((int)threadIdx.x) + 7) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 7) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1456) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1568)] = (((((1 <= ((((((int)threadIdx.x) + 56) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 56) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1568) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1680)] = (((((1 <= ((((((int)threadIdx.x) + 42) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 42) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1680) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1792)] = (((((1 <= ((((((int)threadIdx.x) + 28) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 28) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1792) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1904)] = (((((1 <= ((((((int)threadIdx.x) + 14) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 14) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1904) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
+ kernel_shared[(((int)threadIdx.x) * 4)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 448)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 449)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 450)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 451)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 448) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 896)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 32) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 897)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 11) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 898)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 34) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 899)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 896) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1344)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 64512)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1345)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3)) + 64512)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1346)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3)) + 64512)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1347)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 64512)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1792)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1793)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1794)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 1795)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 1792) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2240)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 32) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2241)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 11) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2242)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 34) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2243)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 2240) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2688)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 129024)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2689)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3)) + 129024)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2690)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3)) + 129024)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 2691)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 129024)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3136)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3137)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3138)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3139)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 3136) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3584)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 32) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3585)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 11) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3586)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 34) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 3587)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 3584) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4032)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 193536)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4033)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3)) + 193536)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4034)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3)) + 193536)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4035)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 193536)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4480)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4481)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4482)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4483)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 4480) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4928)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 32) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4929)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 11) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4930)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 34) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 4931)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 4928) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ kernel_shared[((((int)threadIdx.x) * 4) + 5376)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((int)threadIdx.x) % 24) * 4) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 258048)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 5377)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 1) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3)) + 258048)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 5378)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) % 24) * 4) + 2) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3)) + 258048)];
+ kernel_shared[((((int)threadIdx.x) * 4) + 5379)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 258048)];
+ if (((int)threadIdx.x) < 80) {
+ kernel_shared[((((int)threadIdx.x) * 4) + 5824)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 64) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ }
+ if (((int)threadIdx.x) < 80) {
+ kernel_shared[((((int)threadIdx.x) * 4) + 5825)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) + 65) % 96) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+ }
+ if (((int)threadIdx.x) < 80) {
+ kernel_shared[((((int)threadIdx.x) * 4) + 5826)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 288)) + (((((((int)threadIdx.x) * 4) / 3) + 22) & 31) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+ }
+ if (((int)threadIdx.x) < 80) {
+ kernel_shared[((((int)threadIdx.x) * 4) + 5827)] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 288)) + ((((((((int)threadIdx.x) * 4) + 5824) / 3) + 1) & 31) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ }
+ __syncthreads();
+ for (int rc_outer_inner = 0; rc_outer_inner < 8; ++rc_outer_inner) {
+ for (int rx_outer_inner = 0; rx_outer_inner < 3; ++rx_outer_inner) {
+ for (int ff_outer_inner = 0; ff_outer_inner < 4; ++ff_outer_inner) {
+ conv2d_nchw[(ff_outer_inner * 7)] = (conv2d_nchw[(ff_outer_inner * 7)] + (pad_temp_shared[(((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[(ff_outer_inner * 7)] = (conv2d_nchw[(ff_outer_inner * 7)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 63)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[(ff_outer_inner * 7)] = (conv2d_nchw[(ff_outer_inner * 7)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 126)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[(ff_outer_inner * 7)] = (conv2d_nchw[(ff_outer_inner * 7)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 189)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 1)] = (conv2d_nchw[((ff_outer_inner * 7) + 1)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 1)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 1)] = (conv2d_nchw[((ff_outer_inner * 7) + 1)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 64)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 1)] = (conv2d_nchw[((ff_outer_inner * 7) + 1)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 127)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 1)] = (conv2d_nchw[((ff_outer_inner * 7) + 1)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 190)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 2)] = (conv2d_nchw[((ff_outer_inner * 7) + 2)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 2)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 2)] = (conv2d_nchw[((ff_outer_inner * 7) + 2)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 65)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 2)] = (conv2d_nchw[((ff_outer_inner * 7) + 2)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 128)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 2)] = (conv2d_nchw[((ff_outer_inner * 7) + 2)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 191)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 3)] = (conv2d_nchw[((ff_outer_inner * 7) + 3)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 3)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 3)] = (conv2d_nchw[((ff_outer_inner * 7) + 3)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 66)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 3)] = (conv2d_nchw[((ff_outer_inner * 7) + 3)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 129)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 3)] = (conv2d_nchw[((ff_outer_inner * 7) + 3)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 192)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 4)] = (conv2d_nchw[((ff_outer_inner * 7) + 4)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 4)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 4)] = (conv2d_nchw[((ff_outer_inner * 7) + 4)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 67)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 4)] = (conv2d_nchw[((ff_outer_inner * 7) + 4)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 130)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 4)] = (conv2d_nchw[((ff_outer_inner * 7) + 4)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 193)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 5)] = (conv2d_nchw[((ff_outer_inner * 7) + 5)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 5)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 5)] = (conv2d_nchw[((ff_outer_inner * 7) + 5)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 68)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 5)] = (conv2d_nchw[((ff_outer_inner * 7) + 5)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 131)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 5)] = (conv2d_nchw[((ff_outer_inner * 7) + 5)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 194)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 6)] = (conv2d_nchw[((ff_outer_inner * 7) + 6)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 6)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 6)] = (conv2d_nchw[((ff_outer_inner * 7) + 6)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 69)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 3)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 6)] = (conv2d_nchw[((ff_outer_inner * 7) + 6)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 132)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 6)]));
+ conv2d_nchw[((ff_outer_inner * 7) + 6)] = (conv2d_nchw[((ff_outer_inner * 7) + 6)] + (pad_temp_shared[((((rc_outer_inner * 252) + ((((int)threadIdx.x) % 7) * 9)) + rx_outer_inner) + 195)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 384) + (ff_outer_inner * 96)) + (rc_outer_inner * 12)) + rx_outer_inner) + 9)]));
+ }
+ }
+ }
}
- __syncthreads();
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[((((int)threadIdx.x) / 7) * 72)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 1)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 2)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 81)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 82)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 9)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 82)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 88)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 10)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 88)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 89)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 11)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 9)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 10)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 11)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 12)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 13)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 14)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 15)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 3)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 10)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 11)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 12)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 13)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 14)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 15)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 16)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 4)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 11)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 12)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 13)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 14)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 15)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 16)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 17)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 5)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 90)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 91)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 92)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 93)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 94)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 95)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 96)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 12)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 91)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 92)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 93)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 94)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 95)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 96)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 97)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 13)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 92)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 93)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 94)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 95)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 96)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 97)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 98)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 14)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 18)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 19)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 20)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 21)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 22)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 23)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 24)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 6)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 19)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 20)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 21)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 22)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 23)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 24)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 25)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 7)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 20)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 21)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 22)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 23)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 24)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 25)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 26)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 8)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 99)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 100)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 101)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 102)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 103)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 104)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 105)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 15)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 100)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 101)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 102)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 103)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 104)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 105)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 106)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 16)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 101)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 102)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 103)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 104)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 105)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 106)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 107)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 17)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 162)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 163)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 18)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 163)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 169)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 19)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 169)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 170)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 20)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 243)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 244)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 27)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 244)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 250)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 28)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 250)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 251)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 29)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 171)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 172)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 173)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 174)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 175)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 176)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 177)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 21)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 172)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 173)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 174)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 175)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 176)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 177)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 178)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 22)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 173)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 174)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 175)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 176)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 177)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 178)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 179)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 23)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 30)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 31)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 32)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 180)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 181)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 182)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 183)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 184)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 185)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 186)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 24)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 181)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 182)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 183)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 184)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 185)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 186)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 187)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 25)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 182)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 183)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 184)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 185)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 186)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 187)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 188)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 26)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 261)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 262)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 263)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 264)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 265)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 266)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 267)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 33)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 262)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 263)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 264)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 265)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 266)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 267)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 268)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 34)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 263)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 264)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 265)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 266)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 267)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 268)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 269)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 35)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 324)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 325)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 36)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 325)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 331)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 37)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 331)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 332)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 38)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 405)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 406)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 45)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 406)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 412)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 46)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 412)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 413)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 47)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 333)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 334)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 335)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 336)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 337)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 338)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 339)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 39)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 334)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 335)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 336)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 337)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 338)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 339)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 340)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 40)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 335)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 336)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 337)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 338)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 339)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 340)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 341)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 41)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 414)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 415)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 416)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 417)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 418)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 419)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 420)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 48)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 415)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 416)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 417)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 418)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 419)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 420)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 421)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 49)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 416)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 417)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 418)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 419)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 420)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 421)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 422)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 50)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 342)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 343)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 344)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 345)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 346)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 347)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 348)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 42)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 343)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 344)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 345)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 346)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 347)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 348)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 349)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 43)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 344)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 345)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 346)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 347)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 348)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 349)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 350)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 44)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 423)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 424)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 425)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 426)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 427)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 428)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 429)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 51)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 424)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 425)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 426)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 427)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 428)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 429)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 430)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 52)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 425)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 426)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 427)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 428)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 429)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 430)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 431)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 53)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 486)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 487)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 488)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 489)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 490)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 491)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 492)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 54)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 487)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 488)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 489)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 490)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 491)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 492)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 493)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 55)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 488)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 489)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 490)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 491)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 492)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 493)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 494)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 56)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 567)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 568)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 569)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 570)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 571)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 572)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 573)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 63)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 568)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 569)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 570)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 571)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 572)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 573)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 574)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 64)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 569)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 570)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 571)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 572)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 573)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 574)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 575)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 65)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 495)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 496)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 497)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 498)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 499)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 500)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 501)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 57)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 496)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 497)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 498)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 499)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 500)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 501)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 502)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 58)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 497)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 498)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 499)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 500)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 501)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 502)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 503)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 59)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 576)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 577)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 578)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 579)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 580)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 581)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 582)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 66)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 577)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 578)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 579)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 580)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 581)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 582)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 583)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 67)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 578)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 579)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 580)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 581)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 582)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 583)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 584)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 68)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 504)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 505)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 506)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 507)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 508)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 509)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 510)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 60)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 505)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 506)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 507)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 508)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 509)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 510)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 511)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 61)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 506)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 507)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 508)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 509)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 510)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 511)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 512)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 62)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 585)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 586)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 587)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 588)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 589)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 590)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 591)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 69)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 586)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 587)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 588)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 589)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 590)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 591)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 592)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 70)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 587)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 588)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 589)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 590)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 591)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 592)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 593)] * kernel_shared[(((((int)threadIdx.x) / 7) * 72) + 71)]));
}
- for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
- compute[(((((int)blockIdx.x) * 1568) + (((int)threadIdx.x) * 7)) + i3_inner)] = max((conv2d_nchw[i3_inner] + bias[((((int)blockIdx.x) * 32) + (((int)threadIdx.x) / 7))]), 0.000000e+00f);
+ for (int i1_inner = 0; i1_inner < 4; ++i1_inner) {
+ for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
+ compute[(((((((int)blockIdx.x) * 3136) + ((((int)threadIdx.x) / 7) * 196)) + (i1_inner * 49)) + ((((int)threadIdx.x) % 7) * 7)) + i3_inner)] = max((conv2d_nchw[((i1_inner * 7) + i3_inner)] + bias[(((((int)blockIdx.x) * 64) + ((((int)threadIdx.x) / 7) * 4)) + i1_inner)]), 0.000000e+00f);
+ }
}
}
</pre></div>
@@ -1761,7 +1042,7 @@ 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 20.355 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes 22.148 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 be8c6f5d8..5ba88c6f2 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)
- 9.7893 9.8049 9.8140 9.7491 0.0287
+ 9.9589 9.9571 9.9964 9.9231 0.0299
</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 634d62820..241938815 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)
- 754.1263 754.0259 754.4682 753.8849 0.2485
+ 761.6540 760.7114 764.0088 760.2418 1.6761
</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 22.398 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 23.959 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 81d97d994..7e9881458 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
@@ -625,15 +625,15 @@ 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 = {compute_1: compute_3: Buffer(compute_2, float32, [128, 512], []), placeholder_6: placeholder_15: Buffer(placeholder_11, float32, [4916, 16, 1], []), placeholder_7: placeholder_16: Buffer(placeholder_12, int32, [4916], []), placeholder_5: placeholder_17: Buffer(placeholder_10, float32, [128, 256], []), placeholder_9: placeholder_18: Buffer(placeholder_14, float32, [128, 512], []), placeholder_8: placeholder_19: Buffer(placeholder_13, int32, [33], [])} {
- for (i0.outer.i1.outer.fused: int32, 0, 16) "parallel" {
- allocate(compute_4: Pointer(global float32), float32, [4096]), storage_scope = global {
- for (i.outer.inner: int32, 0, 16) {
+ preflattened_buffer_map = {placeholder_7: placeholder_15: Buffer(placeholder_12, int32, [4916], []), placeholder_8: placeholder_16: Buffer(placeholder_13, int32, [33], []), placeholder_6: placeholder_17: Buffer(placeholder_11, float32, [4916, 16, 1], []), placeholder_5: placeholder_18: Buffer(placeholder_10, float32, [128, 256], []), compute_1: compute_3: Buffer(compute_2, float32, [128, 512], []), placeholder_9: placeholder_19: Buffer(placeholder_14, 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 (i.outer.inner: int32, 0, 2) {
for (nb_j.inner: int32, 0, 2) {
for (i.inner.init: int32, 0, 8) {
let cse_var_1: int32 = (((i.outer.inner*256) + (i.inner.init*32)) + (nb_j.inner*16))
{
- compute_5: Buffer(compute_4, float32, [4096], [])[cse_var_1] = 0f32
+ 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
@@ -651,51 +651,51 @@ layout transformation, parallelization, vectorization, unrolling, and operator f
compute_5[(cse_var_1 + 15)] = 0f32
}
}
- for (elem_idx: int32, 0, let cse_var_2: int32 = ((i0.outer.i1.outer.fused*2) + nb_j.inner) in (placeholder_3[(cse_var_2 + 1)] - placeholder_3[cse_var_2])) {
+ 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, 8) {
let cse_var_21: int32 = (elem_idx*16)
- let cse_var_20: int32 = ((i0.outer.i1.outer.fused*2) + nb_j.inner)
- let cse_var_19: int32 = ((i.outer.inner*2048) + (i.inner*256))
- let cse_var_18: int32 = (((i.outer.inner*256) + (i.inner*32)) + (nb_j.inner*16))
- let cse_var_17: int32 = (cse_var_18 + 9)
- let cse_var_16: int32 = (cse_var_18 + 8)
- let cse_var_15: int32 = (cse_var_18 + 7)
- let cse_var_14: int32 = (cse_var_18 + 6)
- let cse_var_13: int32 = (cse_var_18 + 5)
- let cse_var_12: int32 = (cse_var_18 + 4)
- let cse_var_11: int32 = (cse_var_18 + 3)
- let cse_var_10: int32 = (cse_var_18 + 2)
- let cse_var_9: int32 = (cse_var_18 + 15)
- let cse_var_8: int32 = (cse_var_18 + 14)
- let cse_var_7: int32 = (cse_var_18 + 13)
- let cse_var_6: int32 = (cse_var_18 + 12)
- let cse_var_5: int32 = (cse_var_18 + 11)
- let cse_var_4: int32 = (cse_var_18 + 10)
- let cse_var_3: int32 = (cse_var_18 + 1)
+ let cse_var_20: int32 = ((floormod(i0.outer.i1.outer.fused, 16)*2) + nb_j.inner)
+ let cse_var_19: int32 = (((i.outer.inner*256) + (i.inner*32)) + (nb_j.inner*16))
+ let cse_var_18: int32 = (((floordiv(i0.outer.i1.outer.fused, 16)*4096) + (i.outer.inner*2048)) + (i.inner*256))
+ let cse_var_17: int32 = (cse_var_19 + 9)
+ let cse_var_16: int32 = (cse_var_19 + 8)
+ let cse_var_15: int32 = (cse_var_19 + 7)
+ let cse_var_14: int32 = (cse_var_19 + 6)
+ let cse_var_13: int32 = (cse_var_19 + 5)
+ let cse_var_12: int32 = (cse_var_19 + 4)
+ let cse_var_11: int32 = (cse_var_19 + 3)
+ let cse_var_10: int32 = (cse_var_19 + 2)
+ let cse_var_9: int32 = (cse_var_19 + 15)
+ let cse_var_8: int32 = (cse_var_19 + 14)
+ let cse_var_7: int32 = (cse_var_19 + 13)
+ let cse_var_6: int32 = (cse_var_19 + 12)
+ let cse_var_5: int32 = (cse_var_19 + 11)
+ let cse_var_4: int32 = (cse_var_19 + 10)
+ let cse_var_3: int32 = (cse_var_19 + 1)
{
- compute_5[cse_var_18] = (compute_5[cse_var_18] + (placeholder_1[((placeholder_3[cse_var_20]*16) + cse_var_21)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_3] = (compute_5[cse_var_3] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 1)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_10] = (compute_5[cse_var_10] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 2)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_11] = (compute_5[cse_var_11] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 3)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_12] = (compute_5[cse_var_12] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 4)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_13] = (compute_5[cse_var_13] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 5)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_14] = (compute_5[cse_var_14] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 6)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_15] = (compute_5[cse_var_15] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 7)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_16] = (compute_5[cse_var_16] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 8)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_17] = (compute_5[cse_var_17] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 9)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_4] = (compute_5[cse_var_4] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 10)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_5] = (compute_5[cse_var_5] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 11)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_6] = (compute_5[cse_var_6] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 12)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_7] = (compute_5[cse_var_7] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 13)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_8] = (compute_5[cse_var_8] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 14)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
- compute_5[cse_var_9] = (compute_5[cse_var_9] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 15)]*max(placeholder[(cse_var_19 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_19] = (compute_5[cse_var_19] + (placeholder_1[((placeholder_3[cse_var_20]*16) + cse_var_21)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_3] = (compute_5[cse_var_3] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 1)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_10] = (compute_5[cse_var_10] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 2)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_11] = (compute_5[cse_var_11] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 3)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_12] = (compute_5[cse_var_12] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 4)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_13] = (compute_5[cse_var_13] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 5)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_14] = (compute_5[cse_var_14] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 6)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_15] = (compute_5[cse_var_15] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 7)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_16] = (compute_5[cse_var_16] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 8)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_17] = (compute_5[cse_var_17] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 9)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_4] = (compute_5[cse_var_4] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 10)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_5] = (compute_5[cse_var_5] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 11)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_6] = (compute_5[cse_var_6] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 12)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_7] = (compute_5[cse_var_7] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 13)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_8] = (compute_5[cse_var_8] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 14)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
+ compute_5[cse_var_9] = (compute_5[cse_var_9] + (placeholder_1[(((placeholder_3[cse_var_20]*16) + cse_var_21) + 15)]*max(placeholder[(cse_var_18 + placeholder_2[(placeholder_3[cse_var_20] + elem_idx)])], 0f32)))
}
}
}
}
}
- for (i0.inner: int32, 0, 128) {
- let cse_var_22: int32 = ((i0.inner*512) + (i0.outer.i1.outer.fused*32))
+ 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))
}
}
@@ -734,7 +734,7 @@ 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: 1.846 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 1.852 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 e1990625e..52182833c 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>00:46.493</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
+<p><strong>00:45.864</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -336,11 +336,11 @@
</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>00:46.463</p></td>
+<td><p>00:45.828</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>00:00.016</p></td>
+<td><p>00:00.021</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>
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 bb981242f..20bf73b2b 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: 218.25/218.25 result: MeasureResult(costs=(0.0010607314827586207,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.160534620285034, timestamp=1660697029.041432) [('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: 0.00/218.25 result: Traceback (most recent call last):
+No: 9 GFLOPS: 176.20/176.20 result: MeasureResult(costs=(0.0013138815888888889,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.0211830139160156, timestamp=1660714966.3254552) [('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: 0.00/176.20 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: 260.45/260.45 result: MeasureResult(costs=(0.0008888644861878454,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7346055507659912, timestamp=1660697029.9757693) [('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: 0.00/260.45 result: Traceback (most recent call last):
+No: 11 GFLOPS: 260.82/260.82 result: MeasureResult(costs=(0.0008875853812154695,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7145953178405762, timestamp=1660714967.240541) [('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: 0.00/260.82 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: 0.00/260.45 result: Traceback (most recent call last):
+No: 13 GFLOPS: 0.00/260.82 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: 0.00/260.45 result: Traceback (most recent call last):
+No: 14 GFLOPS: 0.00/260.82 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: 5.29/260.45 result: MeasureResult(costs=(0.043759062,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.805462121963501, timestamp=1660697034.50367) [('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: 3.34/260.45 result: MeasureResult(costs=(0.06928359925,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.539597749710083, timestamp=1660697035.7375107) [('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: 0.00/260.45 result: Traceback (most recent call last):
+No: 15 GFLOPS: 5.34/260.82 result: MeasureResult(costs=(0.04338321325,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.8048241138458252, timestamp=1660714971.793172) [('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: 3.35/260.82 result: MeasureResult(costs=(0.069062067,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.55722713470459, timestamp=1660714973.0338068) [('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: 0.00/260.82 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: 0.00/260.45 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: 28.06/260.45 result: MeasureResult(costs=(0.008249963714285715,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.280989170074463, timestamp=1660697046.7702246) [('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: 0.00/260.45 result: Traceback (most recent call last):
+No: 18 GFLOPS: 28.06/260.82 result: MeasureResult(costs=(0.008250961,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.285449504852295, timestamp=1660714984.027661) [('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: 0.00/260.82 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: 0.00/260.45 result: Traceback (most recent call last):
+No: 20 GFLOPS: 0.00/260.82 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: 0.001275
+Time cost of this operator: 0.001243
</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 49f7ed6ea..06f8c8b60 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 313.4 98.744 (1, 2, 10, 10, 3) 2 1 [313.4]
-tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.018 0.951 (1, 6, 10, 10) 1 1 [3.018]
-tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.97 0.306 (1, 1, 10, 10, 3) 1 1 [0.97]
-Total_time - 317.388 - - - - -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 315.6 98.728 (1, 2, 10, 10, 3) 2 1 [315.6]
+tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.094 0.968 (1, 6, 10, 10) 1 1 [3.094]
+tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.971 0.304 (1, 1, 10, 10, 3) 1 1 [0.971]
+Total_time - 319.665 - - - - -
</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 223.2 98.704 (1, 1, 10, 10, 6) 2 1 [223.2]
-tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.971 0.871 (1, 6, 10, 10) 1 1 [1.971]
-tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.96 0.425 (1, 1, 10, 10, 3) 1 1 [0.96]
-Total_time - 226.131 - - - - -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 81.625 96.775 (1, 6, 10, 10, 1) 2 1 [81.625]
+tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.767 2.095 (1, 6, 10, 10) 1 1 [1.767]
+tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.953 1.13 (1, 1, 10, 10, 3) 1 1 [0.953]
+Total_time - 84.345 - - - - -
</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 02825fe85..b803590f6 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/tmp2y4e1w0m/images/random'
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>'/tmp/tmpptt6e863/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/tmp2y4e1w0m/images/target contains 8144 images
-/tmp/tmp2y4e1w0m/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/tmpptt6e863/images/target contains 8144 images
+/tmp/tmpptt6e863/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 - 55s - loss: 0.2112 - accuracy: 0.9246 - val_loss: 0.1566 - val_accuracy: 0.9524
+328/328 - 55s - loss: 0.2233 - accuracy: 0.9226 - val_loss: 0.1452 - val_accuracy: 0.9554
Epoch 2/3
-328/328 - 53s - loss: 0.0963 - accuracy: 0.9641 - val_loss: 0.1274 - val_accuracy: 0.9622
+328/328 - 52s - loss: 0.0975 - accuracy: 0.9641 - val_loss: 0.1087 - val_accuracy: 0.9626
Epoch 3/3
-328/328 - 52s - loss: 0.0642 - accuracy: 0.9767 - val_loss: 0.1251 - val_accuracy: 0.9611
+328/328 - 52s - loss: 0.0657 - accuracy: 0.9744 - val_loss: 0.0945 - val_accuracy: 0.9653
-<keras.callbacks.History object at 0x7f69a974ef50>
+<keras.callbacks.History object at 0x7fc4005cac50>
</pre></div>
</div>
</div>
@@ -957,7 +957,7 @@ 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> ( 5 minutes 25.927 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 5 minutes 15.500 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 1e665e558..1df268b74 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>06:19.787</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
+<p><strong>06:08.879</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -336,26 +336,26 @@
</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>05:25.927</p></td>
+<td><p>05:15.500</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>00:42.430</p></td>
+<td><p>00:42.359</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>00:08.102</p></td>
+<td><p>00:07.572</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>00:03.326</p></td>
+<td><p>00:03.446</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>
+<tr class="row-odd"><td><p><a class="reference internal" href="micro_reference_vm.html#sphx-glr-how-to-work-with-microtvm-micro-reference-vm-py"><span class="std std-ref">microTVM Reference Virtual Machines</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_reference_vm.py</span></code>)</p></td>
<td><p>00:00.001</p></td>
<td><p>0.0 MB</p></td>
</tr>
-<tr class="row-even"><td><p><a class="reference internal" href="micro_reference_vm.html#sphx-glr-how-to-work-with-microtvm-micro-reference-vm-py"><span class="std std-ref">microTVM Reference Virtual Machines</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_reference_vm.py</span></code>)</p></td>
+<tr class="row-even"><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>
<td><p>00:00.001</p></td>
<td><p>0.0 MB</p></td>
</tr>
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 4d976fbb4..802acb55f 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>00:42.409</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
+<p><strong>00:42.763</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>00:30.599</p></td>
+<td><p>00:31.189</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>00:10.171</p></td>
+<td><p>00:10.045</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>00:01.631</p></td>
+<td><p>00:01.522</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 ca121c47d..3c3bd8d0e 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 0x7f69307e8a70>
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span><function my_cuda_math_rule at 0x7fc3980a5440>
</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 b1e6b733f..ba194492f 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>00:04.372</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
+<p><strong>00:03.998</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -336,19 +336,19 @@
</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>00:02.018</p></td>
+<td><p>00:01.854</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>00:01.062</p></td>
+<td><p>00:00.955</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>00:00.558</p></td>
+<td><p>00:00.508</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>00:00.551</p></td>
+<td><p>00:00.492</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>
@@ -360,11 +360,11 @@
<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>
-<td><p>00:00.026</p></td>
+<td><p>00:00.029</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="tuple_inputs.html#sphx-glr-how-to-work-with-schedules-tuple-inputs-py"><span class="std std-ref">Compute and Reduce with Tuple Inputs</span></a> (<code class="docutils literal notranslate"><span class="pre">tuple_inputs.py</span></code>)</p></td>
-<td><p>00:00.015</p></td>
+<td><p>00:00.016</p></td>
<td><p>0.0 MB</p></td>
</tr>
</tbody>
diff --git a/docs/how_to/work_with_schedules/tensorize.html b/docs/how_to/work_with_schedules/tensorize.html
index 9df6ebb50..b746325ea 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/tmpd4a8juer/input0.cc'\nsource_filename = \"/tmp/tmpd4a8juer/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/tmpqz7f6hvr/input0.cc'\nsource_filename = \"/tmp/tmpqz7f6hvr/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/affine__type_8h__incl.svg b/docs/reference/api/doxygen/affine__type_8h__incl.svg
index 4681de814..52439ff7c 100644
--- a/docs/reference/api/doxygen/affine__type_8h__incl.svg
+++ b/docs/reference/api/doxygen/affine__type_8h__incl.svg
@@ -12,1332 +12,1344 @@
<!-- Node0 -->
<g id="node1" class="node">
<title>Node0</title>
-<polygon fill="#bfbfbf" stroke="#000000" points="2508.5,-772.5 2508.5,-802.5 2623.5,-802.5 2623.5,-772.5 2508.5,-772.5"/>
-<text text-anchor="start" x="2516.5" y="-790.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/affine</text>
-<text text-anchor="middle" x="2566" y="-779.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_type.h</text>
+<polygon fill="#bfbfbf" stroke="#000000" points="1525.5,-772.5 1525.5,-802.5 1640.5,-802.5 1640.5,-772.5 1525.5,-772.5"/>
+<text text-anchor="start" x="1533.5" y="-790.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">include/tvm/ir/affine</text>
+<text text-anchor="middle" x="1583" y="-779.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_type.h</text>
</g>
<!-- Node1 -->
<g id="node2" class="node">
<title>Node1</title>
<g id="a_node2"><a xlink:href="ir_2expr_8h.html" target="_top" xlink:title="Base expr nodes in TVM. ">
-<polygon fill="#ffffff" stroke="#000000" points="2492.5,-716.5 2492.5,-735.5 2571.5,-735.5 2571.5,-716.5 2492.5,-716.5"/>
-<text text-anchor="middle" x="2532" y="-723.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/ir/expr.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1509.5,-716.5 1509.5,-735.5 1588.5,-735.5 1588.5,-716.5 1509.5,-716.5"/>
+<text text-anchor="middle" x="1549" y="-723.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/ir/expr.h</text>
</a>
</g>
</g>
<!-- Node0->Node1 -->
<g id="edge1" class="edge">
<title>Node0->Node1</title>
-<path fill="none" stroke="#191970" d="M2557.5955,-772.2977C2552.9743,-763.9388 2547.2031,-753.4997 2542.3111,-744.6509"/>
-<polygon fill="#191970" stroke="#191970" points="2545.3452,-742.9051 2537.4438,-735.8469 2539.2191,-746.292 2545.3452,-742.9051"/>
-</g>
-<!-- Node48 -->
-<g id="node49" class="node">
-<title>Node48</title>
-<g id="a_node49"><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="#000000" points="2598,-660.5 2598,-679.5 2678,-679.5 2678,-660.5 2598,-660.5"/>
-<text text-anchor="middle" x="2638" y="-667.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/ir/type.h</text>
+<path fill="none" stroke="#191970" d="M1574.5955,-772.2977C1569.9743,-763.9388 1564.2031,-753.4997 1559.3111,-744.6509"/>
+<polygon fill="#191970" stroke="#191970" points="1562.3452,-742.9051 1554.4438,-735.8469 1556.2191,-746.292 1562.3452,-742.9051"/>
+</g>
+<!-- Node49 -->
+<g id="node50" class="node">
+<title>Node49</title>
+<g id="a_node50"><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="#000000" points="2397,-660.5 2397,-679.5 2477,-679.5 2477,-660.5 2397,-660.5"/>
+<text text-anchor="middle" x="2437" y="-667.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/ir/type.h</text>
</a>
</g>
</g>
-<!-- Node0->Node48 -->
-<g id="edge159" class="edge">
-<title>Node0->Node48</title>
-<path fill="none" stroke="#191970" d="M2575.4127,-772.1389C2588.6207,-750.5843 2612.6165,-711.4245 2626.7494,-688.3604"/>
-<polygon fill="#191970" stroke="#191970" points="2629.9018,-689.9147 2632.1423,-679.5595 2623.9332,-686.2573 2629.9018,-689.9147"/>
+<!-- Node0->Node49 -->
+<g id="edge160" class="edge">
+<title>Node0->Node49</title>
+<path fill="none" stroke="#191970" d="M1640.693,-779.5621C1798.769,-757.8128 2236.9838,-697.5198 2386.3597,-676.9675"/>
+<polygon fill="#191970" stroke="#191970" points="2387.1782,-680.3879 2396.6078,-675.5575 2386.224,-673.4533 2387.1782,-680.3879"/>
</g>
<!-- Node2 -->
<g id="node3" class="node">
<title>Node2</title>
<g id="a_node3"><a xlink:href="ir_2span_8h.html" target="_top" xlink:title="Span information for debugging purposes. ">
-<polygon fill="#ffffff" stroke="#000000" points="2668.5,-604.5 2668.5,-623.5 2749.5,-623.5 2749.5,-604.5 2668.5,-604.5"/>
-<text text-anchor="middle" x="2709" y="-611.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/ir/span.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="2711.5,-604.5 2711.5,-623.5 2792.5,-623.5 2792.5,-604.5 2711.5,-604.5"/>
+<text text-anchor="middle" x="2752" y="-611.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/ir/span.h</text>
</a>
</g>
</g>
<!-- Node1->Node2 -->
<g id="edge2" class="edge">
<title>Node1->Node2</title>
-<path fill="none" stroke="#191970" d="M2538.1915,-716.1442C2547.5783,-701.9287 2566.698,-675.662 2589,-660 2610.5683,-644.8533 2637.7819,-633.8907 2660.8388,-626.4758"/>
-<polygon fill="#191970" stroke="#191970" points="2661.9143,-629.8069 2670.4373,-623.5134 2659.8499,-623.1182 2661.9143,-629.8069"/>
+<path fill="none" stroke="#191970" d="M1588.5975,-725.0809C1743.5093,-721.3265 2308.6523,-706.1021 2486,-680 2567.4228,-668.0162 2660.2506,-642.0963 2711.9008,-626.5432"/>
+<polygon fill="#191970" stroke="#191970" points="2713.1145,-629.8325 2721.6662,-623.578 2711.0807,-623.1345 2713.1145,-629.8325"/>
</g>
<!-- Node3 -->
<g id="node4" class="node">
<title>Node3</title>
<g id="a_node4"><a xlink:href="node_8h.html" target="_top" xlink:title="Definitions and helper macros for IR/AST nodes. ">
-<polygon fill="#ffffff" stroke="#000000" points="2461.5,-548.5 2461.5,-567.5 2560.5,-567.5 2560.5,-548.5 2461.5,-548.5"/>
-<text text-anchor="middle" x="2511" y="-555.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/node.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="2222.5,-548.5 2222.5,-567.5 2321.5,-567.5 2321.5,-548.5 2222.5,-548.5"/>
+<text text-anchor="middle" x="2272" y="-555.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/node.h</text>
</a>
</g>
</g>
<!-- Node1->Node3 -->
-<g id="edge152" class="edge">
+<g id="edge153" class="edge">
<title>Node1->Node3</title>
-<path fill="none" stroke="#191970" d="M2530.8108,-716.4862C2527.4912,-689.9293 2518.0706,-614.5651 2513.5207,-578.1655"/>
-<polygon fill="#191970" stroke="#191970" points="2516.9419,-577.3165 2512.2285,-567.8279 2509.9959,-578.1848 2516.9419,-577.3165"/>
+<path fill="none" stroke="#191970" d="M1588.6429,-716.7884C1711.2484,-688.2991 2084.2435,-601.6281 2221.0203,-569.8459"/>
+<polygon fill="#191970" stroke="#191970" points="2221.8891,-573.2374 2230.8374,-567.5648 2220.3047,-566.419 2221.8891,-573.2374"/>
</g>
<!-- Node8 -->
<g id="node9" class="node">
<title>Node8</title>
<g id="a_node9"><a xlink:href="object_8h.html" target="_top" xlink:title="A managed object in the TVM runtime. ">
-<polygon fill="#ffffff" stroke="#000000" points="1547.5,-123.5 1547.5,-142.5 1666.5,-142.5 1666.5,-123.5 1547.5,-123.5"/>
-<text text-anchor="middle" x="1607" y="-130.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/object.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1350.5,-123.5 1350.5,-142.5 1469.5,-142.5 1469.5,-123.5 1350.5,-123.5"/>
+<text text-anchor="middle" x="1410" y="-130.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/object.h</text>
</a>
</g>
</g>
<!-- Node1->Node8 -->
-<g id="edge154" class="edge">
+<g id="edge155" class="edge">
<title>Node1->Node8</title>
-<path fill="none" stroke="#191970" d="M2492.2543,-725.3145C2247.5635,-720.8405 956,-693.4616 956,-614 956,-614 956,-614 956,-558 956,-438.5428 1073.252,-465.5005 1165,-389 1216.3164,-346.2119 1224.7187,-329.9581 1279,-291 1358.3254,-234.0675 1379.1367,-219.4616 1468,-179 1498.1843,-165.2564 1533.708,-153.573 1561.2875,-145.4199"/>
-<polygon fill="#191970" stroke="#191970" points="1562.4384,-148.7303 1571.0643,-142.5785 1560.4848,-142.0084 1562.4384,-148.7303"/>
+<path fill="none" stroke="#191970" d="M1509.2205,-725.1433C1338.4362,-720.9488 676,-699.1886 676,-614 676,-614 676,-614 676,-502 676,-405.3346 751.268,-398.5869 839,-358 940.8829,-310.8665 984.4019,-362.7544 1089,-322 1141.4358,-301.5695 1248.4762,-205.7266 1298,-179 1322.0099,-166.0425 1350.4635,-154.3996 1372.6831,-146.069"/>
+<polygon fill="#191970" stroke="#191970" points="1373.9321,-149.3389 1382.1032,-142.5947 1371.5098,-142.7713 1373.9321,-149.3389"/>
</g>
<!-- Node14 -->
<g id="node15" class="node">
<title>Node14</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="2317,-62 2317,-81 2361,-81 2361,-62 2317,-62"/>
-<text text-anchor="middle" x="2339" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">string</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="2470,-62 2470,-81 2514,-81 2514,-62 2470,-62"/>
+<text text-anchor="middle" x="2492" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">string</text>
</g>
<!-- Node1->Node14 -->
-<g id="edge157" class="edge">
+<g id="edge158" class="edge">
<title>Node1->Node14</title>
-<path fill="none" stroke="#191970" d="M2571.5113,-723.7942C2675.3519,-716.9239 2948,-691.5052 2948,-614 2948,-614 2948,-614 2948,-373.5 2948,-243.3587 2495.9872,-113.3059 2370.8002,-79.7828"/>
-<polygon fill="#191970" stroke="#191970" points="2371.598,-76.3734 2361.0344,-77.1865 2369.7995,-83.1384 2371.598,-76.3734"/>
+<path fill="none" stroke="#191970" d="M1588.6158,-724.957C1816.285,-718.7097 2948,-684.1281 2948,-614 2948,-614 2948,-614 2948,-245 2948,-200.5978 2909.5501,-204.212 2873,-179 2829.1409,-148.7462 2817.3094,-140.5465 2767,-123 2682.2576,-93.4442 2576.5848,-79.7245 2524.3405,-74.3779"/>
+<polygon fill="#191970" stroke="#191970" points="2524.6231,-70.8889 2514.3277,-73.3897 2523.9355,-77.8551 2524.6231,-70.8889"/>
</g>
<!-- Node15 -->
<g id="node16" class="node">
<title>Node15</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="2836.5,-62 2836.5,-81 2905.5,-81 2905.5,-62 2836.5,-62"/>
-<text text-anchor="middle" x="2871" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">type_traits</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="2761.5,-62 2761.5,-81 2830.5,-81 2830.5,-62 2761.5,-62"/>
+<text text-anchor="middle" x="2796" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">type_traits</text>
</g>
<!-- Node1->Node15 -->
-<g id="edge158" class="edge">
+<g id="edge159" class="edge">
<title>Node1->Node15</title>
-<path fill="none" stroke="#191970" d="M2571.6835,-722.8735C2653.7395,-716.11 2839.4632,-699.1741 2900,-680 2959.5173,-661.1488 3024,-676.4314 3024,-614 3024,-614 3024,-614 3024,-189 3024,-133.6159 2960.1582,-100.755 2914.9533,-84.429"/>
-<polygon fill="#191970" stroke="#191970" points="2915.931,-81.0637 2905.3362,-81.0999 2913.6411,-87.6786 2915.931,-81.0637"/>
+<path fill="none" stroke="#191970" d="M1588.5174,-725.148C1761.8018,-721.294 2455.459,-704.6053 2672,-680 2830.153,-662.0293 3024,-773.1707 3024,-614 3024,-614 3024,-614 3024,-189 3024,-107.4203 2907.2378,-82.4276 2840.8286,-74.8102"/>
+<polygon fill="#191970" stroke="#191970" points="2840.9411,-71.3027 2830.6283,-73.7309 2840.2044,-78.2638 2840.9411,-71.3027"/>
</g>
<!-- Node24 -->
<g id="node25" class="node">
<title>Node24</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="714,-179.5 714,-198.5 778,-198.5 778,-179.5 714,-179.5"/>
-<text text-anchor="middle" x="746" y="-186.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">algorithm</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="222,-179.5 222,-198.5 286,-198.5 286,-179.5 222,-179.5"/>
+<text text-anchor="middle" x="254" y="-186.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">algorithm</text>
</g>
<!-- Node1->Node24 -->
-<g id="edge155" class="edge">
+<g id="edge156" class="edge">
<title>Node1->Node24</title>
-<path fill="none" stroke="#191970" d="M2492.1622,-725.3959C2265.7116,-721.8941 1138.53,-703.5359 792,-680 657.73,-670.8806 190,-748.5794 190,-614 190,-614 190,-614 190,-373.5 190,-310.9576 248.203,-319.2533 304,-291 350.6468,-267.38 367.0044,-273.2555 416,-255 437.2729,-247.0739 441.1534,-241.1711 463,-235 567.2213,-205.5602 597.6702,-219.561 704,-199 704.1012,-198.9804 704.2025,-198.9608 704.3039,-198.9411"/>
-<polygon fill="#191970" stroke="#191970" points="704.8831,-202.3957 713.9575,-196.9271 703.4535,-195.5432 704.8831,-202.3957"/>
+<path fill="none" stroke="#191970" d="M1509.1533,-725.2628C1266.6991,-720.5224 0,-692.0229 0,-614 0,-614 0,-614 0,-373.5 0,-269.9115 140.0031,-218.441 212.126,-198.7985"/>
+<polygon fill="#191970" stroke="#191970" points="213.0557,-202.173 221.8307,-196.236 211.2685,-195.405 213.0557,-202.173"/>
</g>
<!-- Node26 -->
<g id="node27" class="node">
<title>Node26</title>
<g id="a_node27"><a xlink:href="string_8h.html" target="_top" xlink:title="Runtime String container types. ">
-<polygon fill="#ffffff" stroke="#000000" points="992,-291.5 992,-321.5 1118,-321.5 1118,-291.5 992,-291.5"/>
-<text text-anchor="start" x="1000" y="-309.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
-<text text-anchor="middle" x="1055" y="-298.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/string.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="954,-291.5 954,-321.5 1080,-321.5 1080,-291.5 954,-291.5"/>
+<text text-anchor="start" x="962" y="-309.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
+<text text-anchor="middle" x="1017" y="-298.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/string.h</text>
</a>
</g>
</g>
<!-- Node1->Node26 -->
-<g id="edge153" class="edge">
+<g id="edge154" class="edge">
<title>Node1->Node26</title>
-<path fill="none" stroke="#191970" d="M2492.2127,-725.5028C2247.011,-722.3635 948.9629,-704.6267 869,-680 817.0389,-663.9972 766,-668.3696 766,-614 766,-614 766,-614 766,-440.5 766,-402.8735 757.8949,-384.0973 785,-358 812.5634,-331.4614 911.5425,-317.9627 981.6492,-311.5808"/>
-<polygon fill="#191970" stroke="#191970" points="982.1166,-315.0533 991.7713,-310.6902 981.503,-308.0803 982.1166,-315.0533"/>
+<path fill="none" stroke="#191970" d="M1509.4333,-725.1717C1347.8991,-721.6241 743.118,-706.6651 662,-680 611.4113,-663.3705 562,-667.2518 562,-614 562,-614 562,-614 562,-440.5 562,-320.3399 871.3503,-327.1004 943.5165,-321.521"/>
+<polygon fill="#191970" stroke="#191970" points="944.2896,-324.9559 953.832,-320.3524 943.5015,-318.0004 944.2896,-324.9559"/>
</g>
-<!-- Node44 -->
-<g id="node45" class="node">
-<title>Node44</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="794,-364 794,-383 838,-383 838,-364 794,-364"/>
-<text text-anchor="middle" x="816" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">limits</text>
+<!-- Node45 -->
+<g id="node46" class="node">
+<title>Node45</title>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="848,-364 848,-383 892,-383 892,-364 848,-364"/>
+<text text-anchor="middle" x="870" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">limits</text>
</g>
-<!-- Node1->Node44 -->
-<g id="edge156" class="edge">
-<title>Node1->Node44</title>
-<path fill="none" stroke="#191970" d="M2492.423,-725.4679C2254.1378,-722.1874 1018.1149,-704.0588 942,-680 891.2243,-663.9506 842,-667.2518 842,-614 842,-614 842,-614 842,-502 842,-462.5442 830.0785,-417.603 822.3806,-392.6964"/>
-<polygon fill="#191970" stroke="#191970" points="825.6719,-391.4974 819.2929,-383.0381 819.0043,-393.629 825.6719,-391.4974"/>
+<!-- Node1->Node45 -->
+<g id="edge157" class="edge">
+<title>Node1->Node45</title>
+<path fill="none" stroke="#191970" d="M1509.2615,-724.372C1353.0916,-717.4553 790,-687.4253 790,-614 790,-614 790,-614 790,-502 790,-455.9602 827.4675,-412.7534 851.1203,-390.0646"/>
+<polygon fill="#191970" stroke="#191970" points="853.5878,-392.5496 858.538,-383.1823 848.8267,-387.4182 853.5878,-392.5496"/>
</g>
-<!-- Node1->Node48 -->
-<g id="edge145" class="edge">
-<title>Node1->Node48</title>
-<path fill="none" stroke="#191970" d="M2550.4638,-716.2455C2567.1074,-707.4527 2591.7993,-694.4079 2610.7192,-684.4125"/>
-<polygon fill="#191970" stroke="#191970" points="2612.5408,-687.4086 2619.7478,-679.6427 2609.2709,-681.2193 2612.5408,-687.4086"/>
+<!-- Node1->Node49 -->
+<g id="edge146" class="edge">
+<title>Node1->Node49</title>
+<path fill="none" stroke="#191970" d="M1588.5909,-723.5033C1732.4288,-714.4324 2227.2514,-683.2274 2386.909,-673.1589"/>
+<polygon fill="#191970" stroke="#191970" points="2387.1735,-676.6493 2396.9334,-672.5267 2386.7329,-669.6631 2387.1735,-676.6493"/>
</g>
<!-- Node2->Node3 -->
<g id="edge3" class="edge">
<title>Node2->Node3</title>
-<path fill="none" stroke="#191970" d="M2674.9626,-604.3733C2641.6106,-594.9404 2590.6473,-580.5265 2554.3998,-570.2747"/>
-<polygon fill="#191970" stroke="#191970" points="2555.1684,-566.8548 2544.5933,-567.5011 2553.2632,-573.5906 2555.1684,-566.8548"/>
+<path fill="none" stroke="#191970" d="M2711.4892,-609.2737C2626.9004,-599.405 2431.1499,-576.5675 2331.9458,-564.9937"/>
+<polygon fill="#191970" stroke="#191970" points="2332.1238,-561.4908 2321.7855,-563.8083 2331.3125,-568.4436 2332.1238,-561.4908"/>
</g>
<!-- Node2->Node8 -->
-<g id="edge143" class="edge">
+<g id="edge144" class="edge">
<title>Node2->Node8</title>
-<path fill="none" stroke="#191970" d="M2707.631,-604.4512C2700.6813,-559.8513 2664.9643,-370.8615 2549,-291 2406.4421,-192.8242 1869.9446,-149.6976 1676.8164,-137.1251"/>
-<polygon fill="#191970" stroke="#191970" points="1676.9878,-133.629 1666.7837,-136.4795 1676.5382,-140.6146 1676.9878,-133.629"/>
+<path fill="none" stroke="#191970" d="M2747.2998,-604.4593C2729.6931,-569.2703 2664.7634,-444.3316 2589,-358 2524.3602,-284.3437 2498.4507,-267.5026 2406,-235 2273.5045,-188.419 2234.3909,-196.176 2095,-179 1871.2043,-151.4235 1604.0909,-139.4836 1479.8991,-135.1294"/>
+<polygon fill="#191970" stroke="#191970" points="1479.8829,-131.6269 1469.7681,-134.78 1479.6415,-138.6227 1479.8829,-131.6269"/>
</g>
<!-- Node2->Node14 -->
-<g id="edge144" class="edge">
+<g id="edge145" class="edge">
<title>Node2->Node14</title>
-<path fill="none" stroke="#191970" d="M2711.0402,-604.2969C2715.0547,-583.3608 2722.702,-532.5621 2712,-492 2669.639,-331.4451 2622.6277,-299.1367 2508,-179 2482.3671,-152.1352 2476.8214,-143.7076 2446,-123 2422.2564,-107.0476 2392.8186,-93.259 2370.8339,-83.9747"/>
-<polygon fill="#191970" stroke="#191970" points="2371.8831,-80.6212 2361.3045,-80.0343 2369.2082,-87.09 2371.8831,-80.6212"/>
+<path fill="none" stroke="#191970" d="M2775.9954,-604.4657C2811.0144,-588.7839 2872,-554.0082 2872,-502 2872,-502 2872,-502 2872,-245 2872,-168.7085 2615.3387,-100.63 2523.8862,-78.794"/>
+<polygon fill="#191970" stroke="#191970" points="2524.6095,-75.3687 2514.0729,-76.4776 2523.0013,-82.1814 2524.6095,-75.3687"/>
</g>
<!-- Node4 -->
<g id="node5" class="node">
<title>Node4</title>
<g id="a_node5"><a xlink:href="reflection_8h.html" target="_top" xlink:title="Reflection and serialization of compiler IR/AST nodes. ">
-<polygon fill="#ffffff" stroke="#000000" points="2143.5,-492.5 2143.5,-511.5 2264.5,-511.5 2264.5,-492.5 2143.5,-492.5"/>
-<text text-anchor="middle" x="2204" y="-499.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/reflection.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1989.5,-492.5 1989.5,-511.5 2110.5,-511.5 2110.5,-492.5 1989.5,-492.5"/>
+<text text-anchor="middle" x="2050" y="-499.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/reflection.h</text>
</a>
</g>
</g>
<!-- Node3->Node4 -->
<g id="edge4" class="edge">
<title>Node3->Node4</title>
-<path fill="none" stroke="#191970" d="M2461.3339,-548.9404C2408.1839,-539.2453 2323.7672,-523.8468 2266.2242,-513.3503"/>
-<polygon fill="#191970" stroke="#191970" points="2266.6283,-509.8664 2256.1626,-511.515 2265.3721,-516.7528 2266.6283,-509.8664"/>
+<path fill="none" stroke="#191970" d="M2234.0891,-548.4369C2196.335,-538.9133 2138.2735,-524.2672 2097.4736,-513.9753"/>
+<polygon fill="#191970" stroke="#191970" points="2098.3069,-510.576 2087.7546,-511.5237 2096.5947,-517.3633 2098.3069,-510.576"/>
</g>
<!-- Node5 -->
<g id="node6" class="node">
<title>Node5</title>
<g id="a_node6"><a xlink:href="structural__equal_8h.html" target="_top" xlink:title="Structural equality comparison. ">
-<polygon fill="#ffffff" stroke="#000000" points="1984.5,-425.5 1984.5,-455.5 2097.5,-455.5 2097.5,-425.5 1984.5,-425.5"/>
-<text text-anchor="start" x="1992.5" y="-443.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/structural</text>
-<text text-anchor="middle" x="2041" y="-432.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_equal.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1597.5,-425.5 1597.5,-455.5 1710.5,-455.5 1710.5,-425.5 1597.5,-425.5"/>
+<text text-anchor="start" x="1605.5" y="-443.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/structural</text>
+<text text-anchor="middle" x="1654" y="-432.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_equal.h</text>
</a>
</g>
</g>
<!-- Node3->Node5 -->
-<g id="edge134" class="edge">
+<g id="edge135" class="edge">
<title>Node3->Node5</title>
-<path fill="none" stroke="#191970" d="M2461.3529,-555.2001C2372.9799,-549.6805 2192.2025,-535.8174 2134,-512 2106.703,-500.8296 2080.528,-479.416 2062.9413,-462.9231"/>
-<polygon fill="#191970" stroke="#191970" points="2065.1219,-460.1639 2055.4921,-455.7461 2060.2651,-465.2049 2065.1219,-460.1639"/>
+<path fill="none" stroke="#191970" d="M2222.4186,-551.0641C2164.3801,-542.7368 2065.4464,-527.8927 1981,-512 1890.284,-494.9274 1786.2469,-471.5134 1720.3465,-456.2007"/>
+<polygon fill="#191970" stroke="#191970" points="1721.0394,-452.7685 1710.506,-453.9085 1719.4512,-459.586 1721.0394,-452.7685"/>
</g>
<!-- Node3->Node8 -->
-<g id="edge138" class="edge">
+<g id="edge139" class="edge">
<title>Node3->Node8</title>
-<path fill="none" stroke="#191970" d="M2537.9644,-548.4934C2554.1714,-541.2188 2573.4508,-529.4326 2583,-512 2601.5763,-478.0879 2604.1327,-457.3808 2583,-425 2545.021,-366.8064 2075.4364,-195.7699 2008,-179 1894.038,-150.6603 1757.4404,-139.7118 1676.6783,-135.5296"/>
-<polygon fill="#191970" stroke="#191970" points="1676.79,-132.0309 1666.6283,-135.0293 1676.4419,-139.0223 1676.79,-132.0309"/>
+<path fill="none" stroke="#191970" d="M2320.0145,-548.4579C2396.0773,-532.2581 2538.6901,-497.1014 2568,-456 2614.2346,-391.1651 2583.3471,-363.3909 2473,-291 2270.6038,-158.2222 2179.4571,-214.4184 1940,-179 1853.132,-166.1512 1602.8926,-147.0104 1479.9115,-138.0141"/>
+<polygon fill="#191970" stroke="#191970" points="1480.076,-134.5168 1469.8478,-137.2797 1479.5665,-141.4983 1480.076,-134.5168"/>
</g>
<!-- Node9 -->
<g id="node10" class="node">
<title>Node9</title>
<g id="a_node10"><a xlink:href="c__runtime__api_8h.html" target="_top" xlink:title="tvm/runtime/c_runtime\l_api.h">
-<polygon fill="#ffffff" stroke="#000000" points="2483.5,-56.5 2483.5,-86.5 2612.5,-86.5 2612.5,-56.5 2483.5,-56.5"/>
-<text text-anchor="start" x="2491.5" y="-74.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/c_runtime</text>
-<text text-anchor="middle" x="2548" y="-63.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_api.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1921.5,-56.5 1921.5,-86.5 2050.5,-86.5 2050.5,-56.5 1921.5,-56.5"/>
+<text text-anchor="start" x="1929.5" y="-74.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/c_runtime</text>
+<text text-anchor="middle" x="1986" y="-63.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_api.h</text>
</a>
</g>
</g>
<!-- Node3->Node9 -->
-<g id="edge136" class="edge">
+<g id="edge137" class="edge">
<title>Node3->Node9</title>
-<path fill="none" stroke="#191970" d="M2560.6581,-552.2094C2608.5188,-545.4857 2675.9778,-532.6035 2692,-512 2746.9581,-441.3275 2790.7846,-313.8512 2654,-123 2643.8375,-108.8205 2628.6855,-98.3734 2612.9951,-90.7454"/>
-<polygon fill="#191970" stroke="#191970" points="2614.2836,-87.4871 2603.7283,-86.5729 2611.4096,-93.87 2614.2836,-87.4871"/>
+<path fill="none" stroke="#191970" d="M2321.7269,-553.2756C2369.1402,-547.5625 2441.8747,-535.6106 2501,-512 2544.4018,-494.6683 2567.4065,-496.9101 2590,-456 2594.75,-447.3992 2597.7283,-390.0156 2589,-358 2543.8966,-192.5602 2446.4396,-174.8857 2283,-123 2209.3264,-99.6115 2121.7975,-86.0678 2060.9274,-78.7946"/>
+<polygon fill="#191970" stroke="#191970" points="2060.9768,-75.2767 2050.6386,-77.5945 2060.1658,-82.2296 2060.9768,-75.2767"/>
</g>
<!-- Node3->Node14 -->
-<g id="edge139" class="edge">
+<g id="edge140" class="edge">
<title>Node3->Node14</title>
-<path fill="none" stroke="#191970" d="M2549.2484,-548.3952C2568.178,-541.5364 2589.4074,-530.2061 2601,-512 2654.4508,-428.0559 2603.5461,-382.7486 2565,-291 2523.7154,-192.7333 2414.3284,-116.8818 2364.4424,-86.2605"/>
-<polygon fill="#191970" stroke="#191970" points="2366.2198,-83.2452 2355.8499,-81.074 2362.6024,-89.2381 2366.2198,-83.2452"/>
+<path fill="none" stroke="#191970" d="M2321.58,-554.8533C2423.1381,-548.0513 2647.7604,-531.0462 2677,-512 2736.7514,-473.0789 2758,-444.8098 2758,-373.5 2758,-373.5 2758,-373.5 2758,-306.5 2758,-208.0696 2711.6246,-183.5224 2634,-123 2601.348,-97.5419 2554.9596,-83.8969 2524.1979,-77.1772"/>
+<polygon fill="#191970" stroke="#191970" points="2524.5174,-73.6697 2514.017,-75.0812 2523.1058,-80.5259 2524.5174,-73.6697"/>
</g>
<!-- Node3->Node15 -->
-<g id="edge140" class="edge">
+<g id="edge141" class="edge">
<title>Node3->Node15</title>
-<path fill="none" stroke="#191970" d="M2560.5893,-554.7031C2603.1521,-550.0607 2664.8997,-538.8866 2712,-512 2813.0217,-454.3331 2817.5165,-411.348 2892,-322 2915.6224,-293.6633 2927.6906,-289.4071 2941,-255 2953.224,-223.399 2953.0455,-212.3392 2947,-179 2942.2572,-152.8452 2940.7559,-145.1095 2926,-123 2916.9729,-109.4742 2903.7436,-96.9406 2892.522,-87.6251"/>
-<polygon fill="#191970" stroke="#191970" points="2894.4327,-84.6719 2884.4349,-81.1657 2890.064,-90.1414 2894.4327,-84.6719"/>
+<path fill="none" stroke="#191970" d="M2321.7609,-556.6799C2420.415,-553.4198 2637.0489,-542.8741 2704,-512 2860.7359,-439.7222 2986,-417.5984 2986,-245 2986,-245 2986,-245 2986,-189 2986,-121.0882 2897.1098,-91.1069 2840.9296,-78.94"/>
+<polygon fill="#191970" stroke="#191970" points="2841.3485,-75.4531 2830.8489,-76.8703 2839.9407,-82.31 2841.3485,-75.4531"/>
</g>
<!-- Node16 -->
<g id="node17" class="node">
<title>Node16</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="410.5,-62 410.5,-81 455.5,-81 455.5,-62 410.5,-62"/>
-<text text-anchor="middle" x="433" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">utility</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="564.5,-62 564.5,-81 609.5,-81 609.5,-62 564.5,-62"/>
+<text text-anchor="middle" x="587" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">utility</text>
</g>
<!-- Node3->Node16 -->
-<g id="edge141" class="edge">
+<g id="edge142" class="edge">
<title>Node3->Node16</title>
-<path fill="none" stroke="#191970" d="M2461.0369,-557.5282C2169.1354,-554.2837 682.465,-530.9287 252,-389 129.491,-348.6075 0,-373.9961 0,-245 0,-245 0,-245 0,-189 0,-106.1429 299.0468,-79.852 400.0964,-73.3333"/>
-<polygon fill="#191970" stroke="#191970" points="400.5659,-76.811 410.3297,-72.6977 400.1319,-69.8244 400.5659,-76.811"/>
+<path fill="none" stroke="#191970" d="M2222.2197,-557.6925C2020.5047,-555.8665 1244.6196,-543.2304 614,-456 452.3548,-433.6404 396.5272,-466.6471 253,-389 146.9749,-331.6412 101.3767,-192.3167 200,-123 256.9761,-82.9547 471.8404,-74.0153 554.4055,-72.0466"/>
+<polygon fill="#191970" stroke="#191970" points="554.5548,-75.5443 564.4757,-71.8259 554.4014,-68.546 554.5548,-75.5443"/>
</g>
<!-- Node18 -->
<g id="node19" class="node">
<title>Node18</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1402.5,-235.5 1402.5,-254.5 1449.5,-254.5 1449.5,-235.5 1402.5,-235.5"/>
-<text text-anchor="middle" x="1426" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">vector</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1384.5,-235.5 1384.5,-254.5 1431.5,-254.5 1431.5,-235.5 1384.5,-235.5"/>
+<text text-anchor="middle" x="1408" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">vector</text>
</g>
<!-- Node3->Node18 -->
-<g id="edge142" class="edge">
+<g id="edge143" class="edge">
<title>Node3->Node18</title>
-<path fill="none" stroke="#191970" d="M2461.2832,-556.1715C2251.0549,-548.1025 1441.4203,-513.5089 1344,-456 1278.5426,-417.3593 1255.3529,-353.2311 1299,-291 1304.6311,-282.9713 1357.1256,-265.7952 1392.8371,-254.8535"/>
-<polygon fill="#191970" stroke="#191970" points="1393.8605,-258.2006 1402.4117,-251.9455 1391.8261,-251.5027 1393.8605,-258.2006"/>
+<path fill="none" stroke="#191970" d="M2222.3317,-555.9133C2048.5058,-547.9595 1478,-516.0965 1478,-440.5 1478,-440.5 1478,-440.5 1478,-373.5 1478,-329.1535 1445.4523,-285.5371 1424.7445,-262.2749"/>
+<polygon fill="#191970" stroke="#191970" points="1427.0821,-259.6466 1417.7332,-254.6618 1421.9331,-264.3887 1427.0821,-259.6466"/>
</g>
<!-- Node22 -->
<g id="node23" class="node">
<title>Node22</title>
<g id="a_node23"><a xlink:href="runtime_2memory_8h.html" target="_top" xlink:title="Runtime memory management. ">
-<polygon fill="#ffffff" stroke="#000000" points="566.5,-179.5 566.5,-198.5 695.5,-198.5 695.5,-179.5 566.5,-179.5"/>
-<text text-anchor="middle" x="631" y="-186.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/memory.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="723.5,-179.5 723.5,-198.5 852.5,-198.5 852.5,-179.5 723.5,-179.5"/>
+<text text-anchor="middle" x="788" y="-186.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/memory.h</text>
</a>
</g>
</g>
<!-- Node3->Node22 -->
-<g id="edge137" class="edge">
+<g id="edge138" class="edge">
<title>Node3->Node22</title>
-<path fill="none" stroke="#191970" d="M2461.4477,-557.142C2314.9913,-554.3035 1871.4557,-543.5305 1505,-512 1198.361,-485.6162 1124.3733,-454.3079 818,-425 756.4846,-419.1154 305.2587,-434.0886 263,-389 251.3527,-376.5727 252.955,-315.5428 269,-291 297.2989,-247.7133 322.2202,-252.1784 371,-235 431.5227,-213.6861 503.5938,-201.9276 556.489,-195.6473"/>
-<polygon fill="#191970" stroke="#191970" points="556.9341,-199.1193 566.469,-194.5002 556.1347,-192.1651 556.9341,-199.1193"/>
+<path fill="none" stroke="#191970" d="M2222.291,-554.7138C2109.6117,-547.2512 1826.1236,-528.4001 1589,-512 1238.289,-487.7439 1149.104,-497.3876 800,-456 716.7759,-446.1335 697.0265,-436.4108 614,-425 540.4808,-414.8958 337.3352,-438.4487 282,-389 230.9544,-343.3846 231.5761,-277.8055 285,-235 317.6393,-208.848 580.9834,-196.1275 713.0475,-191.3435"/>
+<polygon fill="#191970" stroke="#191970" points="713.2263,-194.8394 723.0954,-190.9856 712.9771,-187.8438 713.2263,-194.8394"/>
</g>
-<!-- Node33 -->
-<g id="node34" class="node">
-<title>Node33</title>
-<g id="a_node34"><a xlink:href="structural__hash_8h.html" target="_top" xlink:title="tvm/node/structural\l_hash.h">
-<polygon fill="#ffffff" stroke="#000000" points="2192.5,-425.5 2192.5,-455.5 2305.5,-455.5 2305.5,-425.5 2192.5,-425.5"/>
-<text text-anchor="start" x="2200.5" y="-443.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/structural</text>
-<text text-anchor="middle" x="2249" y="-432.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_hash.h</text>
+<!-- Node34 -->
+<g id="node35" class="node">
+<title>Node34</title>
+<g id="a_node35"><a xlink:href="structural__hash_8h.html" target="_top" xlink:title="tvm/node/structural\l_hash.h">
+<polygon fill="#ffffff" stroke="#000000" points="2113.5,-425.5 2113.5,-455.5 2226.5,-455.5 2226.5,-425.5 2113.5,-425.5"/>
+<text text-anchor="start" x="2121.5" y="-443.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/structural</text>
+<text text-anchor="middle" x="2170" y="-432.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_hash.h</text>
</a>
</g>
</g>
-<!-- Node3->Node33 -->
-<g id="edge135" class="edge">
-<title>Node3->Node33</title>
-<path fill="none" stroke="#191970" d="M2489.5594,-548.3845C2446.4668,-529.0586 2348.6369,-485.1845 2291.8998,-459.7394"/>
-<polygon fill="#191970" stroke="#191970" points="2293.0407,-456.4152 2282.484,-455.5167 2290.1762,-462.8023 2293.0407,-456.4152"/>
+<!-- Node3->Node34 -->
+<g id="edge136" class="edge">
+<title>Node3->Node34</title>
+<path fill="none" stroke="#191970" d="M2263.6529,-548.3845C2247.6819,-529.9865 2212.3982,-489.341 2189.9896,-463.5272"/>
+<polygon fill="#191970" stroke="#191970" points="2192.4058,-460.9715 2183.2073,-455.7143 2187.1197,-465.5603 2192.4058,-460.9715"/>
</g>
-<!-- Node46 -->
-<g id="node47" class="node">
-<title>Node46</title>
-<g id="a_node47"><a xlink:href="repr__printer_8h.html" target="_top" xlink:title="Printer class to print repr string of each AST/IR nodes. ">
-<polygon fill="#ffffff" stroke="#000000" points="2442.5,-492.5 2442.5,-511.5 2573.5,-511.5 2573.5,-492.5 2442.5,-492.5"/>
-<text text-anchor="middle" x="2508" y="-499.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/repr_printer.h</text>
+<!-- Node47 -->
+<g id="node48" class="node">
+<title>Node47</title>
+<g id="a_node48"><a xlink:href="repr__printer_8h.html" target="_top" xlink:title="Printer class to print repr string of each AST/IR nodes. ">
+<polygon fill="#ffffff" stroke="#000000" points="2253.5,-492.5 2253.5,-511.5 2384.5,-511.5 2384.5,-492.5 2253.5,-492.5"/>
+<text text-anchor="middle" x="2319" y="-499.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/repr_printer.h</text>
</a>
</g>
</g>
-<!-- Node3->Node46 -->
-<g id="edge131" class="edge">
-<title>Node3->Node46</title>
-<path fill="none" stroke="#191970" d="M2510.4774,-548.2455C2510.086,-540.9382 2509.5372,-530.6944 2509.0556,-521.7046"/>
-<polygon fill="#191970" stroke="#191970" points="2512.5466,-521.4411 2508.5166,-511.6427 2505.5566,-521.8156 2512.5466,-521.4411"/>
+<!-- Node3->Node47 -->
+<g id="edge132" class="edge">
+<title>Node3->Node47</title>
+<path fill="none" stroke="#191970" d="M2280.1868,-548.2455C2286.8589,-540.2958 2296.4479,-528.8706 2304.4253,-519.3656"/>
+<polygon fill="#191970" stroke="#191970" points="2307.1592,-521.5525 2310.907,-511.6427 2301.7974,-517.0524 2307.1592,-521.5525"/>
</g>
<!-- Node4->Node5 -->
<g id="edge5" class="edge">
<title>Node4->Node5</title>
-<path fill="none" stroke="#191970" d="M2178.5313,-492.3906C2155.0414,-483.5279 2119.5155,-470.1239 2090.455,-459.1594"/>
-<polygon fill="#191970" stroke="#191970" points="2091.466,-455.8001 2080.8743,-455.5446 2088.9949,-462.3494 2091.466,-455.8001"/>
+<path fill="none" stroke="#191970" d="M1989.2872,-494.7763C1923.8383,-486.6701 1816.7037,-472.4762 1725,-456 1723.5254,-455.7351 1722.0348,-455.4613 1720.5329,-455.1802"/>
+<polygon fill="#191970" stroke="#191970" points="1720.9896,-451.7036 1710.5056,-453.2324 1719.6547,-458.5752 1720.9896,-451.7036"/>
</g>
<!-- Node4->Node8 -->
-<g id="edge94" class="edge">
+<g id="edge95" class="edge">
<title>Node4->Node8</title>
-<path fill="none" stroke="#191970" d="M2198.0235,-492.2032C2192.8888,-483.1737 2185.8729,-469.2108 2183,-456 2167.3137,-383.868 2238.1652,-355.3519 2202,-291 2178.5482,-249.2703 2153.105,-255.8162 2110,-235 2045.1693,-203.692 2027.9027,-196.2287 1958,-179 1862.1645,-155.3797 1748.4314,-143.2917 1676.9747,-137.5625"/>
-<polygon fill="#191970" stroke="#191970" points="1676.9882,-134.053 1666.7458,-136.7628 1676.4425,-141.0317 1676.9882,-134.053"/>
+<path fill="none" stroke="#191970" d="M2050.8168,-492.4238C2053.0699,-459.4244 2055.8392,-349.2738 1999,-291 1926.52,-216.6906 1619.4896,-163.5169 1479.7608,-142.6996"/>
+<polygon fill="#191970" stroke="#191970" points="1479.9953,-139.1964 1469.5913,-141.1977 1478.9726,-146.1213 1479.9953,-139.1964"/>
</g>
<!-- Node4->Node9 -->
-<g id="edge90" class="edge">
+<g id="edge91" class="edge">
<title>Node4->Node9</title>
-<path fill="none" stroke="#191970" d="M2264.5224,-493.236C2333.0045,-480.6197 2444.6593,-451.5768 2519,-389 2595.0488,-324.9854 2621.6049,-297.2053 2637,-199 2642.2313,-165.6298 2653.8746,-152.2606 2637,-123 2629.1721,-109.4263 2616.2494,-99.1064 2602.646,-91.4046"/>
-<polygon fill="#191970" stroke="#191970" points="2604.1398,-88.2363 2593.6549,-86.7137 2600.9018,-94.4424 2604.1398,-88.2363"/>
+<path fill="none" stroke="#191970" d="M2110.6837,-494.8962C2158.6974,-487.9041 2220.2054,-475.4194 2236,-456 2282.2935,-399.0823 2271.587,-357.6869 2241,-291 2196.6172,-194.235 2086.1883,-123.9994 2026.2458,-91.5625"/>
+<polygon fill="#191970" stroke="#191970" points="2027.5488,-88.2911 2017.0768,-86.6828 2024.2602,-94.4706 2027.5488,-88.2911"/>
</g>
<!-- Node4->Node14 -->
-<g id="edge128" class="edge">
+<g id="edge129" class="edge">
<title>Node4->Node14</title>
-<path fill="none" stroke="#191970" d="M2245.1641,-492.4234C2275.0923,-484.6583 2316.1247,-472.3266 2350,-456 2372.2088,-445.2962 2374.4872,-437.0416 2396,-425 2467.53,-384.9617 2515.7617,-396.9321 2549,-322 2554.5866,-309.4057 2554.1315,-303.7865 2549,-291 2528.8025,-240.6721 2405.5345,-129.5522 2357.8293,-87.7909"/>
-<polygon fill="#191970" stroke="#191970" points="2359.9304,-84.9793 2350.0923,-81.0469 2355.3308,-90.2561 2359.9304,-84.9793"/>
+<path fill="none" stroke="#191970" d="M2110.7464,-499.3223C2198.8399,-494.579 2356.9903,-482.5767 2407,-456 2492.8715,-410.3652 2592.6764,-238.1406 2620,-143 2632.4688,-99.5838 2566.1421,-82.0021 2524.2714,-75.3006"/>
+<polygon fill="#191970" stroke="#191970" points="2524.6188,-71.8142 2514.216,-73.8222 2523.6005,-78.7397 2524.6188,-71.8142"/>
</g>
<!-- Node4->Node15 -->
-<g id="edge129" class="edge">
+<g id="edge130" class="edge">
<title>Node4->Node15</title>
-<path fill="none" stroke="#191970" d="M2264.7506,-495.6938C2314.7305,-489.2845 2386.8671,-477.2015 2447,-456 2474.5656,-446.281 2478.5373,-437.4116 2505,-425 2686.5747,-339.8374 2811.0423,-421.3956 2923,-255 2958.8029,-201.7883 2910.6607,-124.052 2885.1048,-89.3633"/>
-<polygon fill="#191970" stroke="#191970" points="2887.7023,-86.996 2878.8708,-81.1432 2882.1248,-91.226 2887.7023,-86.996"/>
+<path fill="none" stroke="#191970" d="M2110.8601,-498.9535C2205.4366,-493.5321 2382.8666,-480.4169 2441,-456 2482.9649,-438.3741 2727.3288,-230.6911 2760,-199 2783.5296,-176.1762 2796.6269,-173.7444 2808,-143 2814.3252,-125.9012 2809.6514,-105.2624 2804.4001,-90.5631"/>
+<polygon fill="#191970" stroke="#191970" points="2807.5648,-89.049 2800.6232,-81.045 2801.0584,-91.6309 2807.5648,-89.049"/>
</g>
<!-- Node4->Node18 -->
-<g id="edge130" class="edge">
+<g id="edge131" class="edge">
<title>Node4->Node18</title>
-<path fill="none" stroke="#191970" d="M2143.3896,-499.9041C1974.219,-493.8372 1504.9246,-475.4502 1438,-456 1378.9708,-438.8444 1348.2112,-441.9587 1317,-389 1294.8854,-351.4763 1292.9223,-327.2953 1317,-291 1325.8111,-277.7179 1364.0011,-263.6334 1392.7092,-254.6012"/>
-<polygon fill="#191970" stroke="#191970" points="1393.8709,-257.9059 1402.4045,-251.6267 1391.8177,-251.2138 1393.8709,-257.9059"/>
+<path fill="none" stroke="#191970" d="M1989.4027,-494.3709C1937.8701,-487.0712 1862.2289,-474.4292 1798,-456 1762.1426,-445.7115 1755.6476,-435.9935 1720,-425 1649.9337,-403.392 1619.6994,-430.1312 1559,-389 1515.3216,-359.4026 1531.1027,-326.4229 1492,-291 1477.0639,-277.4694 1457.4402,-266.4393 1440.9164,-258.5752"/>
+<polygon fill="#191970" stroke="#191970" points="1442.1542,-255.2926 1431.6045,-254.3158 1439.2424,-261.6583 1442.1542,-255.2926"/>
</g>
<!-- Node4->Node22 -->
-<g id="edge92" class="edge">
+<g id="edge93" class="edge">
<title>Node4->Node22</title>
-<path fill="none" stroke="#191970" d="M2143.4946,-499.2904C2003.018,-492.8844 1643.8773,-475.844 1344,-456 1169.1731,-444.431 1125.8236,-436.6183 951,-425 914.0275,-422.5429 310.4842,-415.899 285,-389 191.4761,-290.2843 443.9522,-225.4948 567.4587,-200.5442"/>
-<polygon fill="#191970" stroke="#191970" points="568.4339,-203.9189 577.5586,-198.5345 567.0678,-197.0535 568.4339,-203.9189"/>
+<path fill="none" stroke="#191970" d="M1989.2753,-498.4239C1684.4784,-480.428 335.2323,-400.1932 317,-389 258.6264,-353.1632 252.9894,-271.4215 311,-235 344.1961,-214.1581 587.0063,-199.0738 712.8304,-192.5646"/>
+<polygon fill="#191970" stroke="#191970" points="713.3494,-196.0427 723.1574,-192.036 712.9915,-189.0518 713.3494,-196.0427"/>
</g>
-<!-- Node32 -->
-<g id="node33" class="node">
-<title>Node32</title>
-<g id="a_node33"><a xlink:href="data__type_8h.html" target="_top" xlink:title="tvm/runtime/data_type.h">
-<polygon fill="#ffffff" stroke="#000000" points="2402,-297 2402,-316 2540,-316 2540,-297 2402,-297"/>
-<text text-anchor="middle" x="2471" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/data_type.h</text>
+<!-- Node33 -->
+<g id="node34" class="node">
+<title>Node33</title>
+<g id="a_node34"><a xlink:href="data__type_8h.html" target="_top" xlink:title="tvm/runtime/data_type.h">
+<polygon fill="#ffffff" stroke="#000000" points="2326,-297 2326,-316 2464,-316 2464,-297 2326,-297"/>
+<text text-anchor="middle" x="2395" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/data_type.h</text>
</a>
</g>
</g>
-<!-- Node4->Node32 -->
-<g id="edge91" class="edge">
-<title>Node4->Node32</title>
-<path fill="none" stroke="#191970" d="M2231.439,-492.4288C2254.1712,-484.0551 2287.0061,-470.9199 2314,-456 2334.9365,-444.4281 2336.8706,-436.2159 2358,-425 2397.8817,-403.83 2423.5104,-423.1918 2453,-389 2467.9857,-371.6247 2471.3756,-344.6573 2471.7471,-326.3217"/>
-<polygon fill="#191970" stroke="#191970" points="2475.246,-326.1548 2471.6651,-316.1835 2468.2462,-326.2115 2475.246,-326.1548"/>
-</g>
<!-- Node4->Node33 -->
-<g id="edge66" class="edge">
+<g id="edge92" class="edge">
<title>Node4->Node33</title>
-<path fill="none" stroke="#191970" d="M2211.0313,-492.3906C2216.6139,-484.761 2224.6591,-473.7658 2231.9082,-463.8588"/>
-<polygon fill="#191970" stroke="#191970" points="2234.9112,-465.6817 2237.9918,-455.5446 2229.262,-461.5481 2234.9112,-465.6817"/>
+<path fill="none" stroke="#191970" d="M2110.5357,-497.3455C2190.274,-490.5292 2325.6928,-476.3887 2371,-456 2391.9915,-446.5536 2440.3266,-409.8879 2450,-389 2455.7899,-376.4978 2455.4809,-370.6407 2450,-358 2443.6351,-343.3206 2430.7971,-330.9121 2419.1244,-321.9554"/>
+<polygon fill="#191970" stroke="#191970" points="2421.0999,-319.065 2410.94,-316.0606 2417.0088,-324.7451 2421.0999,-319.065"/>
</g>
-<!-- Node34 -->
-<g id="node35" class="node">
-<title>Node34</title>
-<g id="a_node35"><a xlink:href="ndarray_8h.html" target="_top" xlink:title="A device-independent managed NDArray abstraction. ">
-<polygon fill="#ffffff" stroke="#000000" points="1643.5,-364 1643.5,-383 1768.5,-383 1768.5,-364 1643.5,-364"/>
-<text text-anchor="middle" x="1706" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/ndarray.h</text>
+<!-- Node4->Node34 -->
+<g id="edge67" class="edge">
+<title>Node4->Node34</title>
+<path fill="none" stroke="#191970" d="M2068.75,-492.3906C2085.5169,-483.7976 2110.6134,-470.9356 2131.6252,-460.1671"/>
+<polygon fill="#191970" stroke="#191970" points="2133.3417,-463.2203 2140.6447,-455.5446 2130.149,-456.9908 2133.3417,-463.2203"/>
+</g>
+<!-- Node35 -->
+<g id="node36" class="node">
+<title>Node35</title>
+<g id="a_node36"><a xlink:href="ndarray_8h.html" target="_top" xlink:title="A device-independent managed NDArray abstraction. ">
+<polygon fill="#ffffff" stroke="#000000" points="1567.5,-364 1567.5,-383 1692.5,-383 1692.5,-364 1567.5,-364"/>
+<text text-anchor="middle" x="1630" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/ndarray.h</text>
</a>
</g>
</g>
-<!-- Node4->Node34 -->
-<g id="edge93" class="edge">
-<title>Node4->Node34</title>
-<path fill="none" stroke="#191970" d="M2192.3493,-492.272C2168.5634,-472.497 2115.8276,-429.1204 2106,-425 2048.5464,-400.9114 1877.6875,-385.4272 1778.847,-378.2248"/>
-<polygon fill="#191970" stroke="#191970" points="1778.8572,-374.7166 1768.6321,-377.4912 1778.3557,-381.6986 1778.8572,-374.7166"/>
+<!-- Node4->Node35 -->
+<g id="edge94" class="edge">
+<title>Node4->Node35</title>
+<path fill="none" stroke="#191970" d="M2018.7749,-492.4466C1943.9661,-469.5587 1754.5031,-411.592 1670.8123,-385.9866"/>
+<polygon fill="#191970" stroke="#191970" points="1671.7317,-382.6078 1661.1453,-383.029 1669.6837,-389.3016 1671.7317,-382.6078"/>
</g>
-<!-- Node40 -->
-<g id="node41" class="node">
-<title>Node40</title>
-<g id="a_node41"><a xlink:href="packed__func_8h.html" target="_top" xlink:title="Type-erased function used across TVM API. ">
-<polygon fill="#ffffff" stroke="#000000" points="1447,-425.5 1447,-455.5 1563,-455.5 1563,-425.5 1447,-425.5"/>
-<text text-anchor="start" x="1455" y="-443.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/packed</text>
-<text text-anchor="middle" x="1505" y="-432.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_func.h</text>
+<!-- Node41 -->
+<g id="node42" class="node">
+<title>Node41</title>
+<g id="a_node42"><a xlink:href="packed__func_8h.html" target="_top" xlink:title="Type-erased function used across TVM API. ">
+<polygon fill="#ffffff" stroke="#000000" points="1301,-425.5 1301,-455.5 1417,-455.5 1417,-425.5 1301,-425.5"/>
+<text text-anchor="start" x="1309" y="-443.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/packed</text>
+<text text-anchor="middle" x="1359" y="-432.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">_func.h</text>
</a>
</g>
</g>
-<!-- Node4->Node40 -->
-<g id="edge95" class="edge">
-<title>Node4->Node40</title>
-<path fill="none" stroke="#191970" d="M2143.4952,-496.6766C2013.8781,-485.2725 1710.1178,-458.5468 1573.589,-446.5347"/>
-<polygon fill="#191970" stroke="#191970" points="1573.5333,-443.0163 1563.2649,-445.6263 1572.9197,-449.9894 1573.5333,-443.0163"/>
+<!-- Node4->Node41 -->
+<g id="edge96" class="edge">
+<title>Node4->Node41</title>
+<path fill="none" stroke="#191970" d="M1989.284,-496.5962C1860.8969,-485.1696 1562.6136,-458.6219 1427.5892,-446.6045"/>
+<polygon fill="#191970" stroke="#191970" points="1427.6454,-443.0958 1417.3745,-445.6954 1427.0248,-450.0682 1427.6454,-443.0958"/>
</g>
<!-- Node6 -->
<g id="node7" class="node">
<title>Node6</title>
<g id="a_node7"><a xlink:href="functor_8h.html" target="_top" xlink:title="Defines the Functor data structures. ">
-<polygon fill="#ffffff" stroke="#000000" points="2084.5,-297 2084.5,-316 2193.5,-316 2193.5,-297 2084.5,-297"/>
-<text text-anchor="middle" x="2139" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/functor.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="2122.5,-297 2122.5,-316 2231.5,-316 2231.5,-297 2122.5,-297"/>
+<text text-anchor="middle" x="2177" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/functor.h</text>
</a>
</g>
</g>
<!-- Node5->Node6 -->
<g id="edge6" class="edge">
<title>Node5->Node6</title>
-<path fill="none" stroke="#191970" d="M2052.0513,-425.389C2070.2313,-400.5307 2106.3686,-351.1184 2125.8404,-324.4937"/>
-<polygon fill="#191970" stroke="#191970" points="2128.8979,-326.242 2131.9761,-316.1042 2123.2477,-322.1098 2128.8979,-326.242"/>
+<path fill="none" stroke="#191970" d="M1710.7187,-437.1869C1808.9208,-430.8388 2006.2247,-415.215 2070,-389 2089.1997,-381.1079 2132.997,-344.6936 2158.2021,-322.9686"/>
+<polygon fill="#191970" stroke="#191970" points="2160.714,-325.423 2165.9748,-316.2266 2156.1272,-320.1351 2160.714,-325.423"/>
</g>
<!-- Node5->Node14 -->
-<g id="edge65" class="edge">
+<g id="edge66" class="edge">
<title>Node5->Node14</title>
-<path fill="none" stroke="#191970" d="M2068.3358,-425.463C2102.8498,-405.4285 2162.4824,-367.2038 2202,-322 2267.704,-246.8418 2314.818,-134.9045 2331.9276,-90.571"/>
-<polygon fill="#191970" stroke="#191970" points="2335.2679,-91.6334 2335.5415,-81.0421 2328.7228,-89.1511 2335.2679,-91.6334"/>
+<path fill="none" stroke="#191970" d="M1710.8585,-436.4513C1815.0185,-428.6305 2032.4121,-410.3235 2105,-389 2169.6496,-370.0085 2186.2019,-361.2098 2241,-322 2343.1132,-248.9347 2441.8321,-133.2975 2477.7544,-89.2945"/>
+<polygon fill="#191970" stroke="#191970" points="2480.6902,-91.2308 2484.2653,-81.2573 2475.251,-86.8245 2480.6902,-91.2308"/>
</g>
<!-- Node19 -->
<g id="node20" class="node">
<title>Node19</title>
<g id="a_node20"><a xlink:href="object__path_8h.html" target="_top" xlink:title="tvm/node/object_path.h">
-<polygon fill="#ffffff" stroke="#000000" points="1024,-364 1024,-383 1156,-383 1156,-364 1024,-364"/>
-<text text-anchor="middle" x="1090" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/object_path.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="1030,-364 1030,-383 1162,-383 1162,-364 1030,-364"/>
+<text text-anchor="middle" x="1096" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/object_path.h</text>
</a>
</g>
</g>
<!-- Node5->Node19 -->
<g id="edge21" class="edge">
<title>Node5->Node19</title>
-<path fill="none" stroke="#191970" d="M1984.3586,-436.5095C1821.4922,-425.0352 1351.7699,-391.9423 1166.471,-378.8875"/>
-<polygon fill="#191970" stroke="#191970" points="1166.4918,-375.3804 1156.2705,-378.1689 1165.9998,-382.3631 1166.4918,-375.3804"/>
+<path fill="none" stroke="#191970" d="M1597.2529,-436.2552C1506.7531,-429.0804 1324.3058,-412.9562 1171,-389 1163.7749,-387.871 1156.1925,-386.5116 1148.7398,-385.073"/>
+<polygon fill="#191970" stroke="#191970" points="1149.1939,-381.5949 1138.7032,-383.0764 1147.8281,-388.4603 1149.1939,-381.5949"/>
</g>
-<!-- Node31 -->
-<g id="node32" class="node">
-<title>Node31</title>
-<g id="a_node32"><a xlink:href="array_8h.html" target="_top" xlink:title="Runtime Array container types. ">
-<polygon fill="#ffffff" stroke="#000000" points="552,-291.5 552,-321.5 678,-321.5 678,-291.5 552,-291.5"/>
-<text text-anchor="start" x="560" y="-309.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
-<text text-anchor="middle" x="615" y="-298.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/array.h</text>
+<!-- Node32 -->
+<g id="node33" class="node">
+<title>Node32</title>
+<g id="a_node33"><a xlink:href="array_8h.html" target="_top" xlink:title="Runtime Array container types. ">
+<polygon fill="#ffffff" stroke="#000000" points="294,-291.5 294,-321.5 420,-321.5 420,-291.5 294,-291.5"/>
+<text text-anchor="start" x="302" y="-309.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
+<text text-anchor="middle" x="357" y="-298.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/array.h</text>
</a>
</g>
</g>
-<!-- Node5->Node31 -->
-<g id="edge54" class="edge">
-<title>Node5->Node31</title>
-<path fill="none" stroke="#191970" d="M1984.186,-438.2491C1782.9383,-430.1931 1110.9126,-402.5564 1015,-389 955.3303,-380.5662 941.9579,-370.4729 883,-358 798.487,-340.1207 776.7885,-338.5238 692,-322 690.7201,-321.7506 689.4284,-321.4982 688.1276,-321.2433"/>
-<polygon fill="#191970" stroke="#191970" points="688.5076,-317.751 678.0196,-319.2516 687.1542,-324.6189 688.5076,-317.751"/>
-</g>
<!-- Node5->Node32 -->
-<g id="edge60" class="edge">
+<g id="edge55" class="edge">
<title>Node5->Node32</title>
-<path fill="none" stroke="#191970" d="M2089.4906,-425.389C2175.1124,-398.7068 2351.514,-343.7352 2430.8243,-319.0199"/>
-<polygon fill="#191970" stroke="#191970" points="2431.9472,-322.336 2440.4531,-316.0193 2429.8646,-315.653 2431.9472,-322.336"/>
+<path fill="none" stroke="#191970" d="M1597.3993,-436.4538C1452.7538,-426.0343 1076.9934,-398.4384 1021,-389 966.6821,-379.844 955.3273,-367.1003 901,-358 701.3702,-324.5603 647.8359,-347.2033 447,-322 441.5922,-321.3214 436.014,-320.5455 430.4086,-319.7108"/>
+<polygon fill="#191970" stroke="#191970" points="430.7376,-316.2202 420.3205,-318.153 429.6692,-323.1382 430.7376,-316.2202"/>
+</g>
+<!-- Node5->Node33 -->
+<g id="edge61" class="edge">
+<title>Node5->Node33</title>
+<path fill="none" stroke="#191970" d="M1710.9554,-439.544C1835.948,-436.7707 2130.435,-426.3558 2222,-389 2242.1811,-380.7667 2241.1326,-368.9125 2260,-358 2288.9126,-341.2776 2324.2693,-328.0187 2351.5455,-319.1441"/>
+<polygon fill="#191970" stroke="#191970" points="2352.7322,-322.4395 2361.2021,-316.0746 2350.6117,-315.7684 2352.7322,-322.4395"/>
</g>
<!-- Node7 -->
<g id="node8" class="node">
<title>Node7</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1104.5,-179.5 1104.5,-198.5 1193.5,-198.5 1193.5,-179.5 1104.5,-179.5"/>
-<text text-anchor="middle" x="1149" y="-186.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">dmlc/logging.h</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1086.5,-179.5 1086.5,-198.5 1175.5,-198.5 1175.5,-179.5 1086.5,-179.5"/>
+<text text-anchor="middle" x="1131" y="-186.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">dmlc/logging.h</text>
</g>
<!-- Node6->Node7 -->
<g id="edge7" class="edge">
<title>Node6->Node7</title>
-<path fill="none" stroke="#191970" d="M2109.1362,-296.9215C2058.3535,-281.0767 1951.6606,-249.7023 1859,-235 1732.9899,-215.0062 1348.0141,-197.3265 1204.0743,-191.2456"/>
-<polygon fill="#191970" stroke="#191970" points="1203.8668,-187.7339 1193.7287,-190.811 1203.5729,-194.7277 1203.8668,-187.7339"/>
+<path fill="none" stroke="#191970" d="M2128.9754,-296.9515C2050.3069,-281.6463 1888.9222,-251.6549 1751,-235 1543.1252,-209.8979 1294.2731,-196.4249 1185.9162,-191.3708"/>
+<polygon fill="#191970" stroke="#191970" points="1186.0253,-187.8722 1175.8747,-190.9079 1185.7029,-194.8648 1186.0253,-187.8722"/>
</g>
<!-- Node6->Node8 -->
<g id="edge8" class="edge">
<title>Node6->Node8</title>
-<path fill="none" stroke="#191970" d="M2134.4195,-296.9015C2126.4614,-281.316 2108.5528,-250.6739 2084,-235 2017.1362,-192.3158 1792.6751,-157.4788 1677.0916,-141.8634"/>
-<polygon fill="#191970" stroke="#191970" points="1677.3073,-138.3612 1666.9311,-140.5024 1676.3779,-145.2992 1677.3073,-138.3612"/>
+<path fill="none" stroke="#191970" d="M2161.0551,-296.8737C2119.4945,-272.3139 2003.725,-207.2744 1898,-179 1820.9927,-158.4057 1595.3433,-143.3183 1479.8521,-136.7071"/>
+<polygon fill="#191970" stroke="#191970" points="1479.8854,-133.2035 1469.7033,-136.1318 1479.4891,-140.1923 1479.8854,-133.2035"/>
</g>
<!-- Node6->Node15 -->
<g id="edge18" class="edge">
<title>Node6->Node15</title>
-<path fill="none" stroke="#191970" d="M2181.0287,-296.9399C2254.9535,-279.7657 2412.764,-241.4955 2543,-199 2653.9087,-162.8109 2781.855,-109.6707 2839.7141,-85.0024"/>
-<polygon fill="#191970" stroke="#191970" points="2841.2172,-88.1662 2849.0352,-81.0158 2838.4645,-81.7302 2841.2172,-88.1662"/>
+<path fill="none" stroke="#191970" d="M2202.0322,-296.9967C2298.8561,-260.238 2648.9872,-127.3126 2761.1633,-84.7256"/>
+<polygon fill="#191970" stroke="#191970" points="2762.5496,-87.9431 2770.6563,-81.1216 2760.0651,-81.3988 2762.5496,-87.9431"/>
</g>
<!-- Node6->Node16 -->
<g id="edge19" class="edge">
<title>Node6->Node16</title>
-<path fill="none" stroke="#191970" d="M2115.1431,-296.9058C2075.3764,-281.3264 1992.7183,-250.6928 1920,-235 1346.1088,-111.1527 626.2663,-78.5393 465.7194,-72.6042"/>
-<polygon fill="#191970" stroke="#191970" points="465.6671,-69.1002 455.5471,-72.2366 465.4142,-76.0956 465.6671,-69.1002"/>
+<path fill="none" stroke="#191970" d="M2136.3904,-296.9671C2072.5261,-282.2117 1944.5887,-253.6066 1835,-235 1354.7051,-153.4528 764.4616,-89.9392 620.0306,-74.8956"/>
+<polygon fill="#191970" stroke="#191970" points="619.9679,-71.3704 609.66,-73.8191 619.2451,-78.333 619.9679,-71.3704"/>
</g>
<!-- Node6->Node18 -->
<g id="edge20" class="edge">
<title>Node6->Node18</title>
-<path fill="none" stroke="#191970" d="M2084.3086,-297.4994C2068.7311,-295.1453 2051.7232,-292.7734 2036,-291 1816.9815,-266.2968 1552.18,-251.4152 1459.7637,-246.6657"/>
-<polygon fill="#191970" stroke="#191970" points="1459.7755,-243.1619 1449.6106,-246.1493 1459.4198,-250.1529 1459.7755,-243.1619"/>
+<path fill="none" stroke="#191970" d="M2122.4817,-301.5777C2087.2587,-298.4407 2040.4311,-294.351 1999,-291 1784.8803,-273.6819 1731.079,-272.8135 1517,-255 1491.7745,-252.901 1463.2895,-250.2789 1441.7664,-248.2464"/>
+<polygon fill="#191970" stroke="#191970" points="1442.0907,-244.7616 1431.8046,-247.3009 1441.4291,-251.7302 1442.0907,-244.7616"/>
</g>
<!-- Node8->Node9 -->
<g id="edge9" class="edge">
<title>Node8->Node9</title>
-<path fill="none" stroke="#191970" d="M1666.8832,-129.9337C1795.6196,-123.2223 2108.2835,-106.2807 2370,-87 2403.9165,-84.5014 2441.4293,-81.3 2473.2375,-78.4562"/>
-<polygon fill="#191970" stroke="#191970" points="2473.7062,-81.9282 2483.3526,-77.5468 2473.0794,-74.9564 2473.7062,-81.9282"/>
+<path fill="none" stroke="#191970" d="M1469.7421,-126.6213C1575.4789,-115.3317 1795.3091,-91.8602 1911.0367,-79.5039"/>
+<polygon fill="#191970" stroke="#191970" points="1911.6527,-82.9581 1921.2246,-78.4161 1910.9095,-75.9977 1911.6527,-82.9581"/>
</g>
<!-- Node13 -->
<g id="node14" class="node">
<title>Node13</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="867.5,-62 867.5,-81 992.5,-81 992.5,-62 867.5,-62"/>
-<text text-anchor="middle" x="930" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/logging.h</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="894.5,-62 894.5,-81 1019.5,-81 1019.5,-62 894.5,-62"/>
+<text text-anchor="middle" x="957" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/logging.h</text>
</g>
<!-- Node8->Node13 -->
<g id="edge13" class="edge">
<title>Node8->Node13</title>
-<path fill="none" stroke="#191970" d="M1547.2177,-127.5693C1423.1357,-116.2974 1137.9969,-90.3948 1003.1788,-78.1477"/>
-<polygon fill="#191970" stroke="#191970" points="1003.2291,-74.638 992.9535,-77.2188 1002.5958,-81.6092 1003.2291,-74.638"/>
+<path fill="none" stroke="#191970" d="M1350.3123,-124.8967C1267.7252,-113.6845 1118.743,-93.4585 1029.7471,-81.3763"/>
+<polygon fill="#191970" stroke="#191970" points="1030.1121,-77.8938 1019.7321,-80.0166 1029.1703,-84.8302 1030.1121,-77.8938"/>
</g>
<!-- Node8->Node14 -->
<g id="edge14" class="edge">
<title>Node8->Node14</title>
-<path fill="none" stroke="#191970" d="M1666.5923,-127.9933C1814.3805,-115.5766 2193.9624,-83.6855 2306.8618,-74.2001"/>
-<polygon fill="#191970" stroke="#191970" points="2307.2665,-77.6786 2316.9383,-73.3535 2306.6804,-70.7031 2307.2665,-77.6786"/>
+<path fill="none" stroke="#191970" d="M1469.7142,-129.6059C1669.1625,-118.2694 2308.243,-81.9446 2459.6265,-73.3401"/>
+<polygon fill="#191970" stroke="#191970" points="2459.8853,-76.8311 2469.6705,-72.7692 2459.488,-69.8424 2459.8853,-76.8311"/>
</g>
<!-- Node8->Node15 -->
<g id="edge15" class="edge">
<title>Node8->Node15</title>
-<path fill="none" stroke="#191970" d="M1666.5078,-130.6544C1823.0263,-124.4088 2259.1483,-106.5144 2622,-87 2693.4695,-83.1563 2776.4214,-77.8165 2826.2536,-74.5136"/>
-<polygon fill="#191970" stroke="#191970" points="2826.6228,-77.9969 2836.3685,-73.841 2826.1583,-71.0123 2826.6228,-77.9969"/>
+<path fill="none" stroke="#191970" d="M1469.7368,-130.8552C1636.4171,-124.8008 2120.7403,-106.7309 2523,-87 2603.2353,-83.0644 2696.6811,-77.5593 2750.8153,-74.2791"/>
+<polygon fill="#191970" stroke="#191970" points="2751.3331,-77.7542 2761.1024,-73.6539 2750.9084,-70.7671 2751.3331,-77.7542"/>
</g>
<!-- Node8->Node16 -->
<g id="edge16" class="edge">
<title>Node8->Node16</title>
-<path fill="none" stroke="#191970" d="M1547.1671,-129.8657C1335.6839,-118.7871 627.4528,-81.6864 465.9691,-73.2271"/>
-<polygon fill="#191970" stroke="#191970" points="465.8859,-69.718 455.7165,-72.69 465.5197,-76.7084 465.8859,-69.718"/>
+<path fill="none" stroke="#191970" d="M1350.4808,-128.5523C1188.6729,-116.461 744.5357,-83.2721 619.9063,-73.959"/>
+<polygon fill="#191970" stroke="#191970" points="620.0712,-70.4616 609.8382,-73.2066 619.5495,-77.4422 620.0712,-70.4616"/>
</g>
<!-- Node17 -->
<g id="node18" class="node">
<title>Node17</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1641,-62 1641,-81 1691,-81 1691,-62 1641,-62"/>
-<text text-anchor="middle" x="1666" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">atomic</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1429,-62 1429,-81 1479,-81 1479,-62 1429,-62"/>
+<text text-anchor="middle" x="1454" y="-69" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">atomic</text>
</g>
<!-- Node8->Node17 -->
<g id="edge17" class="edge">
<title>Node8->Node17</title>
-<path fill="none" stroke="#191970" d="M1616.2188,-123.3906C1625.0858,-114.1478 1638.6912,-99.966 1649.4389,-88.7628"/>
-<polygon fill="#191970" stroke="#191970" points="1652.2541,-90.8841 1656.6513,-81.2449 1647.2027,-86.0381 1652.2541,-90.8841"/>
+<path fill="none" stroke="#191970" d="M1416.875,-123.3906C1423.3593,-114.3273 1433.2414,-100.5149 1441.1801,-89.4188"/>
+<polygon fill="#191970" stroke="#191970" points="1444.0559,-91.4143 1447.0281,-81.2449 1438.3629,-87.3412 1444.0559,-91.4143"/>
</g>
<!-- Node10 -->
<g id="node11" class="node">
<title>Node10</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="2408.5,-.5 2408.5,-19.5 2501.5,-19.5 2501.5,-.5 2408.5,-.5"/>
-<text text-anchor="middle" x="2455" y="-7.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">dlpack/dlpack.h</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1846.5,-.5 1846.5,-19.5 1939.5,-19.5 1939.5,-.5 1846.5,-.5"/>
+<text text-anchor="middle" x="1893" y="-7.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">dlpack/dlpack.h</text>
</g>
<!-- Node9->Node10 -->
<g id="edge10" class="edge">
<title>Node9->Node10</title>
-<path fill="none" stroke="#191970" d="M2525.0112,-56.2977C2510.8034,-46.9022 2492.6215,-34.8787 2478.3475,-25.4395"/>
-<polygon fill="#191970" stroke="#191970" points="2479.8845,-22.2598 2469.6128,-19.6633 2476.0233,-28.0986 2479.8845,-22.2598"/>
+<path fill="none" stroke="#191970" d="M1963.0112,-56.2977C1948.8034,-46.9022 1930.6215,-34.8787 1916.3475,-25.4395"/>
+<polygon fill="#191970" stroke="#191970" points="1917.8845,-22.2598 1907.6128,-19.6633 1914.0233,-28.0986 1917.8845,-22.2598"/>
</g>
<!-- Node11 -->
<g id="node12" class="node">
<title>Node11</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="2520,-.5 2520,-19.5 2576,-19.5 2576,-.5 2520,-.5"/>
-<text text-anchor="middle" x="2548" y="-7.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">stddef.h</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1958,-.5 1958,-19.5 2014,-19.5 2014,-.5 1958,-.5"/>
+<text text-anchor="middle" x="1986" y="-7.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">stddef.h</text>
</g>
<!-- Node9->Node11 -->
<g id="edge11" class="edge">
<title>Node9->Node11</title>
-<path fill="none" stroke="#191970" d="M2548,-56.2977C2548,-48.3834 2548,-38.6043 2548,-30.0759"/>
-<polygon fill="#191970" stroke="#191970" points="2551.5001,-29.8469 2548,-19.8469 2544.5001,-29.847 2551.5001,-29.8469"/>
+<path fill="none" stroke="#191970" d="M1986,-56.2977C1986,-48.3834 1986,-38.6043 1986,-30.0759"/>
+<polygon fill="#191970" stroke="#191970" points="1989.5001,-29.8469 1986,-19.8469 1982.5001,-29.847 1989.5001,-29.8469"/>
</g>
<!-- Node12 -->
<g id="node13" class="node">
<title>Node12</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="2594.5,-.5 2594.5,-19.5 2647.5,-19.5 2647.5,-.5 2594.5,-.5"/>
-<text text-anchor="middle" x="2621" y="-7.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">stdint.h</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="2032.5,-.5 2032.5,-19.5 2085.5,-19.5 2085.5,-.5 2032.5,-.5"/>
+<text text-anchor="middle" x="2059" y="-7.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">stdint.h</text>
</g>
<!-- Node9->Node12 -->
<g id="edge12" class="edge">
<title>Node9->Node12</title>
-<path fill="none" stroke="#191970" d="M2566.0449,-56.2977C2576.8114,-47.2274 2590.4851,-35.7077 2601.4995,-26.4285"/>
-<polygon fill="#191970" stroke="#191970" points="2603.9191,-28.9667 2609.3118,-19.8469 2599.4089,-23.6132 2603.9191,-28.9667"/>
+<path fill="none" stroke="#191970" d="M2004.0449,-56.2977C2014.8114,-47.2274 2028.4851,-35.7077 2039.4995,-26.4285"/>
+<polygon fill="#191970" stroke="#191970" points="2041.9191,-28.9667 2047.3118,-19.8469 2037.4089,-23.6132 2041.9191,-28.9667"/>
</g>
<!-- Node19->Node8 -->
-<g id="edge52" class="edge">
+<g id="edge53" class="edge">
<title>Node19->Node8</title>
-<path fill="none" stroke="#191970" d="M1105.991,-363.9877C1133.8032,-347.6857 1193.1719,-313.9514 1246,-291 1290.8622,-271.5095 1304.7269,-273.516 1350,-255 1361.4685,-250.3096 1512.911,-177.9905 1577.8069,-146.9637"/>
-<polygon fill="#191970" stroke="#191970" points="1579.5745,-149.9982 1587.0863,-142.5267 1576.5547,-143.683 1579.5745,-149.9982"/>
+<path fill="none" stroke="#191970" d="M1105.9669,-363.9267C1123.0013,-347.8233 1159.1616,-314.7496 1193,-291 1219.3964,-272.4736 1230.9992,-275.3706 1256,-255 1290.9087,-226.5564 1287.6299,-206.8677 1323,-179 1339.3792,-166.095 1359.9149,-155.0396 1376.9212,-147.0034"/>
+<polygon fill="#191970" stroke="#191970" points="1378.8347,-149.9766 1386.4682,-142.6295 1375.9191,-143.6127 1378.8347,-149.9766"/>
</g>
<!-- Node19->Node14 -->
-<g id="edge53" class="edge">
+<g id="edge54" class="edge">
<title>Node19->Node14</title>
-<path fill="none" stroke="#191970" d="M1097.6626,-363.7665C1131.5281,-321.0407 1268.791,-150.9902 1330,-123 1375.4117,-102.2337 2138.6169,-77.656 2306.5085,-72.4844"/>
-<polygon fill="#191970" stroke="#191970" points="2306.8044,-75.9771 2316.6924,-72.1721 2306.5898,-68.9804 2306.8044,-75.9771"/>
+<path fill="none" stroke="#191970" d="M1136.6864,-363.9593C1147.7617,-361.6726 1159.7944,-359.47 1171,-358 1449.0743,-321.5211 1529.5943,-392.6751 1801,-322 1822.4,-316.4274 1959.9626,-244.355 1980,-235 2035.1603,-209.2469 2048.4426,-201.5197 2105,-179 2174.9962,-151.1293 2192.5909,-143.8126 2265,-123 2333.6762,-103.2604 2415.7807,-86.2195 2459.9981,-77.5669"/>
+<polygon fill="#191970" stroke="#191970" points="2460.7254,-80.9911 2469.8757,-75.6506 2459.3921,-74.1193 2460.7254,-80.9911"/>
</g>
<!-- Node20 -->
<g id="node21" class="node">
<title>Node20</title>
<g id="a_node21"><a xlink:href="optional_8h.html" target="_top" xlink:title="Runtime Optional container types. ">
-<polygon fill="#ffffff" stroke="#000000" points="332,-291.5 332,-321.5 458,-321.5 458,-291.5 332,-291.5"/>
-<text text-anchor="start" x="340" y="-309.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
-<text text-anchor="middle" x="395" y="-298.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/optional.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="552,-291.5 552,-321.5 678,-321.5 678,-291.5 552,-291.5"/>
+<text text-anchor="start" x="560" y="-309.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
+<text text-anchor="middle" x="615" y="-298.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/optional.h</text>
</a>
</g>
</g>
<!-- Node19->Node20 -->
<g id="edge22" class="edge">
<title>Node19->Node20</title>
-<path fill="none" stroke="#191970" d="M1023.9057,-368.9639C887.1492,-359.4185 577.131,-336.911 472,-322 470.8862,-321.842 469.7641,-321.6781 468.6356,-321.5089"/>
-<polygon fill="#191970" stroke="#191970" points="468.7779,-317.9869 458.3491,-319.8552 467.6668,-324.8981 468.7779,-317.9869"/>
+<path fill="none" stroke="#191970" d="M1029.7592,-365.6431C932.4143,-353.9965 755.4057,-332.4093 692,-322 690.8899,-321.8178 689.7713,-321.6312 688.6461,-321.4407"/>
+<polygon fill="#191970" stroke="#191970" points="688.8383,-317.9208 678.383,-319.6342 687.6248,-324.8148 688.8383,-317.9208"/>
</g>
<!-- Node19->Node26 -->
<g id="edge36" class="edge">
<title>Node19->Node26</title>
-<path fill="none" stroke="#191970" d="M1084.9854,-363.9005C1080.4666,-355.2503 1073.6304,-342.164 1067.6261,-330.6699"/>
-<polygon fill="#191970" stroke="#191970" points="1070.5796,-328.7647 1062.8472,-321.5218 1064.3752,-332.0058 1070.5796,-328.7647"/>
+<path fill="none" stroke="#191970" d="M1084.6812,-363.9005C1073.7678,-354.6448 1056.8657,-340.3101 1042.6828,-328.2816"/>
+<polygon fill="#191970" stroke="#191970" points="1044.6026,-325.3206 1034.7122,-321.5218 1040.0749,-330.6592 1044.6026,-325.3206"/>
</g>
<!-- Node20->Node16 -->
<g id="edge23" class="edge">
<title>Node20->Node16</title>
-<path fill="none" stroke="#191970" d="M359.9243,-291.3088C343.6785,-282.7244 325.2959,-270.5664 313,-255 275.1586,-207.0935 239.1391,-172.4066 275,-123 289.8812,-102.4978 358.7004,-85.8888 400.4511,-77.4798"/>
-<polygon fill="#191970" stroke="#191970" points="401.2941,-80.8812 410.4331,-75.5213 399.9463,-74.0122 401.2941,-80.8812"/>
+<path fill="none" stroke="#191970" d="M551.7699,-297.7116C502.7125,-289.4455 440.5226,-275.2901 425,-255 374.1241,-188.4983 497.5251,-115.6639 557.0746,-85.6381"/>
+<polygon fill="#191970" stroke="#191970" points="558.9754,-88.6026 566.3853,-81.03 555.8704,-82.329 558.9754,-88.6026"/>
</g>
<!-- Node21 -->
<g id="node22" class="node">
<title>Node21</title>
<g id="a_node22"><a xlink:href="runtime_2container_2base_8h.html" target="_top" xlink:title="Base utilities for common POD(plain old data) container types. ">
-<polygon fill="#ffffff" stroke="#000000" points="659.5,-235.5 659.5,-254.5 714.5,-254.5 714.5,-235.5 659.5,-235.5"/>
-<text text-anchor="middle" x="687" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">./base.h</text>
+<polygon fill="#ffffff" stroke="#000000" points="736.5,-235.5 736.5,-254.5 791.5,-254.5 791.5,-235.5 736.5,-235.5"/>
+<text text-anchor="middle" x="764" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">./base.h</text>
</a>
</g>
</g>
<!-- Node20->Node21 -->
<g id="edge24" class="edge">
<title>Node20->Node21</title>
-<path fill="none" stroke="#191970" d="M458.0002,-293.2311C516.261,-280.9604 600.9265,-263.1285 649.3161,-252.9368"/>
-<polygon fill="#191970" stroke="#191970" points="650.2933,-256.3079 659.3572,-250.822 648.8505,-249.4582 650.2933,-256.3079"/>
+<path fill="none" stroke="#191970" d="M651.4495,-291.4554C675.8744,-281.374 707.7581,-268.2139 731.3521,-258.4755"/>
+<polygon fill="#191970" stroke="#191970" points="733.0021,-261.5809 740.9103,-254.5303 730.3313,-255.1104 733.0021,-261.5809"/>
</g>
<!-- Node21->Node7 -->
<g id="edge25" class="edge">
<title>Node21->Node7</title>
-<path fill="none" stroke="#191970" d="M714.7989,-236.7794C717.5553,-236.1147 720.323,-235.5058 723,-235 876.6581,-205.9705 917.6867,-217.2018 1073,-199 1079.8381,-198.1986 1087.0162,-197.3127 1094.1176,-196.4092"/>
-<polygon fill="#191970" stroke="#191970" points="1094.9485,-199.8312 1104.4186,-195.0807 1094.053,-192.8887 1094.9485,-199.8312"/>
+<path fill="none" stroke="#191970" d="M791.7822,-241.3199C845.0084,-234.1858 965.2632,-217.6648 1076.4503,-199.211"/>
+<polygon fill="#191970" stroke="#191970" points="1077.2034,-202.6338 1086.4906,-197.5351 1076.0508,-195.7294 1077.2034,-202.6338"/>
</g>
<!-- Node21->Node8 -->
<g id="edge32" class="edge">
<title>Node21->Node8</title>
-<path fill="none" stroke="#191970" d="M714.6274,-237.0278C765.2049,-222.7875 876.2879,-193.1491 972,-179 1175.7892,-148.8739 1419.4153,-138.232 1536.8605,-134.6717"/>
-<polygon fill="#191970" stroke="#191970" points="1537.2525,-138.1618 1547.145,-134.3684 1537.0461,-131.1648 1537.2525,-138.1618"/>
+<path fill="none" stroke="#191970" d="M789.155,-235.4475C828.5832,-220.8782 907.6109,-193.2941 977,-179 1103.4685,-152.9475 1253.9343,-141.1822 1340.0743,-136.2367"/>
+<polygon fill="#191970" stroke="#191970" points="1340.4656,-139.7204 1350.2543,-135.6669 1340.0743,-132.7313 1340.4656,-139.7204"/>
</g>
<!-- Node21->Node13 -->
<g id="edge26" class="edge">
<title>Node21->Node13</title>
-<path fill="none" stroke="#191970" d="M714.6522,-237.2861C748.2352,-227.6022 802.56,-210.8284 820,-199 864.1521,-169.0546 901.2522,-117.0151 918.9192,-89.5961"/>
-<polygon fill="#191970" stroke="#191970" points="921.9655,-91.327 924.3375,-81.0011 916.0439,-87.594 921.9655,-91.327"/>
+<path fill="none" stroke="#191970" d="M791.8216,-236.1182C812.6474,-228.531 841.0346,-216.1158 862,-199 900.6605,-167.4383 932.0973,-116.8633 947.231,-89.8729"/>
+<polygon fill="#191970" stroke="#191970" points="950.3158,-91.5267 952.0521,-81.0751 944.177,-88.1627 950.3158,-91.5267"/>
</g>
<!-- Node21->Node16 -->
<g id="edge35" class="edge">
<title>Node21->Node16</title>
-<path fill="none" stroke="#191970" d="M714.6638,-236.0487C741.2882,-226.8074 778.7008,-211.9527 787,-199 791.7954,-191.5156 791.81,-186.475 787,-179 751.4502,-123.7538 546.6736,-88.1578 465.8228,-76.0843"/>
-<polygon fill="#191970" stroke="#191970" points="466.1107,-72.5892 455.7086,-74.6007 465.0947,-79.515 466.1107,-72.5892"/>
+<path fill="none" stroke="#191970" d="M751.468,-235.3456C740.684,-226.6262 725.1471,-213.0423 714,-199 689.8706,-168.6035 697.9565,-150.9199 671,-123 655.8621,-107.3211 634.9863,-94.4926 617.8333,-85.5607"/>
+<polygon fill="#191970" stroke="#191970" points="619.2551,-82.3583 608.7469,-81.006 616.1182,-88.6161 619.2551,-82.3583"/>
</g>
<!-- Node21->Node22 -->
<g id="edge27" class="edge">
<title>Node21->Node22</title>
-<path fill="none" stroke="#191970" d="M677.2455,-235.2455C669.1352,-227.1352 657.4075,-215.4075 647.7919,-205.7919"/>
-<polygon fill="#191970" stroke="#191970" points="650.1887,-203.2389 640.6427,-198.6427 645.2389,-208.1887 650.1887,-203.2389"/>
+<path fill="none" stroke="#191970" d="M768.1805,-235.2455C771.381,-227.7776 775.896,-217.2427 779.8084,-208.1137"/>
+<polygon fill="#191970" stroke="#191970" points="783.1452,-209.2129 783.8674,-198.6427 776.7112,-206.4554 783.1452,-209.2129"/>
</g>
<!-- Node21->Node24 -->
<g id="edge33" class="edge">
<title>Node21->Node24</title>
-<path fill="none" stroke="#191970" d="M697.277,-235.2455C705.8218,-227.1352 718.1778,-215.4075 728.3085,-205.7919"/>
-<polygon fill="#191970" stroke="#191970" points="730.9972,-208.0655 735.8408,-198.6427 726.1782,-202.9884 730.9972,-208.0655"/>
+<path fill="none" stroke="#191970" d="M736.218,-241.9494C651.7993,-232.6799 397.8966,-204.8004 296.5011,-193.6668"/>
+<polygon fill="#191970" stroke="#191970" points="296.6457,-190.1617 286.3234,-192.5492 295.8816,-197.1199 296.6457,-190.1617"/>
</g>
<!-- Node25 -->
<g id="node26" class="node">
<title>Node25</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="981.5,-179.5 981.5,-198.5 1064.5,-198.5 1064.5,-179.5 981.5,-179.5"/>
-<text text-anchor="middle" x="1023" y="-186.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">initializer_list</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="985.5,-179.5 985.5,-198.5 1068.5,-198.5 1068.5,-179.5 985.5,-179.5"/>
+<text text-anchor="middle" x="1027" y="-186.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">initializer_list</text>
</g>
<!-- Node21->Node25 -->
<g id="edge34" class="edge">
<title>Node21->Node25</title>
-<path fill="none" stroke="#191970" d="M714.8269,-236.9184C717.5777,-236.2256 720.3363,-235.572 723,-235 809.4426,-216.4355 911.4501,-202.4451 971.4827,-195.0081"/>
-<polygon fill="#191970" stroke="#191970" points="971.9693,-198.4747 981.469,-193.7836 971.1173,-191.5268 971.9693,-198.4747"/>
+<path fill="none" stroke="#191970" d="M791.6485,-239.1129C835.103,-229.8602 920.0973,-211.7625 975.17,-200.036"/>
+<polygon fill="#191970" stroke="#191970" points="976.0366,-203.4301 985.0884,-197.9242 974.5787,-196.5836 976.0366,-203.4301"/>
</g>
<!-- Node22->Node8 -->
<g id="edge28" class="edge">
<title>Node22->Node8</title>
-<path fill="none" stroke="#191970" d="M695.5377,-179.9563C698.7345,-179.6096 701.902,-179.2878 705,-179 866.2761,-164.0169 1354.6891,-143.2587 1537.2011,-135.8044"/>
-<polygon fill="#191970" stroke="#191970" points="1537.3711,-139.3005 1547.2203,-135.396 1537.086,-132.3063 1537.3711,-139.3005"/>
+<path fill="none" stroke="#191970" d="M852.5132,-183.1917C970.1308,-172.6024 1218.4813,-150.2428 1340.0219,-139.3003"/>
+<polygon fill="#191970" stroke="#191970" points="1340.6737,-142.7558 1350.3195,-138.3732 1340.0459,-135.784 1340.6737,-142.7558"/>
</g>
<!-- Node22->Node15 -->
<g id="edge30" class="edge">
<title>Node22->Node15</title>
-<path fill="none" stroke="#191970" d="M695.5437,-180.018C698.7389,-179.6549 701.9044,-179.3129 705,-179 1074.1772,-141.6889 1167.4089,-141.6052 1538,-123 2019.437,-98.8299 2140.5045,-109.9768 2622,-87 2693.4914,-83.5885 2776.437,-78.1229 2826.2619,-74.6765"/>
-<polygon fill="#191970" stroke="#191970" points="2826.6422,-78.1586 2836.3752,-73.9732 2826.1565,-71.1755 2826.6422,-78.1586"/>
+<path fill="none" stroke="#191970" d="M852.5355,-185.2236C1161.9066,-167.1205 2488.9636,-89.4665 2750.9684,-74.1351"/>
+<polygon fill="#191970" stroke="#191970" points="2751.4657,-77.612 2761.2442,-73.5338 2751.0567,-70.624 2751.4657,-77.612"/>
</g>
<!-- Node22->Node16 -->
<g id="edge31" class="edge">
<title>Node22->Node16</title>
-<path fill="none" stroke="#191970" d="M621.4876,-179.4649C607.318,-165.6489 579.3879,-139.875 552,-123 524.3747,-105.9787 490.1323,-91.7983 465.4822,-82.6558"/>
-<polygon fill="#191970" stroke="#191970" points="466.4763,-79.2932 455.8821,-79.1696 464.0869,-85.8728 466.4763,-79.2932"/>
+<path fill="none" stroke="#191970" d="M778.5076,-179.4323C764.3629,-165.5757 736.4635,-139.7516 709,-123 680.2368,-105.4556 644.4254,-91.1213 619.0609,-82.0712"/>
+<polygon fill="#191970" stroke="#191970" points="620.148,-78.7436 609.5532,-78.7513 617.8404,-85.3523 620.148,-78.7436"/>
</g>
<!-- Node23 -->
<g id="node24" class="node">
<title>Node23</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="606.5,-123.5 606.5,-142.5 655.5,-142.5 655.5,-123.5 606.5,-123.5"/>
-<text text-anchor="middle" x="631" y="-130.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">cstdlib</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="763.5,-123.5 763.5,-142.5 812.5,-142.5 812.5,-123.5 763.5,-123.5"/>
+<text text-anchor="middle" x="788" y="-130.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">cstdlib</text>
</g>
<!-- Node22->Node23 -->
<g id="edge29" class="edge">
<title>Node22->Node23</title>
-<path fill="none" stroke="#191970" d="M631,-179.2455C631,-171.9382 631,-161.6944 631,-152.7046"/>
-<polygon fill="#191970" stroke="#191970" points="634.5001,-152.6426 631,-142.6427 627.5001,-152.6427 634.5001,-152.6426"/>
+<path fill="none" stroke="#191970" d="M788,-179.2455C788,-171.9382 788,-161.6944 788,-152.7046"/>
+<polygon fill="#191970" stroke="#191970" points="791.5001,-152.6426 788,-142.6427 784.5001,-152.6427 791.5001,-152.6426"/>
</g>
<!-- Node26->Node7 -->
<g id="edge37" class="edge">
<title>Node26->Node7</title>
-<path fill="none" stroke="#191970" d="M1093.349,-291.3669C1109.6033,-283.0301 1127.2394,-271.0608 1138,-255 1147.0814,-241.4455 1149.509,-222.8974 1149.8504,-208.9153"/>
-<polygon fill="#191970" stroke="#191970" points="1153.3499,-208.833 1149.7641,-198.8634 1146.3501,-208.8931 1153.3499,-208.833"/>
+<path fill="none" stroke="#191970" d="M1073.9843,-291.4414C1092.8191,-283.6261 1112.0045,-272.0198 1124,-255 1133.4281,-241.6229 1134.5493,-222.7736 1133.679,-208.6418"/>
+<polygon fill="#191970" stroke="#191970" points="1137.1445,-208.1202 1132.6747,-198.5144 1130.1787,-208.811 1137.1445,-208.1202"/>
</g>
<!-- Node26->Node8 -->
<g id="edge41" class="edge">
<title>Node26->Node8</title>
-<path fill="none" stroke="#191970" d="M1086.5521,-291.2965C1106.8603,-281.3329 1133.7155,-267.8237 1157,-255 1214.6188,-223.267 1221.9942,-200.959 1284,-179 1329.3146,-162.9521 1456.539,-147.9622 1537.1311,-139.6824"/>
-<polygon fill="#191970" stroke="#191970" points="1537.6116,-143.1517 1547.2059,-138.6573 1536.903,-136.1876 1537.6116,-143.1517"/>
+<path fill="none" stroke="#191970" d="M1063.6179,-291.4973C1087.9275,-282.6521 1117.5982,-270.2063 1142,-255 1185.2909,-228.0227 1182.8896,-202.8102 1228,-179 1262.7585,-160.6538 1305.1574,-149.3845 1340.2849,-142.5959"/>
+<polygon fill="#191970" stroke="#191970" points="1341.2638,-145.9743 1350.4605,-140.714 1339.9908,-139.091 1341.2638,-145.9743"/>
</g>
<!-- Node26->Node13 -->
<g id="edge39" class="edge">
<title>Node26->Node13</title>
-<path fill="none" stroke="#191970" d="M998.7913,-291.4557C980.0402,-283.6289 960.9111,-272.0112 949,-255 914.0242,-205.0481 920.5639,-127.8978 926.2128,-91.6355"/>
-<polygon fill="#191970" stroke="#191970" points="929.7299,-91.8313 927.9618,-81.3851 922.8296,-90.6539 929.7299,-91.8313"/>
+<path fill="none" stroke="#191970" d="M956.4515,-291.4417C923.5253,-281.8966 887.5919,-268.783 878,-255 855.22,-222.2667 890.6145,-237.8484 934,-143 941.7801,-125.9912 948.0624,-105.5959 952.0828,-90.9488"/>
+<polygon fill="#191970" stroke="#191970" points="955.5355,-91.5819 954.7038,-81.0198 948.7674,-89.7953 955.5355,-91.5819"/>
</g>
<!-- Node26->Node14 -->
<g id="edge47" class="edge">
<title>Node26->Node14</title>
-<path fill="none" stroke="#191970" d="M1086.0123,-291.3604C1106.3376,-281.3109 1133.413,-267.6828 1157,-255 1261.1455,-199.0006 1274.5775,-156.43 1388,-123 1477.542,-96.6085 2149.1605,-76.7029 2306.2191,-72.3776"/>
-<polygon fill="#191970" stroke="#191970" points="2306.7224,-75.8652 2316.623,-72.0932 2306.531,-68.8679 2306.7224,-75.8652"/>
+<path fill="none" stroke="#191970" d="M1080.254,-300.9519C1116.2272,-297.8661 1162.1228,-294.0457 1203,-291 1446.4059,-272.8644 1513.1689,-309.8789 1751,-255 1902.1098,-220.1318 1923.9449,-162.1592 2074,-123 2214.6449,-86.2965 2388.918,-75.5412 2459.7804,-72.5672"/>
+<polygon fill="#191970" stroke="#191970" points="2460.0037,-76.0612 2469.8574,-72.1683 2459.7267,-69.0667 2460.0037,-76.0612"/>
</g>
<!-- Node26->Node15 -->
-<g id="edge50" class="edge">
+<g id="edge49" class="edge">
<title>Node26->Node15</title>
-<path fill="none" stroke="#191970" d="M1118.0506,-302.6609C1301.1773,-291.484 1822.4256,-259.4721 1859,-255 2235.9485,-208.9085 2685.0452,-112.7516 2826.2492,-81.5267"/>
-<polygon fill="#191970" stroke="#191970" points="2827.2602,-84.8876 2836.2657,-79.3063 2825.7452,-78.0535 2827.2602,-84.8876"/>
+<path fill="none" stroke="#191970" d="M1080.2482,-300.8724C1116.2195,-297.7613 1162.1152,-293.9421 1203,-291 1466.8023,-272.0167 1534.4524,-286.9496 1797,-255 2168.8606,-209.748 2611.4788,-113.2398 2751.3591,-81.7027"/>
+<polygon fill="#191970" stroke="#191970" points="2752.3037,-85.0776 2761.2859,-79.459 2750.7603,-78.2498 2752.3037,-85.0776"/>
</g>
<!-- Node26->Node16 -->
-<g id="edge49" class="edge">
+<g id="edge51" class="edge">
<title>Node26->Node16</title>
-<path fill="none" stroke="#191970" d="M991.8119,-300.5773C882.2548,-290.0438 665.5806,-267.9463 633,-255 552.3162,-222.9391 549.9705,-184.8092 489,-123 477.0256,-110.8609 462.8876,-97.8686 451.8131,-87.9705"/>
-<polygon fill="#191970" stroke="#191970" points="453.8331,-85.084 444.0281,-81.0699 449.1899,-90.3223 453.8331,-85.084"/>
+<path fill="none" stroke="#191970" d="M953.819,-292.7177C950.8403,-292.1248 947.8891,-291.5492 945,-291 848.5265,-272.6611 811.1591,-305.6034 727,-255 709.4469,-244.4456 630.6316,-133.703 599.6058,-89.537"/>
+<polygon fill="#191970" stroke="#191970" points="602.3559,-87.3626 593.7499,-81.1831 596.6239,-91.3806 602.3559,-87.3626"/>
</g>
<!-- Node26->Node18 -->
-<g id="edge51" class="edge">
+<g id="edge52" class="edge">
<title>Node26->Node18</title>
-<path fill="none" stroke="#191970" d="M1118.3556,-295.9977C1197.5895,-282.8632 1330.3444,-260.8567 1392.2761,-250.5904"/>
-<polygon fill="#191970" stroke="#191970" points="1392.9964,-254.0188 1402.2894,-248.9305 1391.8516,-247.113 1392.9964,-254.0188"/>
+<path fill="none" stroke="#191970" d="M1080.2555,-296.5506C1163.9933,-283.3796 1308.6679,-260.6238 1374.1092,-250.3307"/>
+<polygon fill="#191970" stroke="#191970" points="1374.943,-253.7426 1384.2777,-248.7313 1373.8553,-246.8276 1374.943,-253.7426"/>
</g>
<!-- Node26->Node21 -->
<g id="edge38" class="edge">
<title>Node26->Node21</title>
-<path fill="none" stroke="#191970" d="M991.8762,-298.9449C926.9489,-290.6704 822.7768,-275.9979 724.6128,-255.1771"/>
-<polygon fill="#191970" stroke="#191970" points="725.3399,-251.7535 714.828,-253.0767 723.8707,-258.5976 725.3399,-251.7535"/>
+<path fill="none" stroke="#191970" d="M955.1092,-291.4554C906.9639,-279.7521 841.7507,-263.8999 801.3045,-254.0681"/>
+<polygon fill="#191970" stroke="#191970" points="802.1244,-250.6655 791.5807,-251.7044 800.471,-257.4675 802.1244,-250.6655"/>
</g>
<!-- Node26->Node22 -->
<g id="edge40" class="edge">
<title>Node26->Node22</title>
-<path fill="none" stroke="#191970" d="M991.9807,-300.3116C881.3977,-289.1985 662.833,-265.9967 650,-255 636.8686,-243.7476 632.5177,-224.0618 631.213,-209.1097"/>
-<polygon fill="#191970" stroke="#191970" points="634.6943,-208.6353 630.7142,-198.8164 627.7025,-208.9742 634.6943,-208.6353"/>
+<path fill="none" stroke="#191970" d="M953.9435,-291.7456C912.387,-281.2677 862.3696,-266.9895 844,-255 825.0336,-242.621 808.9892,-221.9412 799.0003,-207.0255"/>
+<polygon fill="#191970" stroke="#191970" points="801.9422,-205.1293 793.5914,-198.609 796.0534,-208.9138 801.9422,-205.1293"/>
</g>
<!-- Node26->Node24 -->
<g id="edge42" class="edge">
<title>Node26->Node24</title>
-<path fill="none" stroke="#191970" d="M1071.3162,-291.4379C1086.1517,-275.9398 1103.6616,-251.7591 1089,-235 1069.4462,-212.6487 875.5059,-197.3842 788.2054,-191.592"/>
-<polygon fill="#191970" stroke="#191970" points="788.3133,-188.0916 778.1065,-190.9327 787.8573,-195.0768 788.3133,-188.0916"/>
+<path fill="none" stroke="#191970" d="M953.8331,-292.6415C950.8506,-292.069 947.8948,-291.5185 945,-291 834.0685,-271.1291 805.5035,-271.3589 694,-255 547.7171,-233.5386 374.2084,-207.274 295.9623,-195.3862"/>
+<polygon fill="#191970" stroke="#191970" points="296.4226,-191.9161 286.0102,-193.8737 295.3707,-198.8366 296.4226,-191.9161"/>
</g>
<!-- Node26->Node25 -->
<g id="edge45" class="edge">
<title>Node26->Node25</title>
-<path fill="none" stroke="#191970" d="M1086.0747,-291.4796C1109.6802,-277.9178 1135.9081,-256.7984 1122,-235 1111.2358,-218.1291 1092.5081,-207.3496 1074.2177,-200.5067"/>
-<polygon fill="#191970" stroke="#191970" points="1075.3319,-197.1888 1064.7374,-197.2758 1073.0737,-203.8146 1075.3319,-197.1888"/>
+<path fill="none" stroke="#191970" d="M1064.6358,-291.4168C1081.7276,-283.445 1099.1873,-271.7385 1109,-255 1122.5172,-231.9424 1093.8247,-213.6887 1066.7404,-202.2759"/>
+<polygon fill="#191970" stroke="#191970" points="1067.8851,-198.9646 1057.2992,-198.5276 1065.3021,-205.4706 1067.8851,-198.9646"/>
</g>
<!-- Node27 -->
<g id="node28" class="node">
<title>Node27</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="958.5,-235.5 958.5,-254.5 1011.5,-254.5 1011.5,-235.5 958.5,-235.5"/>
-<text text-anchor="middle" x="985" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">cstddef</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="886.5,-235.5 886.5,-254.5 939.5,-254.5 939.5,-235.5 886.5,-235.5"/>
+<text text-anchor="middle" x="913" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">cstddef</text>
</g>
<!-- Node26->Node27 -->
<g id="edge43" class="edge">
<title>Node26->Node27</title>
-<path fill="none" stroke="#191970" d="M1037.6966,-291.2977C1027.4738,-282.3163 1014.5177,-270.9334 1004.0104,-261.702"/>
-<polygon fill="#191970" stroke="#191970" points="1006.0305,-258.8178 996.2079,-254.8469 1001.4103,-264.0766 1006.0305,-258.8178"/>
+<path fill="none" stroke="#191970" d="M991.5587,-291.4554C975.2716,-281.8241 954.2327,-269.3828 938.027,-259.7996"/>
+<polygon fill="#191970" stroke="#191970" points="939.5055,-256.6078 929.1163,-254.5303 935.9424,-262.6331 939.5055,-256.6078"/>
</g>
<!-- Node28 -->
<g id="node29" class="node">
<title>Node28</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1030,-235.5 1030,-254.5 1080,-254.5 1080,-235.5 1030,-235.5"/>
-<text text-anchor="middle" x="1055" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">cstring</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="958,-235.5 958,-254.5 1008,-254.5 1008,-235.5 958,-235.5"/>
+<text text-anchor="middle" x="983" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">cstring</text>
</g>
<!-- Node26->Node28 -->
<g id="edge44" class="edge">
<title>Node26->Node28</title>
-<path fill="none" stroke="#191970" d="M1055,-291.2977C1055,-283.3834 1055,-273.6043 1055,-265.0759"/>
-<polygon fill="#191970" stroke="#191970" points="1058.5001,-264.8469 1055,-254.8469 1051.5001,-264.847 1058.5001,-264.8469"/>
+<path fill="none" stroke="#191970" d="M1008.5955,-291.2977C1003.9743,-282.9388 998.2031,-272.4997 993.3111,-263.6509"/>
+<polygon fill="#191970" stroke="#191970" points="996.3452,-261.9051 988.4438,-254.8469 990.2191,-265.292 996.3452,-261.9051"/>
</g>
<!-- Node29 -->
<g id="node30" class="node">
<title>Node29</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="844,-235.5 844,-254.5 902,-254.5 902,-235.5 844,-235.5"/>
-<text text-anchor="middle" x="873" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">memory</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="320,-235.5 320,-254.5 378,-254.5 378,-235.5 320,-235.5"/>
+<text text-anchor="middle" x="349" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">memory</text>
</g>
<!-- Node26->Node29 -->
<g id="edge46" class="edge">
<title>Node26->Node29</title>
-<path fill="none" stroke="#191970" d="M1010.4778,-291.4554C979.9774,-281.149 939.955,-267.6249 910.9642,-257.8286"/>
-<polygon fill="#191970" stroke="#191970" points="911.7978,-254.4159 901.2035,-254.5303 909.5568,-261.0475 911.7978,-254.4159"/>
+<path fill="none" stroke="#191970" d="M953.8755,-292.3769C950.8817,-291.8752 947.9119,-291.4121 945,-291 715.6211,-258.5396 655.5075,-278.1218 425,-255 413.0173,-253.798 400.0097,-252.1963 388.199,-250.6276"/>
+<polygon fill="#191970" stroke="#191970" points="388.5637,-247.1451 378.1838,-249.2685 387.6223,-254.0815 388.5637,-247.1451"/>
</g>
<!-- Node30 -->
<g id="node31" class="node">
<title>Node30</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="732.5,-235.5 732.5,-254.5 825.5,-254.5 825.5,-235.5 732.5,-235.5"/>
-<text text-anchor="middle" x="779" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">unordered_map</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1026.5,-235.5 1026.5,-254.5 1099.5,-254.5 1099.5,-235.5 1026.5,-235.5"/>
+<text text-anchor="middle" x="1063" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">string_view</text>
</g>
<!-- Node26->Node30 -->
<g id="edge48" class="edge">
<title>Node26->Node30</title>
-<path fill="none" stroke="#191970" d="M991.68,-292.3907C943.2973,-281.6097 877.3886,-266.9236 831.8243,-256.7706"/>
-<polygon fill="#191970" stroke="#191970" points="832.3736,-253.3073 821.8517,-254.5485 830.8511,-260.1397 832.3736,-253.3073"/>
+<path fill="none" stroke="#191970" d="M1028.3708,-291.2977C1034.7561,-282.7609 1042.7641,-272.0545 1049.4705,-263.0883"/>
+<polygon fill="#191970" stroke="#191970" points="1052.4479,-264.9511 1055.6348,-254.8469 1046.8425,-260.7584 1052.4479,-264.9511"/>
</g>
-<!-- Node31->Node16 -->
-<g id="edge57" class="edge">
-<title>Node31->Node16</title>
-<path fill="none" stroke="#191970" d="M590.5563,-291.4349C562.083,-272.8085 515.2251,-238.6643 486,-199 460.9913,-165.0582 445.3299,-117.5751 437.9762,-91.1768"/>
-<polygon fill="#191970" stroke="#191970" points="441.2889,-90.0155 435.3302,-81.2551 434.5253,-91.8193 441.2889,-90.0155"/>
+<!-- Node31 -->
+<g id="node32" class="node">
+<title>Node31</title>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="434.5,-235.5 434.5,-254.5 527.5,-254.5 527.5,-235.5 434.5,-235.5"/>
+<text text-anchor="middle" x="481" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">unordered_map</text>
</g>
-<!-- Node31->Node18 -->
+<!-- Node26->Node31 -->
+<g id="edge50" class="edge">
+<title>Node26->Node31</title>
+<path fill="none" stroke="#191970" d="M953.8598,-292.4824C950.8701,-291.9525 947.9055,-291.4545 945,-291 783.9493,-265.8061 742.2033,-271.1911 580,-255 566.3706,-253.6395 551.6869,-252.1625 537.9096,-250.7716"/>
+<polygon fill="#191970" stroke="#191970" points="537.8843,-247.2513 527.5831,-249.7282 537.1805,-254.2159 537.8843,-247.2513"/>
+</g>
+<!-- Node32->Node16 -->
<g id="edge58" class="edge">
-<title>Node31->Node18</title>
-<path fill="none" stroke="#191970" d="M678.0147,-296.7988C693.3614,-294.6635 709.7486,-292.5727 725,-291 980.7644,-264.6255 1290.6755,-250.4946 1392.1016,-246.3282"/>
-<polygon fill="#191970" stroke="#191970" points="1392.4958,-249.8151 1402.3455,-245.9123 1392.2118,-242.8209 1392.4958,-249.8151"/>
+<title>Node32->Node16</title>
+<path fill="none" stroke="#191970" d="M315.0839,-291.4268C260.8302,-269.5739 177.2453,-226.4794 213,-179 254.8486,-123.4283 471.4027,-87.6977 554.4924,-75.8491"/>
+<polygon fill="#191970" stroke="#191970" points="554.98,-79.315 564.397,-74.4602 554.0079,-72.3829 554.98,-79.315"/>
</g>
-<!-- Node31->Node21 -->
+<!-- Node32->Node18 -->
<g id="edge59" class="edge">
-<title>Node31->Node21</title>
-<path fill="none" stroke="#191970" d="M632.7978,-291.2977C643.4167,-282.2274 656.9031,-270.7077 667.7667,-261.4285"/>
-<polygon fill="#191970" stroke="#191970" points="670.1413,-264.0031 675.4719,-254.8469 665.5949,-258.6805 670.1413,-264.0031"/>
+<title>Node32->Node18</title>
+<path fill="none" stroke="#191970" d="M420.24,-300.7562C456.2088,-297.6082 502.1046,-293.7907 543,-291 866.3621,-268.9335 1258.3405,-251.4186 1374.2139,-246.4314"/>
+<polygon fill="#191970" stroke="#191970" points="1374.4075,-249.9264 1384.2483,-246.0012 1374.1076,-242.9329 1374.4075,-249.9264"/>
</g>
-<!-- Node31->Node24 -->
-<g id="edge55" class="edge">
-<title>Node31->Node24</title>
-<path fill="none" stroke="#191970" d="M619.168,-291.3279C624.1885,-275.512 634.004,-250.8626 650,-235 665.0861,-220.0397 685.9823,-209.2394 704.4281,-201.8797"/>
-<polygon fill="#191970" stroke="#191970" points="705.6864,-205.1458 713.8069,-198.3409 703.2152,-198.5966 705.6864,-205.1458"/>
+<!-- Node32->Node21 -->
+<g id="edge60" class="edge">
+<title>Node32->Node21</title>
+<path fill="none" stroke="#191970" d="M420.1472,-296.9581C505.7349,-284.0253 655.9027,-261.3341 725.9781,-250.7453"/>
+<polygon fill="#191970" stroke="#191970" points="726.7525,-254.1681 736.1173,-249.2132 725.7066,-247.2467 726.7525,-254.1681"/>
</g>
-<!-- Node31->Node29 -->
+<!-- Node32->Node24 -->
<g id="edge56" class="edge">
-<title>Node31->Node29</title>
-<path fill="none" stroke="#191970" d="M678.3315,-291.9606C719.7994,-282.3585 775.4208,-269.3254 833.8466,-255.0021"/>
-<polygon fill="#191970" stroke="#191970" points="834.7527,-258.3836 843.6285,-252.5983 833.0821,-251.5859 834.7527,-258.3836"/>
+<title>Node32->Node24</title>
+<path fill="none" stroke="#191970" d="M343.2457,-291.2358C334.0692,-281.019 321.7593,-267.2499 311,-255 296.7393,-238.7635 280.6613,-220.1151 269.2258,-206.7919"/>
+<polygon fill="#191970" stroke="#191970" points="271.6128,-204.1987 262.4471,-198.8844 266.2983,-208.7546 271.6128,-204.1987"/>
</g>
-<!-- Node32->Node9 -->
-<g id="edge61" class="edge">
-<title>Node32->Node9</title>
-<path fill="none" stroke="#191970" d="M2496.1086,-296.9204C2531.1152,-282.0254 2593.1896,-249.648 2619,-199 2634.3367,-168.9047 2634.3465,-153.0902 2619,-123 2612.4864,-110.2286 2601.1608,-99.8854 2589.4436,-91.9119"/>
-<polygon fill="#191970" stroke="#191970" points="2591.2304,-88.9009 2580.9056,-86.5239 2587.4946,-94.8206 2591.2304,-88.9009"/>
+<!-- Node32->Node29 -->
+<g id="edge57" class="edge">
+<title>Node32->Node29</title>
+<path fill="none" stroke="#191970" d="M355.0225,-291.2977C353.9814,-283.2945 352.6923,-273.3843 351.5742,-264.7889"/>
+<polygon fill="#191970" stroke="#191970" points="355.0417,-264.3119 350.2809,-254.8469 348.1002,-265.2149 355.0417,-264.3119"/>
</g>
-<!-- Node32->Node13 -->
+<!-- Node33->Node9 -->
<g id="edge62" class="edge">
-<title>Node32->Node13</title>
-<path fill="none" stroke="#191970" d="M2455.5987,-296.9335C2415.4485,-272.5158 2303.5653,-207.7861 2201,-179 1972.7939,-114.9513 1237.515,-82.8503 1002.7628,-74.0525"/>
-<polygon fill="#191970" stroke="#191970" points="1002.7516,-70.5498 992.6284,-73.6756 1002.4913,-77.5449 1002.7516,-70.5498"/>
+<title>Node33->Node9</title>
+<path fill="none" stroke="#191970" d="M2397.49,-296.7781C2400.6591,-282.0869 2404.3552,-253.938 2392,-235 2319.4918,-123.8597 2156.973,-88.2669 2060.9921,-76.8689"/>
+<polygon fill="#191970" stroke="#191970" points="2061.1021,-73.3592 2050.7729,-75.7168 2060.3178,-80.3151 2061.1021,-73.3592"/>
</g>
-<!-- Node32->Node14 -->
+<!-- Node33->Node13 -->
<g id="edge63" class="edge">
-<title>Node32->Node14</title>
-<path fill="none" stroke="#191970" d="M2468.1748,-296.6867C2459.6571,-268.1164 2432.2497,-183.2072 2391,-123 2381.9581,-109.8027 2369.2321,-97.1296 2358.6328,-87.6573"/>
-<polygon fill="#191970" stroke="#191970" points="2360.8895,-84.9816 2351.0383,-81.0821 2356.3076,-90.2737 2360.8895,-84.9816"/>
-</g>
-<!-- Node32->Node15 -->
-<g id="edge64" class="edge">
-<title>Node32->Node15</title>
-<path fill="none" stroke="#191970" d="M2501.0091,-296.988C2563.1627,-276.2722 2708.5453,-222.4716 2809,-143 2828.6039,-127.491 2846.8045,-105.1627 2858.3459,-89.5983"/>
-<polygon fill="#191970" stroke="#191970" points="2861.3094,-91.473 2864.332,-81.3184 2855.6367,-87.3717 2861.3094,-91.473"/>
-</g>
-<!-- Node33->Node6 -->
-<g id="edge67" class="edge">
-<title>Node33->Node6</title>
-<path fill="none" stroke="#191970" d="M2236.5954,-425.389C2216.1003,-400.4222 2175.2724,-350.6864 2153.4857,-324.1462"/>
-<polygon fill="#191970" stroke="#191970" points="2155.9342,-321.6127 2146.884,-316.1042 2150.5237,-326.0542 2155.9342,-321.6127"/>
+<title>Node33->Node13</title>
+<path fill="none" stroke="#191970" d="M2366.8816,-296.9974C2296.0704,-273.4247 2104.3158,-211.8037 1940,-179 1603.9344,-111.9085 1194.2852,-84.0849 1030.1404,-75.0834"/>
+<polygon fill="#191970" stroke="#191970" points="1029.8615,-71.5633 1019.6869,-74.5175 1029.4831,-78.553 1029.8615,-71.5633"/>
</g>
<!-- Node33->Node14 -->
-<g id="edge89" class="edge">
+<g id="edge64" class="edge">
<title>Node33->Node14</title>
-<path fill="none" stroke="#191970" d="M2276.1067,-425.4723C2289.2776,-416.6568 2303.9624,-404.2459 2312,-389 2365.3587,-287.7878 2350.1229,-142.8791 2342.3006,-90.9419"/>
-<polygon fill="#191970" stroke="#191970" points="2345.757,-90.3903 2340.7285,-81.0648 2338.844,-91.4907 2345.757,-90.3903"/>
-</g>
-<!-- Node33->Node32 -->
-<g id="edge68" class="edge">
-<title>Node33->Node32</title>
-<path fill="none" stroke="#191970" d="M2305.8664,-426.2857C2353.8926,-413.9008 2416.3672,-396.7467 2426,-389 2446.2875,-372.6848 2458.7676,-344.6551 2465.291,-325.8537"/>
-<polygon fill="#191970" stroke="#191970" points="2468.65,-326.8411 2468.386,-316.2495 2461.9874,-324.694 2468.65,-326.8411"/>
+<path fill="none" stroke="#191970" d="M2401.7118,-296.6285C2408.759,-286.2628 2420.1614,-269.4879 2430,-255 2446.8977,-230.1171 2456.5333,-226.8066 2468,-199 2482.893,-162.8849 2488.6221,-117.1834 2490.7684,-91.4615"/>
+<polygon fill="#191970" stroke="#191970" points="2494.2833,-91.3843 2491.5225,-81.1555 2487.302,-90.8734 2494.2833,-91.3843"/>
</g>
-<!-- Node33->Node34 -->
-<g id="edge69" class="edge">
-<title>Node33->Node34</title>
-<path fill="none" stroke="#191970" d="M2192.4608,-432.2064C2175.5943,-429.8008 2157.0581,-427.2246 2140,-425 2012.9851,-408.4354 1864.7588,-391.3222 1778.5296,-381.5873"/>
-<polygon fill="#191970" stroke="#191970" points="1778.9131,-378.1084 1768.5839,-380.4661 1778.1289,-385.0644 1778.9131,-378.1084"/>
+<!-- Node33->Node15 -->
+<g id="edge65" class="edge">
+<title>Node33->Node15</title>
+<path fill="none" stroke="#191970" d="M2451.3073,-296.9768C2533.2122,-280.2326 2684.9127,-237.8573 2772,-143 2785.1113,-128.7189 2791.0845,-107.0397 2793.7907,-91.2974"/>
+<polygon fill="#191970" stroke="#191970" points="2797.3016,-91.4665 2795.2391,-81.0744 2790.3708,-90.4845 2797.3016,-91.4665"/>
</g>
-<!-- Node39 -->
-<g id="node40" class="node">
-<title>Node39</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1962.5,-297 1962.5,-316 2027.5,-316 2027.5,-297 1962.5,-297"/>
-<text text-anchor="middle" x="1995" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">functional</text>
+<!-- Node34->Node6 -->
+<g id="edge68" class="edge">
+<title>Node34->Node6</title>
+<path fill="none" stroke="#191970" d="M2170.7894,-425.389C2172.0596,-401.0735 2174.5571,-353.2644 2175.9673,-326.2698"/>
+<polygon fill="#191970" stroke="#191970" points="2179.4718,-326.2732 2176.4983,-316.1042 2172.4813,-325.9079 2179.4718,-326.2732"/>
</g>
-<!-- Node33->Node39 -->
-<g id="edge88" class="edge">
-<title>Node33->Node39</title>
-<path fill="none" stroke="#191970" d="M2220.3567,-425.389C2170.974,-399.3367 2070.4696,-346.3147 2022.1438,-320.8199"/>
-<polygon fill="#191970" stroke="#191970" points="2023.6827,-317.6747 2013.2049,-316.1042 2020.4164,-323.8659 2023.6827,-317.6747"/>
+<!-- Node34->Node14 -->
+<g id="edge90" class="edge">
+<title>Node34->Node14</title>
+<path fill="none" stroke="#191970" d="M2226.6423,-435.2144C2296.2359,-425.5014 2412.4967,-398.1933 2473,-322 2527.6227,-253.2124 2508.0613,-137.6222 2497.2311,-91.4405"/>
+<polygon fill="#191970" stroke="#191970" points="2500.5714,-90.3676 2494.7767,-81.4979 2493.7754,-92.0453 2500.5714,-90.3676"/>
</g>
-<!-- Node34->Node8 -->
-<g id="edge78" class="edge">
-<title>Node34->Node8</title>
-<path fill="none" stroke="#191970" d="M1697.2927,-363.9023C1680.4675,-344.7636 1643.6003,-299.8236 1626,-255 1612.5407,-220.7224 1608.5893,-177.5096 1607.4461,-152.7901"/>
-<polygon fill="#191970" stroke="#191970" points="1610.9353,-152.4174 1607.0876,-142.546 1603.9395,-152.6623 1610.9353,-152.4174"/>
+<!-- Node34->Node33 -->
+<g id="edge69" class="edge">
+<title>Node34->Node33</title>
+<path fill="none" stroke="#191970" d="M2196.8961,-425.3773C2214.844,-415.2201 2238.9273,-401.4645 2260,-389 2299.1237,-365.8583 2344.0794,-338.1607 2370.978,-321.4678"/>
+<polygon fill="#191970" stroke="#191970" points="2373.0537,-324.2987 2379.7001,-316.0479 2369.3591,-318.3531 2373.0537,-324.2987"/>
</g>
-<!-- Node34->Node9 -->
+<!-- Node34->Node35 -->
<g id="edge70" class="edge">
-<title>Node34->Node9</title>
-<path fill="none" stroke="#191970" d="M1747.1706,-363.9822C1802.3954,-351.081 1895.2655,-328.8641 1910,-322 1931.3561,-312.0512 1932.2017,-302.0676 1953,-291 2135.3328,-193.9736 2369.9345,-121.2827 2482.3001,-89.3609"/>
-<polygon fill="#191970" stroke="#191970" points="2483.4633,-92.6693 2492.1352,-86.5824 2481.5603,-85.933 2483.4633,-92.6693"/>
+<title>Node34->Node35</title>
+<path fill="none" stroke="#191970" d="M2113.4852,-433.488C2014.9701,-421.2648 1811.9226,-396.0719 1702.9379,-382.5497"/>
+<polygon fill="#191970" stroke="#191970" points="1703.0462,-379.0364 1692.6913,-381.2784 1702.1843,-385.9831 1703.0462,-379.0364"/>
</g>
-<!-- Node34->Node16 -->
-<g id="edge86" class="edge">
-<title>Node34->Node16</title>
-<path fill="none" stroke="#191970" d="M1643.4319,-365.9826C1617.092,-363.0688 1586.0914,-359.9668 1558,-358 1501.7134,-354.0591 591.5311,-350.7826 543,-322 459.2387,-272.3231 439.0878,-141.5391 434.384,-91.429"/>
-<polygon fill="#191970" stroke="#191970" points="437.8558,-90.9415 433.5433,-81.264 430.8796,-91.5186 437.8558,-90.9415"/>
+<!-- Node40 -->
+<g id="node41" class="node">
+<title>Node40</title>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1924.5,-297 1924.5,-316 1989.5,-316 1989.5,-297 1924.5,-297"/>
+<text text-anchor="middle" x="1957" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">functional</text>
</g>
-<!-- Node34->Node17 -->
-<g id="edge84" class="edge">
-<title>Node34->Node17</title>
-<path fill="none" stroke="#191970" d="M1706.801,-363.7128C1708.4854,-340.999 1711.7842,-282.9834 1706,-235 1699.6146,-182.0295 1681.9913,-121.4517 1672.4077,-91.0609"/>
-<polygon fill="#191970" stroke="#191970" points="1675.6639,-89.753 1669.2764,-81.3001 1668.9985,-91.8913 1675.6639,-89.753"/>
+<!-- Node34->Node40 -->
+<g id="edge89" class="edge">
+<title>Node34->Node40</title>
+<path fill="none" stroke="#191970" d="M2145.9802,-425.389C2104.9138,-399.5538 2021.6892,-347.1965 1980.7845,-321.463"/>
+<polygon fill="#191970" stroke="#191970" points="1982.5944,-318.4667 1972.2663,-316.1042 1978.8669,-324.3917 1982.5944,-318.4667"/>
</g>
-<!-- Node34->Node18 -->
-<g id="edge87" class="edge">
-<title>Node34->Node18</title>
-<path fill="none" stroke="#191970" d="M1684.9545,-363.8416C1635.7817,-341.2748 1513.2077,-285.0221 1456.1294,-258.8273"/>
-<polygon fill="#191970" stroke="#191970" points="1457.4563,-255.5852 1446.9078,-254.5952 1454.5365,-261.9473 1457.4563,-255.5852"/>
+<!-- Node35->Node8 -->
+<g id="edge79" class="edge">
+<title>Node35->Node8</title>
+<path fill="none" stroke="#191970" d="M1624.5645,-363.8518C1614.6702,-347.0397 1592.2073,-312.1531 1565,-291 1533.8986,-266.8193 1513.5684,-281.0329 1484,-255 1450.9903,-225.9373 1428.1749,-178.5083 1417.2903,-152.1654"/>
+<polygon fill="#191970" stroke="#191970" points="1420.4255,-150.5777 1413.4754,-142.581 1413.9218,-153.1665 1420.4255,-150.5777"/>
</g>
-<!-- Node34->Node20 -->
+<!-- Node35->Node9 -->
<g id="edge71" class="edge">
-<title>Node34->Node20</title>
-<path fill="none" stroke="#191970" d="M1643.4341,-365.9505C1617.0946,-363.0314 1586.0936,-359.9343 1558,-358 1076.2088,-324.8279 951.6924,-377.842 472,-322 470.8826,-321.8699 469.7572,-321.7321 468.6255,-321.5872"/>
-<polygon fill="#191970" stroke="#191970" points="468.7119,-318.0639 458.3164,-320.1092 467.7184,-324.9931 468.7119,-318.0639"/>
+<title>Node35->Node9</title>
+<path fill="none" stroke="#191970" d="M1672.3272,-363.9989C1715.8816,-353.6578 1780.2141,-336.6395 1801,-322 1888.0217,-260.7107 1950.6105,-145.5371 1974.8794,-95.6061"/>
+<polygon fill="#191970" stroke="#191970" points="1978.0651,-97.057 1979.2181,-86.5251 1971.749,-94.0392 1978.0651,-97.057"/>
</g>
-<!-- Node34->Node26 -->
-<g id="edge76" class="edge">
-<title>Node34->Node26</title>
-<path fill="none" stroke="#191970" d="M1643.3517,-366.9064C1616.999,-364.1412 1586.0103,-360.9007 1558,-358 1405.1491,-342.1711 1226.1399,-323.909 1128.2191,-313.9429"/>
-<polygon fill="#191970" stroke="#191970" points="1128.4355,-310.4469 1118.1325,-312.9164 1127.7268,-317.4109 1128.4355,-310.4469"/>
-</g>
-<!-- Node34->Node32 -->
-<g id="edge77" class="edge">
-<title>Node34->Node32</title>
-<path fill="none" stroke="#191970" d="M1768.6032,-368.0171C1907.0012,-355.896 2238.7638,-326.8396 2391.5316,-313.46"/>
-<polygon fill="#191970" stroke="#191970" points="2392.1175,-316.9221 2401.774,-312.5629 2391.5068,-309.9488 2392.1175,-316.9221"/>
+<!-- Node35->Node16 -->
+<g id="edge87" class="edge">
+<title>Node35->Node16</title>
+<path fill="none" stroke="#191970" d="M1567.3408,-368.8153C1524.0061,-365.6267 1465.0119,-361.3872 1413,-358 1277.0529,-349.1467 932.6661,-356.9863 801,-322 768.6317,-313.3991 691.7449,-276.4214 666,-255 642.0848,-235.1011 639.7731,-225.8163 624,-199 610.1812,-175.5063 606.8745,-169.0942 599,-143 593.8926,-126.0753 590.7774,-106.1996 589.0057,-91.7055"/>
+<polygon fill="#191970" stroke="#191970" points="592.4431,-90.947 587.8605,-81.3945 585.4858,-91.7198 592.4431,-90.947"/>
</g>
-<!-- Node35 -->
-<g id="node36" class="node">
-<title>Node35</title>
-<g id="a_node36"><a xlink:href="shape__tuple_8h.html" target="_top" xlink:title="Runtime ShapeTuple container types. ">
-<polygon fill="#ffffff" stroke="#000000" points="734,-291.5 734,-321.5 860,-321.5 860,-291.5 734,-291.5"/>
-<text text-anchor="start" x="742" y="-309.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
-<text text-anchor="middle" x="797" y="-298.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/shape_tuple.h</text>
-</a>
+<!-- Node35->Node17 -->
+<g id="edge85" class="edge">
+<title>Node35->Node17</title>
+<path fill="none" stroke="#191970" d="M1628.1288,-363.8518C1624.4116,-347.0397 1614.6126,-312.1532 1594,-291 1567.635,-263.9436 1541.5353,-283.7259 1517,-255 1477.3507,-208.5786 1500.855,-180.0032 1479,-123 1474.7221,-111.8422 1468.9479,-99.8051 1464.0266,-90.1759"/>
+<polygon fill="#191970" stroke="#191970" points="1467.0419,-88.3891 1459.3053,-81.1507 1460.8394,-91.6339 1467.0419,-88.3891"/>
</g>
+<!-- Node35->Node18 -->
+<g id="edge88" class="edge">
+<title>Node35->Node18</title>
+<path fill="none" stroke="#191970" d="M1620.5994,-363.8875C1603.7711,-347.1309 1566.805,-312.3272 1530,-291 1501.9415,-274.7411 1467.0802,-262.2738 1441.8171,-254.4463"/>
+<polygon fill="#191970" stroke="#191970" points="1442.5491,-251.0115 1431.9645,-251.4774 1440.5294,-257.7138 1442.5491,-251.0115"/>
</g>
-<!-- Node34->Node35 -->
+<!-- Node35->Node20 -->
<g id="edge72" class="edge">
-<title>Node34->Node35</title>
-<path fill="none" stroke="#191970" d="M1643.403,-366.3586C1617.0585,-363.5053 1586.0622,-360.3469 1558,-358 1302.8349,-336.6599 1238.3771,-340.6331 983,-322 945.9508,-319.2968 904.7839,-315.9063 870.6169,-312.9862"/>
-<polygon fill="#191970" stroke="#191970" points="870.5132,-309.4645 860.2504,-312.0963 869.9144,-316.4388 870.5132,-309.4645"/>
+<title>Node35->Node20</title>
+<path fill="none" stroke="#191970" d="M1567.344,-368.7656C1524.0107,-365.5548 1465.0169,-361.3092 1413,-358 1126.4664,-339.7715 1054.2318,-344.475 768,-322 742.0333,-319.9611 713.6379,-317.2251 688.4359,-314.6195"/>
+<polygon fill="#191970" stroke="#191970" points="688.5252,-311.1099 678.2157,-313.5522 687.798,-318.0721 688.5252,-311.1099"/>
+</g>
+<!-- Node35->Node26 -->
+<g id="edge77" class="edge">
+<title>Node35->Node26</title>
+<path fill="none" stroke="#191970" d="M1567.279,-366.6447C1453.0114,-354.1554 1211.5296,-327.7618 1090.0461,-314.4838"/>
+<polygon fill="#191970" stroke="#191970" points="1090.4116,-311.003 1080.0905,-313.3957 1089.651,-317.9616 1090.4116,-311.003"/>
+</g>
+<!-- Node35->Node33 -->
+<g id="edge78" class="edge">
+<title>Node35->Node33</title>
+<path fill="none" stroke="#191970" d="M1692.9194,-368.5352C1803.7314,-359.7158 2040.9243,-340.4995 2241,-322 2265.2319,-319.7595 2291.5917,-317.1644 2315.5414,-314.746"/>
+<polygon fill="#191970" stroke="#191970" points="2316.0487,-318.2126 2325.6447,-313.7219 2315.3427,-311.2483 2316.0487,-318.2126"/>
</g>
<!-- Node36 -->
<g id="node37" class="node">
<title>Node36</title>
-<g id="a_node37"><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="1734,-297 1734,-316 1868,-316 1868,-297 1734,-297"/>
-<text text-anchor="middle" x="1801" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/serializer.h</text>
+<g id="a_node37"><a xlink:href="shape__tuple_8h.html" target="_top" xlink:title="Runtime ShapeTuple container types. ">
+<polygon fill="#ffffff" stroke="#000000" points="810,-291.5 810,-321.5 936,-321.5 936,-291.5 810,-291.5"/>
+<text text-anchor="start" x="818" y="-309.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
+<text text-anchor="middle" x="873" y="-298.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/shape_tuple.h</text>
</a>
</g>
</g>
-<!-- Node34->Node36 -->
-<g id="edge79" class="edge">
-<title>Node34->Node36</title>
-<path fill="none" stroke="#191970" d="M1714.7106,-363.9005C1728.2322,-352.8554 1753.8652,-334.5776 1773.8148,-321.6172"/>
-<polygon fill="#191970" stroke="#191970" points="1775.9579,-324.4021 1782.5169,-316.0817 1772.2008,-318.4958 1775.9579,-324.4021"/>
+<!-- Node35->Node36 -->
+<g id="edge73" class="edge">
+<title>Node35->Node36</title>
+<path fill="none" stroke="#191970" d="M1567.3803,-369.0259C1413.2943,-357.9872 1023.2151,-329.8518 946.4479,-321.8224"/>
+<polygon fill="#191970" stroke="#191970" points="946.4695,-318.3007 936.1276,-320.6017 945.6471,-325.2522 946.4695,-318.3007"/>
</g>
-<!-- Node34->Node39 -->
-<g id="edge85" class="edge">
-<title>Node34->Node39</title>
-<path fill="none" stroke="#191970" d="M1758.7941,-363.9828C1808.0858,-354.6925 1883.5217,-339.4373 1948,-322 1951.3358,-321.0979 1954.7834,-320.0967 1958.2255,-319.0497"/>
-<polygon fill="#191970" stroke="#191970" points="1959.3316,-322.3709 1967.8138,-316.0225 1957.224,-315.6957 1959.3316,-322.3709"/>
+<!-- Node37 -->
+<g id="node38" class="node">
+<title>Node37</title>
+<g id="a_node38"><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="1658,-297 1658,-316 1792,-316 1792,-297 1658,-297"/>
+<text text-anchor="middle" x="1725" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/serializer.h</text>
+</a>
</g>
-<!-- Node35->Node16 -->
-<g id="edge73" class="edge">
-<title>Node35->Node16</title>
-<path fill="none" stroke="#191970" d="M733.9732,-295.6373C679.3953,-285.5046 605.2141,-269.7733 579,-255 543.2244,-234.8382 473.1645,-132.4944 445.074,-90.0176"/>
-<polygon fill="#191970" stroke="#191970" points="447.8718,-87.902 439.4537,-81.4688 442.0226,-91.7475 447.8718,-87.902"/>
</g>
-<!-- Node35->Node18 -->
+<!-- Node35->Node37 -->
+<g id="edge80" class="edge">
+<title>Node35->Node37</title>
+<path fill="none" stroke="#191970" d="M1638.7106,-363.9005C1652.2322,-352.8554 1677.8652,-334.5776 1697.8148,-321.6172"/>
+<polygon fill="#191970" stroke="#191970" points="1699.9579,-324.4021 1706.5169,-316.0817 1696.2008,-318.4958 1699.9579,-324.4021"/>
+</g>
+<!-- Node35->Node40 -->
+<g id="edge86" class="edge">
+<title>Node35->Node40</title>
+<path fill="none" stroke="#191970" d="M1692.5303,-364.5073C1749.6274,-355.7045 1836.134,-340.8788 1910,-322 1913.348,-321.1443 1916.804,-320.1749 1920.2516,-319.1487"/>
+<polygon fill="#191970" stroke="#191970" points="1921.3453,-322.4739 1929.8482,-316.1531 1919.2595,-315.7919 1921.3453,-322.4739"/>
+</g>
+<!-- Node36->Node16 -->
<g id="edge74" class="edge">
-<title>Node35->Node18</title>
-<path fill="none" stroke="#191970" d="M860.1908,-300.3216C991.6956,-287.4638 1290.9537,-258.204 1391.9697,-248.3273"/>
-<polygon fill="#191970" stroke="#191970" points="1392.5927,-251.7832 1402.2047,-247.3266 1391.9115,-244.8164 1392.5927,-251.7832"/>
+<title>Node36->Node16</title>
+<path fill="none" stroke="#191970" d="M809.7255,-292.1681C766.4412,-281.6399 713.4996,-267.1201 694,-255 632.8642,-217.0008 602.7514,-130.2692 591.8952,-91.2098"/>
+<polygon fill="#191970" stroke="#191970" points="595.2235,-90.1054 589.2828,-81.3328 588.4562,-91.8953 595.2235,-90.1054"/>
</g>
-<!-- Node35->Node21 -->
+<!-- Node36->Node18 -->
<g id="edge75" class="edge">
-<title>Node35->Node21</title>
-<path fill="none" stroke="#191970" d="M770.091,-291.4554C752.7032,-281.734 730.1951,-269.15 712.9916,-259.5317"/>
-<polygon fill="#191970" stroke="#191970" points="714.4826,-256.3554 704.0461,-254.5303 711.0665,-262.4653 714.4826,-256.3554"/>
+<title>Node36->Node18</title>
+<path fill="none" stroke="#191970" d="M936.1406,-292.485C939.1301,-291.9544 942.0946,-291.4556 945,-291 1104.6761,-265.9627 1297.5914,-251.9624 1374.2414,-247.0413"/>
+<polygon fill="#191970" stroke="#191970" points="1374.5108,-250.5314 1384.2698,-246.407 1374.0688,-243.5453 1374.5108,-250.5314"/>
</g>
-<!-- Node36->Node9 -->
-<g id="edge82" class="edge">
-<title>Node36->Node9</title>
-<path fill="none" stroke="#191970" d="M1813.9634,-296.7895C1826.409,-287.0735 1845.2888,-271.2692 1859,-255 1903.5077,-202.1886 1882.096,-159.0595 1941,-123 2022.5935,-73.0506 2274.6793,-95.1496 2370,-87 2403.8848,-84.1029 2441.393,-80.8445 2473.2064,-78.0655"/>
-<polygon fill="#191970" stroke="#191970" points="2473.6663,-81.5387 2483.3235,-77.1811 2473.0567,-74.5653 2473.6663,-81.5387"/>
+<!-- Node36->Node21 -->
+<g id="edge76" class="edge">
+<title>Node36->Node21</title>
+<path fill="none" stroke="#191970" d="M846.3356,-291.4554C829.1059,-281.734 806.8025,-269.15 789.7554,-259.5317"/>
+<polygon fill="#191970" stroke="#191970" points="791.3204,-256.3961 780.8911,-254.5303 787.8806,-262.4926 791.3204,-256.3961"/>
</g>
-<!-- Node36->Node34 -->
+<!-- Node37->Node9 -->
<g id="edge83" class="edge">
-<title>Node36->Node34</title>
-<path fill="none" stroke="#191970" d="M1792.3112,-316.0817C1778.8026,-327.119 1753.1731,-345.3957 1733.2174,-358.3619"/>
-<polygon fill="#191970" stroke="#191970" points="1731.0702,-355.5796 1724.5118,-363.9005 1734.8277,-361.4857 1731.0702,-355.5796"/>
-</g>
-<!-- Node37 -->
-<g id="node38" class="node">
-<title>Node37</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1635,-235.5 1635,-254.5 1697,-254.5 1697,-235.5 1635,-235.5"/>
-<text text-anchor="middle" x="1666" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">dmlc/io.h</text>
+<title>Node37->Node9</title>
+<path fill="none" stroke="#191970" d="M1733.5159,-296.6462C1758.5649,-267.966 1834.8395,-182.7887 1908,-123 1921.6055,-111.8812 1937.6562,-100.9547 1951.6494,-92.0469"/>
+<polygon fill="#191970" stroke="#191970" points="1953.8841,-94.7772 1960.5015,-86.5031 1950.1686,-88.8446 1953.8841,-94.7772"/>
</g>
-<!-- Node36->Node37 -->
-<g id="edge80" class="edge">
-<title>Node36->Node37</title>
-<path fill="none" stroke="#191970" d="M1779.9063,-296.8906C1757.4173,-286.6457 1721.6082,-270.3326 1696.1476,-258.7339"/>
-<polygon fill="#191970" stroke="#191970" points="1697.4968,-255.5025 1686.9456,-254.5419 1694.5948,-261.8727 1697.4968,-255.5025"/>
+<!-- Node37->Node35 -->
+<g id="edge84" class="edge">
+<title>Node37->Node35</title>
+<path fill="none" stroke="#191970" d="M1716.3112,-316.0817C1702.8026,-327.119 1677.1731,-345.3957 1657.2174,-358.3619"/>
+<polygon fill="#191970" stroke="#191970" points="1655.0702,-355.5796 1648.5118,-363.9005 1658.8277,-361.4857 1655.0702,-355.5796"/>
</g>
<!-- Node38 -->
<g id="node39" class="node">
<title>Node38</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1753.5,-235.5 1753.5,-254.5 1850.5,-254.5 1850.5,-235.5 1753.5,-235.5"/>
-<text text-anchor="middle" x="1802" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">dmlc/serializer.h</text>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1526,-235.5 1526,-254.5 1588,-254.5 1588,-235.5 1526,-235.5"/>
+<text text-anchor="middle" x="1557" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">dmlc/io.h</text>
</g>
-<!-- Node36->Node38 -->
+<!-- Node37->Node38 -->
<g id="edge81" class="edge">
-<title>Node36->Node38</title>
-<path fill="none" stroke="#191970" d="M1801.1563,-296.8906C1801.2949,-288.3657 1801.5018,-275.6392 1801.676,-264.9235"/>
-<polygon fill="#191970" stroke="#191970" points="1805.1784,-264.8005 1801.8415,-254.7449 1798.1793,-264.6866 1805.1784,-264.8005"/>
+<title>Node37->Node38</title>
+<path fill="none" stroke="#191970" d="M1698.75,-296.8906C1670.2707,-286.4652 1624.6257,-269.7559 1592.8543,-258.1252"/>
+<polygon fill="#191970" stroke="#191970" points="1593.6594,-254.6929 1583.0657,-254.5419 1591.2531,-261.2663 1593.6594,-254.6929"/>
</g>
-<!-- Node40->Node8 -->
-<g id="edge119" class="edge">
-<title>Node40->Node8</title>
-<path fill="none" stroke="#191970" d="M1525.9205,-425.3739C1537.1638,-416.1292 1550.2679,-403.3426 1558,-389 1580.256,-347.7163 1598.5967,-205.4413 1604.7704,-152.7937"/>
-<polygon fill="#191970" stroke="#191970" points="1608.2625,-153.0624 1605.9296,-142.7276 1601.3084,-152.2615 1608.2625,-153.0624"/>
+<!-- Node39 -->
+<g id="node40" class="node">
+<title>Node39</title>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1644.5,-235.5 1644.5,-254.5 1741.5,-254.5 1741.5,-235.5 1644.5,-235.5"/>
+<text text-anchor="middle" x="1693" y="-242.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">dmlc/serializer.h</text>
</g>
-<!-- Node40->Node9 -->
-<g id="edge96" class="edge">
-<title>Node40->Node9</title>
-<path fill="none" stroke="#191970" d="M1563.0782,-435.705C1645.2286,-428.4147 1799.8649,-412.8221 1930,-389 1991.2337,-377.7908 2145.5222,-348.1814 2202,-322 2336.3933,-259.6995 2471.7094,-142.1148 2524.6893,-93.4466"/>
-<polygon fill="#191970" stroke="#191970" points="2527.221,-95.8725 2532.1862,-86.5131 2522.4682,-90.7334 2527.221,-95.8725"/>
+<!-- Node37->Node39 -->
+<g id="edge82" class="edge">
+<title>Node37->Node39</title>
+<path fill="none" stroke="#191970" d="M1720,-296.8906C1715.4242,-288.0965 1708.5219,-274.8312 1702.8417,-263.9145"/>
+<polygon fill="#191970" stroke="#191970" points="1705.7912,-262.0003 1698.0705,-254.7449 1699.5815,-265.2314 1705.7912,-262.0003"/>
</g>
-<!-- Node40->Node13 -->
-<g id="edge105" class="edge">
-<title>Node40->Node13</title>
-<path fill="none" stroke="#191970" d="M1446.6954,-438.5778C1205.1608,-430.5487 295.1778,-399.6165 285,-389 127.2395,-224.4395 289.8283,-230.6316 597,-123 683.2526,-92.7775 788.2502,-80.2626 856.9904,-75.0983"/>
-<polygon fill="#191970" stroke="#191970" points="857.5588,-78.5666 867.2825,-74.3597 857.0576,-71.5846 857.5588,-78.5666"/>
+<!-- Node41->Node8 -->
+<g id="edge120" class="edge">
+<title>Node41->Node8</title>
+<path fill="none" stroke="#191970" d="M1356.7211,-425.2756C1354.5571,-408.706 1351.9202,-381.4406 1354,-358 1358.9081,-302.6831 1362.2945,-288.8164 1376,-235 1383.4584,-205.7137 1395.0819,-172.6737 1402.6191,-152.3415"/>
+<polygon fill="#191970" stroke="#191970" points="1405.9985,-153.2981 1406.239,-142.706 1399.4457,-150.8362 1405.9985,-153.2981"/>
</g>
-<!-- Node40->Node14 -->
-<g id="edge123" class="edge">
-<title>Node40->Node14</title>
-<path fill="none" stroke="#191970" d="M1563.0216,-438.2359C1664.9203,-432.027 1880.0057,-408.806 2036,-322 2044.0445,-317.5235 2253.8575,-142.5727 2319.5364,-87.7519"/>
-<polygon fill="#191970" stroke="#191970" points="2322.0462,-90.216 2327.48,-81.1207 2317.5603,-84.8423 2322.0462,-90.216"/>
+<!-- Node41->Node9 -->
+<g id="edge97" class="edge">
+<title>Node41->Node9</title>
+<path fill="none" stroke="#191970" d="M1417.0433,-436.0806C1496.7523,-429.3572 1644.2967,-414.5628 1768,-389 1872.6861,-367.3671 1934.7973,-407.4703 1999,-322 2049.9117,-254.2235 2015.884,-144.6881 1996.6523,-96.0838"/>
+<polygon fill="#191970" stroke="#191970" points="1999.8303,-94.6081 1992.8089,-86.6739 1993.35,-97.255 1999.8303,-94.6081"/>
</g>
-<!-- Node40->Node15 -->
-<g id="edge125" class="edge">
-<title>Node40->Node15</title>
-<path fill="none" stroke="#191970" d="M1563.2842,-437.5807C1755.1789,-427.3961 2365.4239,-390.291 2549,-322 2688.8391,-269.9793 2734.47,-255.0401 2833,-143 2846.7276,-127.3901 2857.2753,-105.9569 2863.7364,-90.6364"/>
-<polygon fill="#191970" stroke="#191970" points="2867.066,-91.7371 2867.5454,-81.1532 2860.5703,-89.1281 2867.066,-91.7371"/>
+<!-- Node41->Node13 -->
+<g id="edge106" class="edge">
+<title>Node41->Node13</title>
+<path fill="none" stroke="#191970" d="M1300.9027,-439.1231C1087.9758,-433.8834 360.4784,-414.2058 318,-389 292.651,-373.9584 220.082,-227.9535 213,-199 210.8881,-190.3656 207.1814,-185.7198 213,-179 256.6825,-128.5522 705.4685,-90.0959 884.1654,-76.6714"/>
+<polygon fill="#191970" stroke="#191970" points="884.5502,-80.1525 894.262,-75.918 884.0292,-73.1719 884.5502,-80.1525"/>
+</g>
+<!-- Node41->Node14 -->
+<g id="edge124" class="edge">
+<title>Node41->Node14</title>
+<path fill="none" stroke="#191970" d="M1417.322,-435.8411C1543.597,-425.5909 1834.4745,-401.0405 1878,-389 2005.4538,-353.7423 2368.5237,-143.8761 2467.0631,-86.1739"/>
+<polygon fill="#191970" stroke="#191970" points="2468.9009,-89.1536 2475.7566,-81.0758 2465.3599,-83.1153 2468.9009,-89.1536"/>
</g>
-<!-- Node40->Node16 -->
+<!-- Node41->Node15 -->
<g id="edge126" class="edge">
-<title>Node40->Node16</title>
-<path fill="none" stroke="#191970" d="M1446.6612,-438.5879C1203.7919,-430.5661 284.2192,-399.5566 256,-389 160.0445,-353.1037 76,-347.45 76,-245 76,-245 76,-245 76,-189 76,-121.0477 312.318,-85.8885 400.4144,-75.1424"/>
-<polygon fill="#191970" stroke="#191970" points="400.8818,-78.6116 410.3972,-73.9524 400.0532,-71.6608 400.8818,-78.6116"/>
+<title>Node41->Node15</title>
+<path fill="none" stroke="#191970" d="M1417.13,-436.5217C1463.6281,-433.3542 1530.4944,-428.832 1589,-425 1983.7011,-399.148 2095.3672,-439.6891 2473,-322 2594.6611,-284.0844 2629.246,-274.37 2732,-199 2761.4129,-177.4257 2776.5331,-176.0355 2792,-143 2799.5615,-126.8495 2799.9053,-106.3506 2798.7457,-91.4466"/>
+<polygon fill="#191970" stroke="#191970" points="2802.2047,-90.8754 2797.6606,-81.3045 2795.2444,-91.6202 2802.2047,-90.8754"/>
</g>
-<!-- Node40->Node18 -->
+<!-- Node41->Node16 -->
<g id="edge127" class="edge">
-<title>Node40->Node18</title>
-<path fill="none" stroke="#191970" d="M1498.9322,-425.4841C1484.7972,-390.5046 1449.682,-303.6055 1433.9012,-264.553"/>
-<polygon fill="#191970" stroke="#191970" points="1436.9881,-262.8501 1429.9964,-254.8898 1430.498,-265.4728 1436.9881,-262.8501"/>
-</g>
-<!-- Node40->Node29 -->
-<g id="edge122" class="edge">
-<title>Node40->Node29</title>
-<path fill="none" stroke="#191970" d="M1446.9362,-436.3307C1325.7617,-427.3505 1053.9318,-405.6465 1015,-389 953.5897,-362.7422 905.0391,-295.9992 884.1439,-263.4344"/>
-<polygon fill="#191970" stroke="#191970" points="887.0508,-261.4817 878.7788,-254.8617 881.1171,-265.1953 887.0508,-261.4817"/>
-</g>
-<!-- Node40->Node31 -->
-<g id="edge97" class="edge">
-<title>Node40->Node31</title>
-<path fill="none" stroke="#191970" d="M1446.7446,-436.4685C1298.3347,-426.0842 915.8781,-398.5664 890,-389 867.9018,-380.831 868.4973,-367.6398 847,-358 839.581,-354.6732 753.3099,-335.9117 688.1222,-321.9881"/>
-<polygon fill="#191970" stroke="#191970" points="688.7734,-318.5484 678.2632,-319.8848 687.3129,-325.3943 688.7734,-318.5484"/>
+<title>Node41->Node16</title>
+<path fill="none" stroke="#191970" d="M1300.9139,-439.0578C1084.4797,-433.5089 334.9927,-412.6618 290,-389 182.3792,-332.402 123.6664,-188.6765 226,-123 280.0663,-88.3009 475.3002,-76.2514 553.8878,-72.7477"/>
+<polygon fill="#191970" stroke="#191970" points="554.3343,-76.2319 564.1759,-72.3086 554.0357,-69.2383 554.3343,-76.2319"/>
</g>
-<!-- Node40->Node32 -->
-<g id="edge104" class="edge">
-<title>Node40->Node32</title>
-<path fill="none" stroke="#191970" d="M1563.2865,-438.1208C1668.0775,-433.2212 1894.9975,-419.7729 2084,-389 2142.6258,-379.4547 2155.9816,-370.7263 2214,-358 2281.0892,-343.284 2358.4072,-328.0183 2410.5701,-317.9667"/>
-<polygon fill="#191970" stroke="#191970" points="2411.4974,-321.3527 2420.6568,-316.0277 2410.1759,-314.4785 2411.4974,-321.3527"/>
+<!-- Node41->Node18 -->
+<g id="edge128" class="edge">
+<title>Node41->Node18</title>
+<path fill="none" stroke="#191970" d="M1382.222,-425.322C1393.8101,-416.3119 1406.6771,-403.7675 1413,-389 1430.7319,-347.5858 1421.0409,-293.1751 1413.7349,-264.5496"/>
+<polygon fill="#191970" stroke="#191970" points="1417.0301,-263.3299 1411.0213,-254.6038 1410.277,-265.1725 1417.0301,-263.3299"/>
</g>
-<!-- Node40->Node34 -->
-<g id="edge118" class="edge">
-<title>Node40->Node34</title>
-<path fill="none" stroke="#191970" d="M1550.1083,-425.4639C1585.4576,-413.6808 1634.103,-397.4657 1667.6047,-386.2984"/>
-<polygon fill="#191970" stroke="#191970" points="1669.1041,-389.488 1677.4841,-383.0053 1666.8905,-382.8472 1669.1041,-389.488"/>
+<!-- Node41->Node29 -->
+<g id="edge123" class="edge">
+<title>Node41->Node29</title>
+<path fill="none" stroke="#191970" d="M1300.8191,-439.5213C1101.9852,-435.8947 456.9504,-421.7061 370,-389 347.4885,-380.5324 295.4969,-343.6399 285,-322 278.9869,-309.6037 278.7136,-303.26 285,-291 291.933,-277.4789 304.8936,-267.0326 317.3481,-259.5377"/>
+<polygon fill="#191970" stroke="#191970" points="319.1928,-262.5168 326.234,-254.6001 315.7928,-256.3979 319.1928,-262.5168"/>
</g>
-<!-- Node40->Node39 -->
-<g id="edge120" class="edge">
-<title>Node40->Node39</title>
-<path fill="none" stroke="#191970" d="M1563.2509,-432.4864C1618.794,-424.1398 1704.4836,-409.4176 1777,-389 1845.336,-369.7594 1922.3906,-338.1454 1964.1215,-320.1527"/>
-<polygon fill="#191970" stroke="#191970" points="1965.675,-323.294 1973.4543,-316.1014 1962.8876,-316.8729 1965.675,-323.294"/>
+<!-- Node41->Node32 -->
+<g id="edge98" class="edge">
+<title>Node41->Node32</title>
+<path fill="none" stroke="#191970" d="M1300.8926,-438.9784C1098.8154,-433.4801 436.9062,-413.6592 399,-389 379.2209,-376.1331 368.3147,-350.891 362.6094,-331.6995"/>
+<polygon fill="#191970" stroke="#191970" points="365.9199,-330.531 359.9516,-321.7771 359.1583,-332.3422 365.9199,-330.531"/>
</g>
-<!-- Node41 -->
-<g id="node42" class="node">
-<title>Node41</title>
-<g id="a_node42"><a xlink:href="map_8h.html" target="_top" xlink:title="Runtime Map container types. ">
-<polygon fill="#ffffff" stroke="#000000" points="294,-358.5 294,-388.5 420,-388.5 420,-358.5 294,-358.5"/>
-<text text-anchor="start" x="302" y="-376.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
-<text text-anchor="middle" x="357" y="-365.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/map.h</text>
-</a>
+<!-- Node41->Node33 -->
+<g id="edge105" class="edge">
+<title>Node41->Node33</title>
+<path fill="none" stroke="#191970" d="M1417.0983,-436.5722C1581.1236,-425.4457 2037.9682,-394.1841 2070,-389 2174.5152,-372.085 2294.5648,-337.5636 2355.3232,-318.9904"/>
+<polygon fill="#191970" stroke="#191970" points="2356.3774,-322.328 2364.9063,-316.0425 2354.3192,-315.6374 2356.3774,-322.328"/>
</g>
+<!-- Node41->Node35 -->
+<g id="edge119" class="edge">
+<title>Node41->Node35</title>
+<path fill="none" stroke="#191970" d="M1417.1363,-426.1268C1466.2551,-413.9831 1535.7712,-396.7964 1581.8307,-385.409"/>
+<polygon fill="#191970" stroke="#191970" points="1582.7058,-388.7981 1591.5734,-383.0003 1581.0257,-382.0027 1582.7058,-388.7981"/>
</g>
-<!-- Node40->Node41 -->
-<g id="edge98" class="edge">
-<title>Node40->Node41</title>
-<path fill="none" stroke="#191970" d="M1446.6245,-436.4065C1399.9284,-433.1747 1332.7732,-428.6221 1274,-425 960.3206,-405.6684 586.0908,-385.5769 430.4828,-377.3545"/>
-<polygon fill="#191970" stroke="#191970" points="430.2769,-373.8389 420.1062,-376.8067 429.9078,-380.8291 430.2769,-373.8389"/>
+<!-- Node41->Node40 -->
+<g id="edge121" class="edge">
+<title>Node41->Node40</title>
+<path fill="none" stroke="#191970" d="M1417.1992,-434.6817C1485.8731,-427.146 1603.2154,-412.1638 1702,-389 1796.5572,-366.8275 1817.6594,-352.0982 1910,-322 1912.757,-321.1014 1915.604,-320.1703 1918.4693,-319.2309"/>
+<polygon fill="#191970" stroke="#191970" points="1919.7702,-322.4876 1928.1775,-316.0403 1917.5846,-315.8375 1919.7702,-322.4876"/>
</g>
<!-- Node42 -->
<g id="node43" class="node">
<title>Node42</title>
-<g id="a_node43"><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="#000000" points="1326,-364 1326,-383 1452,-383 1452,-364 1326,-364"/>
-<text text-anchor="middle" x="1389" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/module.h</text>
+<g id="a_node43"><a xlink:href="map_8h.html" target="_top" xlink:title="Runtime Map container types. ">
+<polygon fill="#ffffff" stroke="#000000" points="408,-358.5 408,-388.5 534,-388.5 534,-358.5 408,-358.5"/>
+<text text-anchor="start" x="416" y="-376.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
+<text text-anchor="middle" x="471" y="-365.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">/map.h</text>
</a>
</g>
</g>
-<!-- Node40->Node42 -->
-<g id="edge106" class="edge">
-<title>Node40->Node42</title>
-<path fill="none" stroke="#191970" d="M1473.2268,-425.4639C1452.7656,-414.4247 1426.9205,-399.4954 1409.1881,-388.4715"/>
-<polygon fill="#191970" stroke="#191970" points="1411.0455,-385.5049 1400.7262,-383.1039 1407.2959,-391.416 1411.0455,-385.5049"/>
-</g>
-<!-- Node40->Node44 -->
-<g id="edge121" class="edge">
-<title>Node40->Node44</title>
-<path fill="none" stroke="#191970" d="M1446.9298,-436.4378C1292.3565,-425.5555 880.6494,-396.0911 852,-389 849.5147,-388.3849 846.9824,-387.6157 844.4704,-386.7512"/>
-<polygon fill="#191970" stroke="#191970" points="845.6146,-383.4407 835.0263,-383.0675 843.0708,-389.9621 845.6146,-383.4407"/>
+<!-- Node41->Node42 -->
+<g id="edge99" class="edge">
+<title>Node41->Node42</title>
+<path fill="none" stroke="#191970" d="M1300.7358,-436.1039C1144.4915,-424.3152 716.3898,-392.0148 544.178,-379.0213"/>
+<polygon fill="#191970" stroke="#191970" points="544.4303,-375.5305 534.1953,-378.2681 543.9036,-382.5107 544.4303,-375.5305"/>
</g>
-<!-- Node45 -->
-<g id="node46" class="node">
-<title>Node45</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1508.5,-364 1508.5,-383 1549.5,-383 1549.5,-364 1508.5,-364"/>
-<text text-anchor="middle" x="1529" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tuple</text>
+<!-- Node43 -->
+<g id="node44" class="node">
+<title>Node43</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="#000000" points="1180,-364 1180,-383 1306,-383 1306,-364 1180,-364"/>
+<text text-anchor="middle" x="1243" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/module.h</text>
+</a>
</g>
-<!-- Node40->Node45 -->
-<g id="edge124" class="edge">
-<title>Node40->Node45</title>
-<path fill="none" stroke="#191970" d="M1510.446,-425.2967C1513.9312,-415.5672 1518.4803,-402.8675 1522.2018,-392.4784"/>
-<polygon fill="#191970" stroke="#191970" points="1525.5,-393.6495 1525.5773,-383.055 1518.91,-391.2889 1525.5,-393.6495"/>
</g>
-<!-- Node41->Node16 -->
-<g id="edge101" class="edge">
-<title>Node41->Node16</title>
-<path fill="none" stroke="#191970" d="M337.4585,-358.3049C326.3916,-348.8554 313.0041,-335.887 304,-322 254.7726,-246.0769 205.6801,-194.6056 261,-123 278.0609,-100.9165 355.3977,-84.5309 400.3474,-76.6759"/>
-<polygon fill="#191970" stroke="#191970" points="400.9466,-80.1243 410.2181,-74.9969 399.7728,-73.2235 400.9466,-80.1243"/>
+<!-- Node41->Node43 -->
+<g id="edge107" class="edge">
+<title>Node41->Node43</title>
+<path fill="none" stroke="#191970" d="M1327.2268,-425.4639C1306.7656,-414.4247 1280.9205,-399.4954 1263.1881,-388.4715"/>
+<polygon fill="#191970" stroke="#191970" points="1265.0455,-385.5049 1254.7262,-383.1039 1261.2959,-391.416 1265.0455,-385.5049"/>
</g>
-<!-- Node41->Node20 -->
-<g id="edge103" class="edge">
-<title>Node41->Node20</title>
-<path fill="none" stroke="#191970" d="M365.6228,-358.2967C370.287,-350.0729 376.1547,-339.7272 381.4174,-330.4483"/>
-<polygon fill="#191970" stroke="#191970" points="384.5223,-332.0683 386.4113,-321.6432 378.4335,-328.6148 384.5223,-332.0683"/>
+<!-- Node41->Node45 -->
+<g id="edge122" class="edge">
+<title>Node41->Node45</title>
+<path fill="none" stroke="#191970" d="M1300.9656,-434.3647C1201.4445,-423.6958 1002.2523,-401.6572 933,-389 922.8028,-387.1363 911.7954,-384.6094 901.8781,-382.137"/>
+<polygon fill="#191970" stroke="#191970" points="902.5733,-378.7021 892.0175,-379.6103 900.8357,-385.483 902.5733,-378.7021"/>
</g>
-<!-- Node41->Node21 -->
-<g id="edge102" class="edge">
-<title>Node41->Node21</title>
-<path fill="none" stroke="#191970" d="M341.8899,-358.196C326.0498,-340.3023 305.6868,-310.6388 323,-291 344.3517,-266.7802 560.2185,-252.0699 649.0858,-246.9986"/>
-<polygon fill="#191970" stroke="#191970" points="649.4835,-250.4819 659.2721,-246.428 649.0919,-243.4929 649.4835,-250.4819"/>
+<!-- Node46 -->
+<g id="node47" class="node">
+<title>Node46</title>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1362.5,-364 1362.5,-383 1403.5,-383 1403.5,-364 1362.5,-364"/>
+<text text-anchor="middle" x="1383" y="-371" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tuple</text>
</g>
-<!-- Node41->Node24 -->
-<g id="edge99" class="edge">
-<title>Node41->Node24</title>
-<path fill="none" stroke="#191970" d="M338.6408,-358.2245C320.0216,-340.884 296.1691,-312.2507 313,-291 316.4637,-286.6267 698.5711,-200.285 704,-199 704.1003,-198.9763 704.2007,-198.9525 704.3013,-198.9287"/>
-<polygon fill="#191970" stroke="#191970" points="704.9761,-202.3658 713.8985,-196.6525 703.3607,-195.5547 704.9761,-202.3658"/>
+<!-- Node41->Node46 -->
+<g id="edge125" class="edge">
+<title>Node41->Node46</title>
+<path fill="none" stroke="#191970" d="M1364.446,-425.2967C1367.9312,-415.5672 1372.4803,-402.8675 1376.2018,-392.4784"/>
+<polygon fill="#191970" stroke="#191970" points="1379.5,-393.6495 1379.5773,-383.055 1372.91,-391.2889 1379.5,-393.6495"/>
</g>
-<!-- Node41->Node30 -->
-<g id="edge100" class="edge">
-<title>Node41->Node30</title>
-<path fill="none" stroke="#191970" d="M388.8001,-358.3521C433.4805,-337.2504 512.7584,-300.5204 543,-291 559.8896,-285.683 658.0617,-267.207 722.3515,-255.3509"/>
-<polygon fill="#191970" stroke="#191970" points="723.1601,-258.7609 732.361,-253.5079 721.8925,-251.8766 723.1601,-258.7609"/>
+<!-- Node42->Node16 -->
+<g id="edge102" class="edge">
+<title>Node42->Node16</title>
+<path fill="none" stroke="#191970" d="M407.8023,-363.1689C355.2051,-353.6308 285.301,-338.4825 262,-322 205.59,-282.0972 171.7382,-241.5945 201,-179 217.4355,-143.8425 232.6934,-139.1126 268,-123 318.7282,-99.8495 483.0341,-81.5983 553.9318,-74.5975"/>
+<polygon fill="#191970" stroke="#191970" points="554.4701,-78.0617 564.0838,-73.6091 553.7917,-71.0947 554.4701,-78.0617"/>
</g>
-<!-- Node42->Node8 -->
-<g id="edge111" class="edge">
-<title>Node42->Node8</title>
-<path fill="none" stroke="#191970" d="M1396.4994,-363.7773C1419.1519,-334.6224 1489.3552,-245.8201 1555,-179 1565.1627,-168.6553 1577.2584,-157.8739 1587.2256,-149.3511"/>
-<polygon fill="#191970" stroke="#191970" points="1589.6238,-151.907 1595.0082,-142.7824 1585.1088,-146.5577 1589.6238,-151.907"/>
+<!-- Node42->Node20 -->
+<g id="edge104" class="edge">
+<title>Node42->Node20</title>
+<path fill="none" stroke="#191970" d="M503.3164,-358.4639C523.9225,-348.8763 550.8351,-336.3545 573.0783,-326.0052"/>
+<polygon fill="#191970" stroke="#191970" points="574.7031,-329.1096 582.2932,-321.7177 571.7501,-322.763 574.7031,-329.1096"/>
</g>
-<!-- Node42->Node9 -->
-<g id="edge108" class="edge">
-<title>Node42->Node9</title>
-<path fill="none" stroke="#191970" d="M1452.0501,-364.1185C1467.3956,-361.9726 1483.7734,-359.7969 1499,-358 1582.7986,-348.1107 1800.111,-356.7575 1877,-322 1895.3366,-313.711 1893.5786,-302.6308 1910,-291 1954.1229,-259.749 1968.8403,-257.5117 2018,-235 2143.2452,-177.6464 2174.8047,-161.7323 2307,-123 2362.0068,-106.8834 2425.4429,-93.531 2473.4213,-84.4635"/>
-<polygon fill="#191970" stroke="#191970" points="2474.23,-87.873 2483.4161,-82.5942 2472.9431,-80.9923 2474.23,-87.873"/>
+<!-- Node42->Node21 -->
+<g id="edge103" class="edge">
+<title>Node42->Node21</title>
+<path fill="none" stroke="#191970" d="M479.8453,-358.343C491.7107,-339.443 514.6113,-307.6043 543,-291 573.6683,-273.0624 671.755,-257.4587 726.3764,-249.881"/>
+<polygon fill="#191970" stroke="#191970" points="726.9313,-253.3378 736.366,-248.5174 725.9845,-246.4021 726.9313,-253.3378"/>
</g>
-<!-- Node42->Node14 -->
-<g id="edge114" class="edge">
-<title>Node42->Node14</title>
-<path fill="none" stroke="#191970" d="M1395.694,-363.7396C1402.6817,-353.4644 1413.8726,-336.7624 1423,-322 1436.4299,-300.2788 1516.3022,-136.4678 1538,-123 1604.1331,-81.9513 2164.9723,-73.3258 2306.6126,-71.7954"/>
-<polygon fill="#191970" stroke="#191970" points="2306.8498,-75.2932 2316.813,-71.6898 2306.7773,-68.2936 2306.8498,-75.2932"/>
+<!-- Node42->Node24 -->
+<g id="edge100" class="edge">
+<title>Node42->Node24</title>
+<path fill="none" stroke="#191970" d="M407.8293,-363.105C360.8505,-354.1119 302.0525,-339.7685 285,-322 255.2896,-291.042 252.0934,-237.3863 252.7718,-208.5964"/>
+<polygon fill="#191970" stroke="#191970" points="256.2706,-208.6953 253.1938,-198.5571 249.2768,-208.4013 256.2706,-208.6953"/>
</g>
-<!-- Node42->Node18 -->
-<g id="edge116" class="edge">
-<title>Node42->Node18</title>
-<path fill="none" stroke="#191970" d="M1391.781,-363.8416C1397.8344,-342.8184 1412.3057,-292.5598 1420.3553,-264.6038"/>
-<polygon fill="#191970" stroke="#191970" points="1423.783,-265.3485 1423.1867,-254.7705 1417.0563,-263.4116 1423.783,-265.3485"/>
+<!-- Node42->Node31 -->
+<g id="edge101" class="edge">
+<title>Node42->Node31</title>
+<path fill="none" stroke="#191970" d="M470.3618,-358.2473C469.8592,-341.6535 469.5925,-314.3676 472,-291 472.8944,-282.3185 474.6191,-272.8374 476.3213,-264.7505"/>
+<polygon fill="#191970" stroke="#191970" points="479.7813,-265.3159 478.5486,-254.793 472.9501,-263.7878 479.7813,-265.3159"/>
</g>
-<!-- Node42->Node22 -->
-<g id="edge110" class="edge">
-<title>Node42->Node22</title>
-<path fill="none" stroke="#191970" d="M1325.974,-369.1111C1162.9146,-357.691 738.8619,-327.5503 725,-322 704.7658,-313.8982 704.5324,-303.949 687,-291 663.7979,-273.8635 646.9941,-280.2223 633,-255 625.1885,-240.9209 625.3464,-222.4239 627.0804,-208.5947"/>
-<polygon fill="#191970" stroke="#191970" points="630.5426,-209.1093 628.6868,-198.6782 623.6327,-207.9899 630.5426,-209.1093"/>
+<!-- Node43->Node8 -->
+<g id="edge112" class="edge">
+<title>Node43->Node8</title>
+<path fill="none" stroke="#191970" d="M1231.9299,-363.8603C1215.013,-347.9414 1186.3551,-315.4849 1203,-291 1227.4524,-255.0301 1260.5133,-282.756 1294,-255 1324.7584,-229.5054 1314.1983,-207.6899 1342,-179 1353.9273,-166.6917 1369.6532,-155.7969 1382.9774,-147.7034"/>
+<polygon fill="#191970" stroke="#191970" points="1384.8855,-150.6419 1391.7419,-142.5648 1381.345,-144.6032 1384.8855,-150.6419"/>
</g>
-<!-- Node42->Node26 -->
+<!-- Node43->Node9 -->
<g id="edge109" class="edge">
-<title>Node42->Node26</title>
-<path fill="none" stroke="#191970" d="M1341.5007,-363.9717C1285.8549,-352.8092 1193.0363,-334.1899 1128.1734,-321.1785"/>
-<polygon fill="#191970" stroke="#191970" points="1128.5064,-317.6756 1118.0133,-319.1404 1127.1296,-324.5389 1128.5064,-317.6756"/>
-</g>
-<!-- Node42->Node29 -->
-<g id="edge112" class="edge">
-<title>Node42->Node29</title>
-<path fill="none" stroke="#191970" d="M1325.9494,-369.316C1227.9447,-362.1738 1043.9799,-346.0318 983,-322 961.081,-313.3618 959.2682,-304.5573 940,-291 925.28,-280.6429 908.594,-269.186 895.5562,-260.2969"/>
-<polygon fill="#191970" stroke="#191970" points="897.4183,-257.3306 887.1818,-254.5982 893.4802,-263.1177 897.4183,-257.3306"/>
+<title>Node43->Node9</title>
+<path fill="none" stroke="#191970" d="M1273.638,-363.9455C1306.0828,-353.7702 1358.7607,-337.0854 1404,-322 1490.1367,-293.2771 1513.51,-290.6926 1597,-255 1614.9115,-247.3427 1618.5059,-243.5682 1636,-235 1745.9354,-181.1563 1877.0271,-120.9776 1943.222,-90.8685"/>
+<polygon fill="#191970" stroke="#191970" points="1944.8515,-93.9725 1952.5073,-86.6487 1941.9553,-87.5998 1944.8515,-93.9725"/>
</g>
-<!-- Node42->Node30 -->
+<!-- Node43->Node14 -->
<g id="edge115" class="edge">
-<title>Node42->Node30</title>
-<path fill="none" stroke="#191970" d="M1325.7932,-368.846C1219.8033,-360.6083 1010.8793,-342.3934 940,-322 926.332,-318.0674 849.4468,-280.1648 807.0595,-259.0444"/>
-<polygon fill="#191970" stroke="#191970" points="808.5871,-255.8952 798.0764,-254.5623 805.4618,-262.1588 808.5871,-255.8952"/>
-</g>
-<!-- Node42->Node37 -->
-<g id="edge107" class="edge">
-<title>Node42->Node37</title>
-<path fill="none" stroke="#191970" d="M1419.5628,-363.9584C1448.5228,-354.5406 1492.9733,-339.1017 1530,-322 1570.7636,-303.1723 1615.657,-276.5095 1642.3064,-260.0092"/>
-<polygon fill="#191970" stroke="#191970" points="1644.3033,-262.8888 1650.9351,-254.6262 1640.5981,-256.9497 1644.3033,-262.8888"/>
+<title>Node43->Node14</title>
+<path fill="none" stroke="#191970" d="M1305.719,-363.9685C1321.456,-361.7901 1338.3244,-359.6372 1354,-358 1468.7645,-346.0138 1763.2776,-360.6523 1872,-322 1894.5397,-313.9868 1895.022,-302.4966 1916,-291 2088.6936,-196.3588 2140.6678,-189.5734 2326,-123 2372.0668,-106.4523 2426.4062,-90.1971 2460.0367,-80.5081"/>
+<polygon fill="#191970" stroke="#191970" points="2461.1267,-83.8367 2469.7768,-77.7189 2459.1996,-77.1072 2461.1267,-83.8367"/>
</g>
-<!-- Node42->Node40 -->
+<!-- Node43->Node18 -->
<g id="edge117" class="edge">
-<title>Node42->Node40</title>
-<path fill="none" stroke="#191970" d="M1410.5291,-383.1039C1429.0892,-392.7457 1455.7723,-407.8959 1476.1316,-420.2103"/>
-<polygon fill="#191970" stroke="#191970" points="1474.3523,-423.2248 1484.7079,-425.4639 1478.0089,-417.2557 1474.3523,-423.2248"/>
+<title>Node43->Node18</title>
+<path fill="none" stroke="#191970" d="M1248.9785,-363.9544C1259.8043,-347.3011 1284.1331,-312.6526 1312,-291 1330.9764,-276.2554 1355.443,-264.6215 1374.9266,-256.7423"/>
+<polygon fill="#191970" stroke="#191970" points="1376.2822,-259.9704 1384.3247,-253.0734 1373.7365,-253.4497 1376.2822,-259.9704"/>
</g>
-<!-- Node43 -->
-<g id="node44" class="node">
-<title>Node43</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="1326,-297 1326,-316 1374,-316 1374,-297 1326,-297"/>
-<text text-anchor="middle" x="1350" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">mutex</text>
+<!-- Node43->Node22 -->
+<g id="edge111" class="edge">
+<title>Node43->Node22</title>
+<path fill="none" stroke="#191970" d="M1202.7873,-363.9885C1192.4702,-361.7838 1181.3599,-359.6136 1171,-358 1007.7475,-332.572 952.1035,-388.826 801,-322 760.4244,-304.0553 744.9897,-295.5557 727,-255 717.783,-234.2212 738.8124,-215.7174 758.6473,-203.6287"/>
+<polygon fill="#191970" stroke="#191970" points="760.613,-206.5369 767.5637,-198.5407 757.1436,-200.4571 760.613,-206.5369"/>
</g>
-<!-- Node42->Node43 -->
+<!-- Node43->Node26 -->
+<g id="edge110" class="edge">
+<title>Node43->Node26</title>
+<path fill="none" stroke="#191970" d="M1210.6197,-363.9005C1175.9706,-353.6284 1120.2199,-337.1006 1077.4202,-324.4122"/>
+<polygon fill="#191970" stroke="#191970" points="1078.2528,-321.0085 1067.6704,-321.5218 1076.2631,-327.7198 1078.2528,-321.0085"/>
+</g>
+<!-- Node43->Node29 -->
<g id="edge113" class="edge">
-<title>Node42->Node43</title>
-<path fill="none" stroke="#191970" d="M1383.4123,-363.9005C1377.5043,-353.751 1368.0413,-337.4941 1360.6918,-324.8679"/>
-<polygon fill="#191970" stroke="#191970" points="1363.633,-322.9634 1355.5774,-316.0817 1357.5833,-326.4849 1363.633,-322.9634"/>
+<title>Node43->Node29</title>
+<path fill="none" stroke="#191970" d="M1203.958,-363.9446C1193.3257,-361.658 1181.77,-359.459 1171,-358 893.9613,-320.4691 815.9797,-382.3416 543,-322 481.1257,-308.3228 413.259,-277.4014 376.4587,-259.1768"/>
+<polygon fill="#191970" stroke="#191970" points="377.8718,-255.9701 367.3637,-254.6166 374.7342,-262.2276 377.8718,-255.9701"/>
</g>
-<!-- Node46->Node6 -->
-<g id="edge132" class="edge">
-<title>Node46->Node6</title>
-<path fill="none" stroke="#191970" d="M2489.9526,-492.4383C2429.9916,-460.6704 2237.1925,-358.5234 2165.8628,-320.7322"/>
-<polygon fill="#191970" stroke="#191970" points="2167.4877,-317.6322 2157.0127,-316.0433 2164.2105,-323.8177 2167.4877,-317.6322"/>
+<!-- Node43->Node31 -->
+<g id="edge116" class="edge">
+<title>Node43->Node31</title>
+<path fill="none" stroke="#191970" d="M1203.1813,-363.9575C1192.758,-361.7267 1181.4986,-359.5518 1171,-358 976.4616,-329.2454 914.5234,-389.9935 730,-322 707.8933,-313.8541 708.4913,-300.6533 687,-291 660.7543,-279.2111 589.1967,-264.5768 537.7914,-255.0236"/>
+<polygon fill="#191970" stroke="#191970" points="538.267,-251.5524 527.7984,-253.183 536.999,-258.4366 538.267,-251.5524"/>
</g>
-<!-- Node47 -->
-<g id="node48" class="node">
-<title>Node47</title>
-<polygon fill="#ffffff" stroke="#bfbfbf" points="2514,-431 2514,-450 2574,-450 2574,-431 2514,-431"/>
-<text text-anchor="middle" x="2544" y="-438" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">iostream</text>
+<!-- Node43->Node38 -->
+<g id="edge108" class="edge">
+<title>Node43->Node38</title>
+<path fill="none" stroke="#191970" d="M1262.7971,-363.8987C1296.1105,-347.8958 1365.6558,-315.1453 1426,-291 1456.5602,-278.7721 1491.7916,-266.5008 1518.0155,-257.7116"/>
+<polygon fill="#191970" stroke="#191970" points="1519.2146,-261.0013 1527.5965,-254.5209 1517.0028,-254.3599 1519.2146,-261.0013"/>
</g>
-<!-- Node46->Node47 -->
+<!-- Node43->Node41 -->
+<g id="edge118" class="edge">
+<title>Node43->Node41</title>
+<path fill="none" stroke="#191970" d="M1264.5291,-383.1039C1283.0892,-392.7457 1309.7723,-407.8959 1330.1316,-420.2103"/>
+<polygon fill="#191970" stroke="#191970" points="1328.3523,-423.2248 1338.7079,-425.4639 1332.0089,-417.2557 1328.3523,-423.2248"/>
+</g>
+<!-- Node44 -->
+<g id="node45" class="node">
+<title>Node44</title>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="1212,-297 1212,-316 1260,-316 1260,-297 1212,-297"/>
+<text text-anchor="middle" x="1236" y="-304" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">mutex</text>
+</g>
+<!-- Node43->Node44 -->
+<g id="edge114" class="edge">
+<title>Node43->Node44</title>
+<path fill="none" stroke="#191970" d="M1241.9971,-363.9005C1240.9783,-354.149 1239.3704,-338.7597 1238.0759,-326.3695"/>
+<polygon fill="#191970" stroke="#191970" points="1241.5213,-325.6638 1237.0011,-316.0817 1234.5592,-326.3912 1241.5213,-325.6638"/>
+</g>
+<!-- Node47->Node6 -->
<g id="edge133" class="edge">
-<title>Node46->Node47</title>
-<path fill="none" stroke="#191970" d="M2513.625,-492.3906C2518.8253,-483.5068 2526.6966,-470.0601 2533.1233,-459.0811"/>
-<polygon fill="#191970" stroke="#191970" points="2536.2644,-460.6432 2538.2957,-450.2449 2530.2233,-457.1069 2536.2644,-460.6432"/>
+<title>Node47->Node6</title>
+<path fill="none" stroke="#191970" d="M2311.9582,-492.3051C2289.4824,-461.3613 2219.2251,-364.6339 2189.9781,-324.3677"/>
+<polygon fill="#191970" stroke="#191970" points="2192.793,-322.2875 2184.0843,-316.2534 2187.1293,-326.4012 2192.793,-322.2875"/>
</g>
-<!-- Node48->Node2 -->
-<g id="edge146" class="edge">
-<title>Node48->Node2</title>
-<path fill="none" stroke="#191970" d="M2650.3673,-660.2455C2660.9554,-651.8943 2676.4064,-639.7076 2688.7893,-629.9408"/>
-<polygon fill="#191970" stroke="#191970" points="2691.0903,-632.5837 2696.7745,-623.6427 2686.7553,-627.0875 2691.0903,-632.5837"/>
+<!-- Node48 -->
+<g id="node49" class="node">
+<title>Node48</title>
+<polygon fill="#ffffff" stroke="#bfbfbf" points="2499,-431 2499,-450 2559,-450 2559,-431 2499,-431"/>
+<text text-anchor="middle" x="2529" y="-438" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">iostream</text>
</g>
-<!-- Node48->Node3 -->
+<!-- Node47->Node48 -->
+<g id="edge134" class="edge">
+<title>Node47->Node48</title>
+<path fill="none" stroke="#191970" d="M2356.1341,-492.4878C2389.9171,-483.6499 2441.0644,-469.8084 2485,-456 2487.5908,-455.1858 2490.2577,-454.3189 2492.9364,-453.4272"/>
+<polygon fill="#191970" stroke="#191970" points="2494.2397,-456.6807 2502.5735,-450.1386 2491.979,-450.0558 2494.2397,-456.6807"/>
+</g>
+<!-- Node49->Node2 -->
<g id="edge147" class="edge">
-<title>Node48->Node3</title>
-<path fill="none" stroke="#191970" d="M2627.1721,-660.4509C2605.4276,-641.2747 2556.2395,-597.8962 2529.3683,-574.1988"/>
-<polygon fill="#191970" stroke="#191970" points="2531.6157,-571.5141 2521.8005,-567.5249 2526.9857,-576.7642 2531.6157,-571.5141"/>
+<title>Node49->Node2</title>
+<path fill="none" stroke="#191970" d="M2477.0972,-662.9667C2528.1754,-653.9932 2619.1683,-637.9606 2697,-624 2698.4628,-623.7376 2699.9473,-623.4709 2701.4458,-623.2012"/>
+<polygon fill="#191970" stroke="#191970" points="2702.2358,-626.6152 2711.4549,-621.3943 2700.9922,-619.7266 2702.2358,-626.6152"/>
</g>
-<!-- Node48->Node8 -->
-<g id="edge150" class="edge">
-<title>Node48->Node8</title>
-<path fill="none" stroke="#191970" d="M2673.9375,-660.4944C2738.625,-642.3184 2868,-600.6367 2868,-558 2868,-558 2868,-558 2868,-440.5 2868,-193.2546 1941.4267,-143.9382 1676.798,-134.8952"/>
-<polygon fill="#191970" stroke="#191970" points="1676.6611,-131.3888 1666.5502,-134.554 1676.4281,-138.385 1676.6611,-131.3888"/>
+<!-- Node49->Node3 -->
+<g id="edge148" class="edge">
+<title>Node49->Node3</title>
+<path fill="none" stroke="#191970" d="M2422.9322,-660.4509C2394.3162,-641.0268 2329.117,-596.7703 2294.5247,-573.2895"/>
+<polygon fill="#191970" stroke="#191970" points="2296.2718,-570.2453 2286.0322,-567.5249 2292.3404,-576.037 2296.2718,-570.2453"/>
</g>
-<!-- Node48->Node14 -->
+<!-- Node49->Node8 -->
<g id="edge151" class="edge">
-<title>Node48->Node14</title>
-<path fill="none" stroke="#191970" d="M2678.3294,-664.2864C2716.035,-657.909 2773.2363,-645.5392 2819,-624 2902.1671,-584.8565 2986,-593.9184 2986,-502 2986,-502 2986,-502 2986,-373.5 2986,-189.7174 2830.7458,-182.9035 2657,-123 2604.3195,-104.837 2440.9736,-83.746 2371.1657,-75.2903"/>
-<polygon fill="#191970" stroke="#191970" points="2371.5284,-71.8088 2361.1821,-74.0901 2370.6929,-78.7588 2371.5284,-71.8088"/>
+<title>Node49->Node8</title>
+<path fill="none" stroke="#191970" d="M2475.984,-660.4794C2576.0765,-633.9856 2829.5275,-552.8322 2760,-425 2681.8822,-281.3739 2598.5713,-291.0957 2445,-235 2338.8618,-196.2304 2308.8881,-194.794 2197,-179 2059.5287,-159.5947 1645.6793,-142.0566 1479.8428,-135.6135"/>
+<polygon fill="#191970" stroke="#191970" points="1479.8851,-132.1126 1469.7573,-135.2235 1479.6145,-139.1074 1479.8851,-132.1126"/>
</g>
-<!-- Node48->Node31 -->
-<g id="edge148" class="edge">
-<title>Node48->Node31</title>
-<path fill="none" stroke="#191970" d="M2597.8083,-668.6362C2397.2766,-660.8664 1486.7408,-614.1792 785,-389 735.7469,-373.1954 682.6749,-345.5624 649.1728,-326.6675"/>
-<polygon fill="#191970" stroke="#191970" points="650.6827,-323.4994 640.2619,-321.5869 647.2156,-329.5805 650.6827,-323.4994"/>
+<!-- Node49->Node14 -->
+<g id="edge152" class="edge">
+<title>Node49->Node14</title>
+<path fill="none" stroke="#191970" d="M2477.3277,-669.0338C2591.7123,-665.2098 2910,-646.6082 2910,-558 2910,-558 2910,-558 2910,-245 2910,-202.8832 2876.751,-202.7949 2842,-179 2739.4054,-108.751 2589.2478,-83.0927 2524.1135,-74.8964"/>
+<polygon fill="#191970" stroke="#191970" points="2524.3953,-71.4052 2514.0487,-73.6853 2523.5589,-78.355 2524.3953,-71.4052"/>
</g>
-<!-- Node48->Node32 -->
+<!-- Node49->Node32 -->
<g id="edge149" class="edge">
-<title>Node48->Node32</title>
-<path fill="none" stroke="#191970" d="M2638.2866,-660.4127C2638.8396,-641.0567 2640,-595.9227 2640,-558 2640,-558 2640,-558 2640,-502 2640,-466.154 2636.059,-454.7082 2616,-425 2584.7051,-378.6509 2529.9973,-341.0856 2497.4715,-321.4269"/>
-<polygon fill="#191970" stroke="#191970" points="2498.9203,-318.2174 2488.5317,-316.1375 2495.3558,-324.2419 2498.9203,-318.2174"/>
+<title>Node49->Node32</title>
+<path fill="none" stroke="#191970" d="M2396.708,-665.123C2212.7969,-642.7816 1434.6334,-547.3766 798,-456 607.4143,-428.645 506.2142,-525.0768 370,-389 355.251,-374.2658 352.9686,-350.0582 353.8017,-331.6663"/>
+<polygon fill="#191970" stroke="#191970" points="357.2966,-331.8661 354.5816,-321.6251 350.3177,-331.324 357.2966,-331.8661"/>
+</g>
+<!-- Node49->Node33 -->
+<g id="edge150" class="edge">
+<title>Node49->Node33</title>
+<path fill="none" stroke="#191970" d="M2472.9287,-660.4171C2517.3049,-646.149 2587,-614.7126 2587,-558 2587,-558 2587,-558 2587,-502 2587,-466.7513 2588.0792,-453.9706 2568,-425 2532.632,-373.9707 2467.6073,-338.3399 2428.1374,-320.2386"/>
+<polygon fill="#191970" stroke="#191970" points="2429.2447,-316.8999 2418.6869,-316.0156 2426.3888,-323.2909 2429.2447,-316.8999"/>
</g>
</g>
</svg>
diff --git a/docs/reference/api/doxygen/algorithm_8h__incl.svg b/docs/reference/api/doxygen/algorithm_8h__incl.svg
index 29d2d61a5..ca1de438c 100644
--- a/docs/reference/api/doxygen/algorithm_8h__incl.svg
+++ b/docs/reference/api/doxygen/algorithm_8h__incl.svg
@@ -43,33 +43,33 @@
<path fill="none" stroke="#191970" d="M2903.005,-1019.8214C3156.7175,-1008.7886 4078,-964.6216 4078,-905 4078,-905 4078,-905 4078,-133 4078,-91.201 4048.7446,-84.9584 4011,-67 3934.0113,-30.3697 3670.5759,-19.1837 3578.2682,-16.3442"/>
<polygon fill="#191970" stroke="#191970" points="3578.2296,-12.8417 3568.1307,-16.0451 3578.0231,-19.8386 3578.2296,-12.8417"/>
</g>
-<!-- Node51 -->
+<!-- Node52 -->
<g id="node36" class="node">
-<title>Node51</title>
+<title>Node52</title>
<g id="a_node36"><a xlink:href="relay_2base_8h.html" target="_top" xlink:title="Base classes for the Relay IR. ">
<polygon fill="#ffffff" stroke="#000000" points="3650.5,-839.5 3650.5,-858.5 3749.5,-858.5 3749.5,-839.5 3650.5,-839.5"/>
<text text-anchor="middle" x="3700" y="-846.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/relay/base.h</text>
</a>
</g>
</g>
-<!-- Node0->Node51 -->
+<!-- Node0->Node52 -->
<g id="edge133" class="edge">
-<title>Node0->Node51</title>
+<title>Node0->Node52</title>
<path fill="none" stroke="#191970" d="M2903.0115,-1009.601C3064.7336,-977.0125 3487.8339,-891.7536 3642.3978,-860.6074"/>
<polygon fill="#191970" stroke="#191970" points="3643.5088,-863.954 3652.6203,-858.5475 3642.1259,-857.0919 3643.5088,-863.954"/>
</g>
-<!-- Node55 -->
+<!-- Node56 -->
<g id="node40" class="node">
-<title>Node55</title>
+<title>Node56</title>
<g id="a_node40"><a xlink:href="relay_2expr_8h.html" target="_top" xlink:title="Relay expression language. ">
<polygon fill="#ffffff" stroke="#000000" points="1905.5,-951.5 1905.5,-970.5 2002.5,-970.5 2002.5,-951.5 1905.5,-951.5"/>
<text text-anchor="middle" x="1954" y="-958.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/relay/expr.h</text>
</a>
</g>
</g>
-<!-- Node0->Node55 -->
+<!-- Node0->Node56 -->
<g id="edge164" class="edge">
-<title>Node0->Node55</title>
+<title>Node0->Node56</title>
<path fill="none" stroke="#191970" d="M2774.997,-1018.0523C2609.974,-1006.5846 2171.863,-976.1396 2012.7758,-965.0844"/>
<polygon fill="#191970" stroke="#191970" points="2012.925,-961.5864 2002.7064,-964.3847 2012.4397,-968.5696 2012.925,-961.5864"/>
</g>
@@ -158,21 +158,21 @@
<path fill="none" stroke="#191970" d="M1545.01,-660.2017C1569.4262,-612.1526 1677.3878,-396.2315 1742,-210 1748.6302,-190.8899 1754.2199,-168.4445 1757.7897,-152.7342"/>
<polygon fill="#191970" stroke="#191970" points="1761.2824,-153.1482 1760.0181,-142.629 1754.4467,-151.6407 1761.2824,-153.1482"/>
</g>
-<!-- Node32 -->
+<!-- Node33 -->
<g id="node22" class="node">
-<title>Node32</title>
+<title>Node33</title>
<polygon fill="#ffffff" stroke="#bfbfbf" points="401.5,-123.5 401.5,-142.5 494.5,-142.5 494.5,-123.5 401.5,-123.5"/>
<text text-anchor="middle" x="448" y="-130.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">unordered_map</text>
</g>
-<!-- Node1->Node32 -->
+<!-- Node1->Node33 -->
<g id="edge130" class="edge">
-<title>Node1->Node32</title>
+<title>Node1->Node33</title>
<path fill="none" stroke="#191970" d="M1499.5252,-668.9139C1355.6115,-664.4375 870.2248,-643.9061 738,-568 566.5201,-469.5588 476.8868,-223.659 453.9779,-152.5477"/>
<polygon fill="#191970" stroke="#191970" points="457.2141,-151.1716 450.8717,-142.6849 450.5374,-153.2744 457.2141,-151.1716"/>
</g>
-<!-- Node35 -->
+<!-- Node36 -->
<g id="node25" class="node">
-<title>Node35</title>
+<title>Node36</title>
<g id="a_node25"><a xlink:href="structural__hash_8h.html" target="_top" xlink:title="tvm/node/structural\l_hash.h">
<polygon fill="#ffffff" stroke="#000000" points="2699.5,-313.5 2699.5,-343.5 2812.5,-343.5 2812.5,-313.5 2699.5,-313.5"/>
<text text-anchor="start" x="2707.5" y="-331.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/structural</text>
@@ -180,27 +180,27 @@
</a>
</g>
</g>
-<!-- Node1->Node35 -->
+<!-- Node1->Node36 -->
<g id="edge125" class="edge">
-<title>Node1->Node35</title>
+<title>Node1->Node36</title>
<path fill="none" stroke="#191970" d="M1580.3516,-661.3331C1621.5749,-652.3777 1687.3868,-637.8006 1744,-624 2064.0664,-545.9776 2142.1362,-518.7999 2461,-436 2523.1841,-419.8525 2541.7821,-424.9167 2601,-400 2616.8077,-393.3487 2618.7263,-387.7999 2634,-380 2657.4695,-368.0147 2684.3809,-356.4374 2706.9467,-347.3057"/>
<polygon fill="#191970" stroke="#191970" points="2708.3473,-350.515 2716.3289,-343.5477 2705.7445,-344.0169 2708.3473,-350.515"/>
</g>
-<!-- Node41 -->
+<!-- Node42 -->
<g id="node27" class="node">
-<title>Node41</title>
+<title>Node42</title>
<polygon fill="#ffffff" stroke="#bfbfbf" points="1668.5,-185 1668.5,-204 1733.5,-204 1733.5,-185 1668.5,-185"/>
<text text-anchor="middle" x="1701" y="-192" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">functional</text>
</g>
-<!-- Node1->Node41 -->
+<!-- Node1->Node42 -->
<g id="edge127" class="edge">
-<title>Node1->Node41</title>
+<title>Node1->Node42</title>
<path fill="none" stroke="#191970" d="M1499.782,-662.1875C1478.4123,-655.7706 1453.9463,-644.2911 1441,-624 1436.2189,-616.5064 1438.0106,-612.3711 1441,-604 1480.4065,-493.6528 1537.922,-493.9067 1608,-400 1625.989,-375.8942 1630.1234,-369.5465 1646,-344 1664.0702,-314.9239 1671.9541,-309.0444 1684,-277 1691.6929,-256.5355 1696.1983,-231.8373 1698.6156,-214.8294"/>
<polygon fill="#191970" stroke="#191970" points="1702.1492,-214.7979 1699.9691,-204.4298 1695.2077,-213.8945 1702.1492,-214.7979"/>
</g>
-<!-- Node42 -->
+<!-- Node43 -->
<g id="node28" class="node">
-<title>Node42</title>
+<title>Node43</title>
<g id="a_node28"><a xlink:href="packed__func_8h.html" target="_top" xlink:title="Type-erased function used across TVM API. ">
<polygon fill="#ffffff" stroke="#000000" points="1164,-313.5 1164,-343.5 1280,-343.5 1280,-313.5 1164,-313.5"/>
<text text-anchor="start" x="1172" y="-331.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/packed</text>
@@ -208,9 +208,9 @@
</a>
</g>
</g>
-<!-- Node1->Node42 -->
+<!-- Node1->Node43 -->
<g id="edge126" class="edge">
-<title>Node1->Node42</title>
+<title>Node1->Node43</title>
<path fill="none" stroke="#191970" d="M1499.9299,-663.3239C1471.8346,-657.014 1434.68,-645.2316 1408,-624 1312.6648,-548.1335 1252.376,-409.246 1230.8681,-353.0603"/>
<polygon fill="#191970" stroke="#191970" points="1234.0951,-351.6974 1227.3037,-343.5655 1227.5416,-354.1576 1234.0951,-351.6974"/>
</g>
@@ -299,30 +299,30 @@
<path fill="none" stroke="#191970" d="M2161.0746,-604.4083C2141.8526,-586.4089 2104,-545.4481 2104,-502 2104,-502 2104,-502 2104,-390 2104,-280.1071 1999.0152,-299.4558 1903,-246 1835.2832,-208.2992 1815.9947,-202.0587 1742,-179 1682.8655,-160.5721 1612.3482,-147.0837 1568.3632,-139.6392"/>
<polygon fill="#191970" stroke="#191970" points="1568.7547,-136.1562 1558.3156,-137.9658 1567.6047,-143.0611 1568.7547,-136.1562"/>
</g>
-<!-- Node46 -->
+<!-- Node47 -->
<g id="node31" class="node">
-<title>Node46</title>
+<title>Node47</title>
<polygon fill="#ffffff" stroke="#bfbfbf" points="1850,-252 1850,-271 1894,-271 1894,-252 1850,-252"/>
<text text-anchor="middle" x="1872" y="-259" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">limits</text>
</g>
-<!-- Node3->Node46 -->
+<!-- Node3->Node47 -->
<g id="edge121" class="edge">
-<title>Node3->Node46</title>
+<title>Node3->Node47</title>
<path fill="none" stroke="#191970" d="M2138.9635,-604.4779C2096.4498,-589.958 2028,-557.9003 2028,-502 2028,-502 2028,-502 2028,-390 2028,-356.7503 1942.5027,-302.3051 1898.0972,-276.2698"/>
<polygon fill="#191970" stroke="#191970" points="1899.6973,-273.1518 1889.2908,-271.1632 1896.1858,-279.2074 1899.6973,-273.1518"/>
</g>
-<!-- Node50 -->
+<!-- Node51 -->
<g id="node35" class="node">
-<title>Node50</title>
+<title>Node51</title>
<g id="a_node35"><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="#000000" points="2398,-548.5 2398,-567.5 2478,-567.5 2478,-548.5 2398,-548.5"/>
<text text-anchor="middle" x="2438" y="-555.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/ir/type.h</text>
</a>
</g>
</g>
-<!-- Node3->Node50 -->
+<!-- Node3->Node51 -->
<g id="edge110" class="edge">
-<title>Node3->Node50</title>
+<title>Node3->Node51</title>
<path fill="none" stroke="#191970" d="M2211.5357,-605.6767C2258.4809,-595.7935 2336.8404,-579.2968 2387.9666,-568.5333"/>
<polygon fill="#191970" stroke="#191970" points="2388.7469,-571.9459 2397.8113,-566.4608 2387.3048,-565.096 2388.7469,-571.9459"/>
</g>
@@ -426,24 +426,24 @@
<path fill="none" stroke="#191970" d="M2469.4182,-444.05C2225.9348,-434.267 1166.4428,-389.2139 1114,-344 1080.6278,-315.228 1076.5075,-283.8896 1099,-246 1116.0563,-217.268 1139.7105,-231.8377 1165,-210 1184.6954,-192.9929 1201.8618,-168.2404 1212.3059,-151.4247"/>
<polygon fill="#191970" stroke="#191970" points="1215.4843,-152.9313 1217.6467,-142.5595 1209.4883,-149.319 1215.4843,-152.9313"/>
</g>
-<!-- Node5->Node35 -->
+<!-- Node5->Node36 -->
<g id="edge100" class="edge">
-<title>Node5->Node35</title>
+<title>Node5->Node36</title>
<path fill="none" stroke="#191970" d="M2546.4511,-436.3385C2570.0677,-427.7168 2604.8379,-414.2783 2634,-400 2665.4775,-384.588 2699.7879,-364.1298 2723.9824,-349.0464"/>
<polygon fill="#191970" stroke="#191970" points="2726.1262,-351.8331 2732.7329,-343.5505 2722.4031,-345.9053 2726.1262,-351.8331"/>
</g>
-<!-- Node48 -->
+<!-- Node49 -->
<g id="node33" class="node">
-<title>Node48</title>
+<title>Node49</title>
<g id="a_node33"><a xlink:href="repr__printer_8h.html" target="_top" xlink:title="Printer class to print repr string of each AST/IR nodes. ">
<polygon fill="#ffffff" stroke="#000000" points="3120.5,-380.5 3120.5,-399.5 3251.5,-399.5 3251.5,-380.5 3120.5,-380.5"/>
<text text-anchor="middle" x="3186" y="-387.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/node/repr_printer.h</text>
</a>
</g>
</g>
-<!-- Node5->Node48 -->
+<!-- Node5->Node49 -->
<g id="edge96" class="edge">
-<title>Node5->Node48</title>
+<title>Node5->Node49</title>
<path fill="none" stroke="#191970" d="M2568.5449,-442.8924C2669.6911,-436.3773 2907.1798,-420.2056 3106,-400 3107.2924,-399.8687 3108.5963,-399.7337 3109.9091,-399.5955"/>
<polygon fill="#191970" stroke="#191970" points="3110.5456,-403.0469 3120.1062,-398.4812 3109.7851,-396.0883 3110.5456,-403.0469"/>
</g>
@@ -489,45 +489,45 @@
<path fill="none" stroke="#191970" d="M2470.3117,-389.5534C2211.7596,-387.449 1212.685,-377.1921 1155,-344 1113.927,-320.3665 1086.8295,-285.5049 1113,-246 1135.6577,-211.7978 1170.2742,-240.2397 1198,-210 1212.5704,-194.1084 1218.6557,-169.6144 1221.193,-152.5333"/>
<polygon fill="#191970" stroke="#191970" points="1224.6751,-152.892 1222.4016,-142.544 1217.7257,-152.0512 1224.6751,-152.892"/>
</g>
-<!-- Node34 -->
+<!-- Node35 -->
<g id="node24" class="node">
-<title>Node34</title>
+<title>Node35</title>
<g id="a_node24"><a xlink:href="data__type_8h.html" target="_top" xlink:title="tvm/runtime/data_type.h">
<polygon fill="#ffffff" stroke="#000000" points="2840,-185 2840,-204 2978,-204 2978,-185 2840,-185"/>
<text text-anchor="middle" x="2909" y="-192" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/data_type.h</text>
</a>
</g>
</g>
-<!-- Node6->Node34 -->
+<!-- Node6->Node35 -->
<g id="edge60" class="edge">
-<title>Node6->Node34</title>
+<title>Node6->Node35</title>
<path fill="none" stroke="#191970" d="M2591.5554,-384.3979C2664.4769,-376.9671 2782.1355,-362.5934 2822,-344 2832.278,-339.2062 2893.6859,-287.0189 2899,-277 2909.3705,-257.4481 2910.9033,-231.7056 2910.4661,-214.2117"/>
<polygon fill="#191970" stroke="#191970" points="2913.9537,-213.8711 2909.9674,-204.0548 2906.9621,-214.2144 2913.9537,-213.8711"/>
</g>
-<!-- Node6->Node35 -->
+<!-- Node6->Node36 -->
<g id="edge46" class="edge">
-<title>Node6->Node35</title>
+<title>Node6->Node36</title>
<path fill="none" stroke="#191970" d="M2566.1563,-380.3906C2599.4973,-371.2774 2650.4047,-357.3627 2691.1129,-346.2358"/>
<polygon fill="#191970" stroke="#191970" points="2692.2355,-349.5574 2700.9588,-343.5446 2690.3898,-342.8051 2692.2355,-349.5574"/>
</g>
-<!-- Node36 -->
+<!-- Node37 -->
<g id="node26" class="node">
-<title>Node36</title>
+<title>Node37</title>
<g id="a_node26"><a xlink:href="ndarray_8h.html" target="_top" xlink:title="A device-independent managed NDArray abstraction. ">
<polygon fill="#ffffff" stroke="#ff0000" points="2102.5,-252 2102.5,-271 2227.5,-271 2227.5,-252 2102.5,-252"/>
<text text-anchor="middle" x="2165" y="-259" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/ndarray.h</text>
</a>
</g>
</g>
-<!-- Node6->Node36 -->
+<!-- Node6->Node37 -->
<g id="edge62" class="edge">
-<title>Node6->Node36</title>
+<title>Node6->Node37</title>
<path fill="none" stroke="#191970" d="M2517.2935,-380.2903C2488.1594,-359.6551 2422.6617,-313.2848 2422,-313 2362.8035,-287.5246 2290.4501,-274.4745 2237.6137,-267.8987"/>
<polygon fill="#191970" stroke="#191970" points="2237.9949,-264.4195 2227.6498,-266.7064 2237.1631,-271.3699 2237.9949,-264.4195"/>
</g>
-<!-- Node6->Node42 -->
+<!-- Node6->Node43 -->
<g id="edge64" class="edge">
-<title>Node6->Node42</title>
+<title>Node6->Node43</title>
<path fill="none" stroke="#191970" d="M2470.4895,-387.1571C2253.9622,-376.9841 1518.4724,-342.429 1290.1688,-331.7027"/>
<polygon fill="#191970" stroke="#191970" points="1290.1585,-328.1985 1280.0053,-331.2252 1289.83,-335.1908 1290.1585,-328.1985"/>
</g>
@@ -567,9 +567,9 @@
<path fill="none" stroke="#191970" d="M2515.7967,-313.2967C2505.3578,-302.8578 2491.4997,-288.9997 2480.7446,-278.2446"/>
<polygon fill="#191970" stroke="#191970" points="2483.101,-275.6512 2473.555,-271.055 2478.1512,-280.601 2483.101,-275.6512"/>
</g>
-<!-- Node33 -->
+<!-- Node34 -->
<g id="node23" class="node">
-<title>Node33</title>
+<title>Node34</title>
<g id="a_node23"><a xlink:href="array_8h.html" target="_top" xlink:title="Runtime Array container types. ">
<polygon fill="#ffffff" stroke="#ff0000" points="1478,-246.5 1478,-276.5 1604,-276.5 1604,-246.5 1478,-246.5"/>
<text text-anchor="start" x="1486" y="-264.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
@@ -577,15 +577,15 @@
</a>
</g>
</g>
-<!-- Node7->Node33 -->
+<!-- Node7->Node34 -->
<g id="edge35" class="edge">
-<title>Node7->Node33</title>
+<title>Node7->Node34</title>
<path fill="none" stroke="#191970" d="M2474.2175,-324.6571C2304.8732,-313.1965 1803.4099,-279.2591 1614.2906,-266.4601"/>
<polygon fill="#191970" stroke="#191970" points="1614.3867,-262.9587 1604.1732,-265.7754 1613.914,-269.9427 1614.3867,-262.9587"/>
</g>
-<!-- Node7->Node34 -->
+<!-- Node7->Node35 -->
<g id="edge40" class="edge">
-<title>Node7->Node34</title>
+<title>Node7->Node35</title>
<path fill="none" stroke="#191970" d="M2587.5771,-325.1793C2643.609,-320.1904 2730.813,-307.7747 2800,-277 2836.4019,-260.8082 2871.9334,-230.3766 2892.083,-211.343"/>
<polygon fill="#191970" stroke="#191970" points="2894.7294,-213.6533 2899.4976,-204.192 2889.87,-208.6148 2894.7294,-213.6533"/>
</g>
@@ -686,13 +686,13 @@
<polygon fill="#191970" stroke="#191970" points="3513.7719,-19.9386 3523.6653,-16.1479 3513.5673,-12.9416 3513.7719,-19.9386"/>
</g>
<!-- Node28->Node17 -->
-<g id="edge31" class="edge">
+<g id="edge29" class="edge">
<title>Node28->Node17</title>
<path fill="none" stroke="#191970" d="M1087.3813,-179.4257C1083.9104,-169.2891 1079.7888,-155.538 1078,-143 1072.9971,-107.9331 1072.8824,-87.1713 1102,-67 1124.5651,-51.368 1974.5273,-23.4685 2181.101,-16.9084"/>
<polygon fill="#191970" stroke="#191970" points="2181.4007,-20.4008 2191.2849,-16.5858 2181.179,-13.4043 2181.4007,-20.4008"/>
</g>
<!-- Node28->Node18 -->
-<g id="edge30" class="edge">
+<g id="edge31" class="edge">
<title>Node28->Node18</title>
<path fill="none" stroke="#191970" d="M1061.1017,-179.4809C1030.1809,-164.9487 981.9324,-142.343 940,-123 885.4575,-97.8402 874.4123,-84.6558 817,-67 722.0259,-37.7929 604.7825,-23.7482 548.7823,-18.3193"/>
<polygon fill="#191970" stroke="#191970" points="549.024,-14.8267 538.7403,-17.3753 548.3688,-21.796 549.024,-14.8267"/>
@@ -727,9 +727,9 @@
<path fill="none" stroke="#191970" d="M1038.2034,-179.4554C997.8323,-168.3714 943.9064,-153.5658 907.8916,-143.6778"/>
<polygon fill="#191970" stroke="#191970" points="908.8045,-140.299 898.2347,-141.0265 906.9512,-147.0492 908.8045,-140.299"/>
</g>
-<!-- Node28->Node32 -->
-<g id="edge29" class="edge">
-<title>Node28->Node32</title>
+<!-- Node28->Node33 -->
+<g id="edge30" class="edge">
+<title>Node28->Node33</title>
<path fill="none" stroke="#191970" d="M1029.9803,-188.4911C904.4652,-176.5234 625.0138,-149.8781 504.7904,-138.4149"/>
<polygon fill="#191970" stroke="#191970" points="504.979,-134.9171 494.692,-137.452 504.3146,-141.8855 504.979,-134.9171"/>
</g>
@@ -751,201 +751,201 @@
<path fill="none" stroke="#191970" d="M1158.3656,-126.3188C1102.4567,-119.5473 1019.8483,-107.1503 950,-87 927.7944,-80.594 924.3121,-73.0243 902,-67 774.3243,-32.5273 616.396,-20.6675 548.9912,-16.9905"/>
<polygon fill="#191970" stroke="#191970" points="548.7646,-13.4741 538.5963,-16.45 548.4011,-20.4646 548.7646,-13.4741"/>
</g>
-<!-- Node33->Node18 -->
+<!-- Node34->Node18 -->
<g id="edge38" class="edge">
-<title>Node33->Node18</title>
+<title>Node34->Node18</title>
<path fill="none" stroke="#191970" d="M1477.935,-254.842C1451.4052,-252.0665 1420.2062,-248.8356 1392,-246 1227.1675,-229.4294 1175.1934,-270.5701 1021,-210 1019.1462,-209.2718 908.7667,-123.9197 907,-123 834.1883,-85.0971 810.0778,-89.0199 731,-67 666.9168,-49.1555 591.0612,-31.8684 548.7444,-22.5627"/>
<polygon fill="#191970" stroke="#191970" points="549.2531,-19.0913 538.7361,-20.3734 547.7571,-25.9296 549.2531,-19.0913"/>
</g>
-<!-- Node33->Node20 -->
+<!-- Node34->Node20 -->
<g id="edge39" class="edge">
-<title>Node33->Node20</title>
+<title>Node34->Node20</title>
<path fill="none" stroke="#191970" d="M1552.964,-246.3166C1568.1455,-227.9824 1595.9893,-197.3302 1626,-179 1658.1112,-159.3868 1699.6431,-146.9092 1728.3188,-139.9719"/>
<polygon fill="#191970" stroke="#191970" points="1729.3113,-143.3346 1738.2607,-137.6637 1727.7282,-136.516 1729.3113,-143.3346"/>
</g>
-<!-- Node33->Node26 -->
+<!-- Node34->Node26 -->
<g id="edge36" class="edge">
-<title>Node33->Node26</title>
+<title>Node34->Node26</title>
<path fill="none" stroke="#191970" d="M1539.2204,-246.2548C1536.5029,-222.9744 1531.3184,-178.5607 1528.3127,-152.8117"/>
<polygon fill="#191970" stroke="#191970" points="1531.7724,-152.2625 1527.1365,-142.7358 1524.8196,-153.0742 1531.7724,-152.2625"/>
</g>
-<!-- Node33->Node31 -->
+<!-- Node34->Node31 -->
<g id="edge37" class="edge">
-<title>Node33->Node31</title>
+<title>Node34->Node31</title>
<path fill="none" stroke="#191970" d="M1477.9458,-254.7323C1451.4177,-251.9392 1420.2171,-248.7247 1392,-246 1212.5676,-228.6735 1157.256,-272.037 988,-210 965.5395,-201.7676 963.9838,-192.149 944,-179 927.6708,-168.2556 908.9259,-156.7586 894.2608,-147.9462"/>
<polygon fill="#191970" stroke="#191970" points="895.6586,-144.7041 885.2793,-142.5778 892.0671,-150.7126 895.6586,-144.7041"/>
</g>
-<!-- Node34->Node11 -->
+<!-- Node35->Node11 -->
<g id="edge41" class="edge">
-<title>Node34->Node11</title>
+<title>Node35->Node11</title>
<path fill="none" stroke="#191970" d="M2908.0322,-184.8762C2905.3232,-157.9361 2897.5977,-81.1108 2893.5478,-40.8362"/>
<polygon fill="#191970" stroke="#191970" points="2897.0163,-40.3465 2892.5332,-30.7469 2890.0514,-41.047 2897.0163,-40.3465"/>
</g>
-<!-- Node34->Node15 -->
+<!-- Node35->Node15 -->
<g id="edge42" class="edge">
-<title>Node34->Node15</title>
+<title>Node35->Node15</title>
<path fill="none" stroke="#191970" d="M2885.9054,-184.9093C2825.0459,-160.1425 2654.7219,-94.0936 2506,-67 2337.2478,-36.2573 1818.0919,-21.5555 1625.7077,-17.0569"/>
<polygon fill="#191970" stroke="#191970" points="1625.777,-13.5577 1615.6986,-16.8252 1625.6149,-20.5558 1625.777,-13.5577"/>
</g>
-<!-- Node34->Node16 -->
+<!-- Node35->Node16 -->
<g id="edge43" class="edge">
-<title>Node34->Node16</title>
+<title>Node35->Node16</title>
<path fill="none" stroke="#191970" d="M2922.8735,-184.8376C2958.6147,-160.4791 3057.5291,-96.414 3150,-67 3217.8914,-45.4045 3432.0046,-25.2907 3513.8018,-18.1984"/>
<polygon fill="#191970" stroke="#191970" points="3514.3303,-21.666 3523.994,-17.3229 3513.7311,-14.6916 3514.3303,-21.666"/>
</g>
-<!-- Node34->Node17 -->
+<!-- Node35->Node17 -->
<g id="edge44" class="edge">
-<title>Node34->Node17</title>
+<title>Node35->Node17</title>
<path fill="none" stroke="#191970" d="M2894.7728,-184.9745C2857.6629,-160.6544 2754.1213,-96.1374 2658,-67 2586.3257,-45.2733 2366.4499,-26.3028 2271,-18.8567"/>
<polygon fill="#191970" stroke="#191970" points="2271.0446,-15.3498 2260.8045,-18.0683 2270.5049,-22.3289 2271.0446,-15.3498"/>
</g>
-<!-- Node35->Node8 -->
+<!-- Node36->Node8 -->
<g id="edge47" class="edge">
-<title>Node35->Node8</title>
+<title>Node36->Node8</title>
<path fill="none" stroke="#191970" d="M2752.9396,-313.4983C2748.7646,-295.8952 2739.9173,-266.5112 2724,-246 2712.218,-230.8175 2694.7671,-218.1794 2679.9966,-209.2226"/>
<polygon fill="#191970" stroke="#191970" points="2681.6271,-206.1222 2671.221,-204.1316 2678.1145,-212.1771 2681.6271,-206.1222"/>
</g>
-<!-- Node35->Node16 -->
+<!-- Node36->Node16 -->
<g id="edge58" class="edge">
-<title>Node35->Node16</title>
+<title>Node36->Node16</title>
<path fill="none" stroke="#191970" d="M2802.6556,-313.4467C2830.9024,-303.9034 2867.4463,-290.7899 2899,-277 3084.791,-195.8038 3113.1214,-135.3848 3304,-67 3376.7456,-40.9379 3466.7536,-26.0732 3513.8408,-19.5384"/>
<polygon fill="#191970" stroke="#191970" points="3514.3208,-23.0054 3523.7619,-18.1974 3513.3831,-16.0685 3514.3208,-23.0054"/>
</g>
-<!-- Node35->Node34 -->
+<!-- Node36->Node35 -->
<g id="edge48" class="edge">
-<title>Node35->Node34</title>
+<title>Node36->Node35</title>
<path fill="none" stroke="#191970" d="M2812.8057,-314.3383C2834.8749,-306.4318 2858.9036,-294.5273 2876,-277 2892.9385,-259.6346 2901.5192,-232.4736 2905.6186,-214.1038"/>
<polygon fill="#191970" stroke="#191970" points="2909.0692,-214.6966 2907.5854,-204.2062 2902.2034,-213.3323 2909.0692,-214.6966"/>
</g>
-<!-- Node35->Node36 -->
+<!-- Node36->Node37 -->
<g id="edge49" class="edge">
-<title>Node35->Node36</title>
+<title>Node36->Node37</title>
<path fill="none" stroke="#191970" d="M2699.3324,-322.0758C2591.5324,-309.8548 2357.074,-283.2749 2237.8159,-269.7549"/>
<polygon fill="#191970" stroke="#191970" points="2238.0185,-266.2556 2227.6878,-268.6067 2237.2299,-273.211 2238.0185,-266.2556"/>
</g>
-<!-- Node35->Node41 -->
+<!-- Node36->Node42 -->
<g id="edge57" class="edge">
-<title>Node35->Node41</title>
+<title>Node36->Node42</title>
<path fill="none" stroke="#191970" d="M2699.3697,-322.6204C2669.0261,-319.5608 2630.9931,-315.8705 2597,-313 2373.2257,-294.104 2315.3932,-308.1971 2093,-277 1964.5025,-258.9745 1815.0906,-223.4027 1743.5018,-205.4346"/>
<polygon fill="#191970" stroke="#191970" points="1744.0683,-201.9679 1733.516,-202.9155 1742.3561,-208.7553 1744.0683,-201.9679"/>
</g>
-<!-- Node36->Node10 -->
+<!-- Node37->Node10 -->
<g id="edge53" class="edge">
-<title>Node36->Node10</title>
+<title>Node37->Node10</title>
<path fill="none" stroke="#191970" d="M2168.1937,-251.8402C2177.8289,-222.6978 2206.737,-135.2626 2219.6494,-96.208"/>
<polygon fill="#191970" stroke="#191970" points="2223.0218,-97.1574 2222.8379,-86.5641 2216.3756,-94.9599 2223.0218,-97.1574"/>
</g>
-<!-- Node36->Node11 -->
+<!-- Node37->Node11 -->
<g id="edge50" class="edge">
-<title>Node36->Node11</title>
+<title>Node37->Node11</title>
<path fill="none" stroke="#191970" d="M2183.6078,-251.9331C2231.4099,-227.6531 2362.8535,-162.7549 2478,-123 2539.5429,-101.752 2556.967,-103.3049 2620,-87 2689.9587,-68.9036 2769.948,-47.7415 2824.7877,-33.1604"/>
<polygon fill="#191970" stroke="#191970" points="2825.9625,-36.4697 2834.7267,-30.5166 2824.163,-29.7049 2825.9625,-36.4697"/>
</g>
-<!-- Node36->Node18 -->
+<!-- Node37->Node18 -->
<g id="edge55" class="edge">
-<title>Node36->Node18</title>
+<title>Node37->Node18</title>
<path fill="none" stroke="#191970" d="M2162.9134,-251.7587C2158.5973,-234.1499 2146.9326,-197.1678 2122,-179 2002.478,-91.9074 1940.7053,-141.6593 1794,-123 1300.5357,-60.2368 695.0423,-25.1484 548.8718,-17.2316"/>
<polygon fill="#191970" stroke="#191970" points="548.9435,-13.7305 538.7699,-16.6883 548.5675,-20.7204 548.9435,-13.7305"/>
</g>
-<!-- Node36->Node20 -->
+<!-- Node37->Node20 -->
<g id="edge56" class="edge">
-<title>Node36->Node20</title>
+<title>Node37->Node20</title>
<path fill="none" stroke="#191970" d="M2159.2248,-251.8665C2147.9953,-234.092 2121.3656,-196.3714 2088,-179 2037.5039,-152.7097 1868.2014,-139.4814 1795.5631,-134.9058"/>
<polygon fill="#191970" stroke="#191970" points="1795.7686,-131.4119 1785.573,-134.293 1795.3399,-138.3988 1795.7686,-131.4119"/>
</g>
-<!-- Node36->Node28 -->
+<!-- Node37->Node28 -->
<g id="edge51" class="edge">
-<title>Node36->Node28</title>
+<title>Node37->Node28</title>
<path fill="none" stroke="#191970" d="M2102.338,-257.5836C1915.9214,-245.9326 1365.4611,-211.5288 1166.1207,-199.07"/>
<polygon fill="#191970" stroke="#191970" points="1166.244,-195.571 1156.0451,-198.4403 1165.8073,-202.5574 1166.244,-195.571"/>
</g>
-<!-- Node36->Node34 -->
+<!-- Node37->Node35 -->
<g id="edge52" class="edge">
-<title>Node36->Node34</title>
+<title>Node37->Node35</title>
<path fill="none" stroke="#191970" d="M2227.7918,-255.8454C2362.978,-243.6713 2681.3381,-215.0018 2829.9735,-201.6166"/>
<polygon fill="#191970" stroke="#191970" points="2830.3003,-205.1015 2839.946,-200.7186 2829.6724,-198.1297 2830.3003,-205.1015"/>
</g>
-<!-- Node36->Node41 -->
+<!-- Node37->Node42 -->
<g id="edge54" class="edge">
-<title>Node36->Node41</title>
+<title>Node37->Node42</title>
<path fill="none" stroke="#191970" d="M2102.4229,-252.4641C2006.5854,-238.6255 1825.9362,-212.5404 1743.705,-200.6665"/>
<polygon fill="#191970" stroke="#191970" points="1744.106,-197.1881 1733.7084,-199.223 1743.1055,-204.1163 1744.106,-197.1881"/>
</g>
-<!-- Node42->Node10 -->
+<!-- Node43->Node10 -->
<g id="edge84" class="edge">
-<title>Node42->Node10</title>
+<title>Node43->Node10</title>
<path fill="none" stroke="#191970" d="M1234.8917,-313.4724C1270.1134,-273.6495 1372.8539,-165.5502 1485,-123 1546.6694,-99.6016 1983.4657,-84.2118 2156.0708,-78.9829"/>
<polygon fill="#191970" stroke="#191970" points="2156.4224,-82.474 2166.3128,-78.6752 2156.2121,-75.4772 2156.4224,-82.474"/>
</g>
-<!-- Node42->Node11 -->
+<!-- Node43->Node11 -->
<g id="edge65" class="edge">
-<title>Node42->Node11</title>
+<title>Node43->Node11</title>
<path fill="none" stroke="#191970" d="M1280.0438,-327.314C1530.7484,-322.0599 2510.8052,-300.1186 2572,-277 2592.3894,-269.2971 2590.9337,-256.5612 2610,-246 2653.523,-221.8918 2674.4581,-237.3812 2716,-210 2787.5768,-162.8221 2848.7284,-79.497 2875.9253,-38.9616"/>
<polygon fill="#191970" stroke="#191970" points="2878.9032,-40.8037 2881.5031,-30.5328 2873.0656,-36.9406 2878.9032,-40.8037"/>
</g>
-<!-- Node42->Node15 -->
+<!-- Node43->Node15 -->
<g id="edge72" class="edge">
-<title>Node42->Node15</title>
+<title>Node43->Node15</title>
<path fill="none" stroke="#191970" d="M1163.7398,-322.8054C1082.1429,-314.2628 938.3853,-296.9834 890,-277 884.156,-274.5864 800.8673,-215.6353 798,-210 780.4651,-175.5379 773.9371,-153.267 798,-123 840.2308,-69.8808 1299.0585,-32.8957 1480.1262,-20.2718"/>
<polygon fill="#191970" stroke="#191970" points="1480.6176,-23.7463 1490.3521,-19.5642 1480.1343,-16.763 1480.6176,-23.7463"/>
</g>
-<!-- Node42->Node16 -->
+<!-- Node43->Node16 -->
<g id="edge88" class="edge">
-<title>Node42->Node16</title>
+<title>Node43->Node16</title>
<path fill="none" stroke="#191970" d="M1280.164,-327.8567C1527.952,-324.9087 2485.328,-311.3303 2539,-277 2580.1914,-250.6527 2550.6488,-209.3331 2589,-179 2605.9093,-165.6259 2945.12,-92.3684 2966,-87 2998.9958,-78.5165 3006.4981,-73.1897 3040,-67 3217.004,-34.2973 3432.9537,-21.0214 3513.5944,-16.9695"/>
<polygon fill="#191970" stroke="#191970" points="3513.8052,-20.4635 3523.6221,-16.4786 3513.4629,-13.4719 3513.8052,-20.4635"/>
</g>
-<!-- Node42->Node17 -->
+<!-- Node43->Node17 -->
<g id="edge90" class="edge">
-<title>Node42->Node17</title>
+<title>Node43->Node17</title>
<path fill="none" stroke="#191970" d="M1163.7864,-321.0565C1096.3937,-311.7911 990.0354,-294.9805 954,-277 875.6193,-237.8905 776.3054,-191.4221 831,-123 839.8194,-111.9671 1065.001,-68.8797 1079,-67 1297.87,-37.6114 1997.1168,-20.5157 2181.2251,-16.4491"/>
<polygon fill="#191970" stroke="#191970" points="2181.4367,-19.9454 2191.3576,-16.227 2181.2832,-12.9471 2181.4367,-19.9454"/>
</g>
-<!-- Node42->Node18 -->
+<!-- Node43->Node18 -->
<g id="edge91" class="edge">
-<title>Node42->Node18</title>
+<title>Node43->Node18</title>
<path fill="none" stroke="#191970" d="M1163.8274,-325.5158C1071.569,-319.9739 896.5529,-305.9287 842,-277 736.4035,-221.0035 771.3122,-136.4014 674,-67 635.9543,-39.8665 582.803,-26.4518 548.8906,-20.258"/>
<polygon fill="#191970" stroke="#191970" points="549.0475,-16.7338 538.5997,-18.4924 547.8638,-23.633 549.0475,-16.7338"/>
</g>
-<!-- Node42->Node20 -->
+<!-- Node43->Node20 -->
<g id="edge92" class="edge">
-<title>Node42->Node20</title>
+<title>Node43->Node20</title>
<path fill="none" stroke="#191970" d="M1280.0603,-313.8517C1321.3191,-302.893 1372.6951,-287.9692 1392,-277 1409.7796,-266.8975 1408.7856,-257.038 1426,-246 1442.0522,-235.7071 1565.9416,-185.1246 1584,-179 1633.3826,-162.2517 1692.2294,-148.1458 1728.3488,-140.1307"/>
<polygon fill="#191970" stroke="#191970" points="1729.3311,-143.4985 1738.3498,-137.9385 1727.8323,-136.6609 1729.3311,-143.4985"/>
</g>
-<!-- Node42->Node31 -->
+<!-- Node43->Node31 -->
<g id="edge87" class="edge">
-<title>Node42->Node31</title>
+<title>Node43->Node31</title>
<path fill="none" stroke="#191970" d="M1163.8625,-320.362C1094.4342,-307.8712 978.4898,-277.8607 905,-210 888.369,-194.6429 878.743,-170.0428 873.6985,-152.7909"/>
<polygon fill="#191970" stroke="#191970" points="876.965,-151.454 871.0175,-142.686 870.1991,-153.2491 876.965,-151.454"/>
</g>
-<!-- Node42->Node33 -->
+<!-- Node43->Node34 -->
<g id="edge66" class="edge">
-<title>Node42->Node33</title>
+<title>Node43->Node34</title>
<path fill="none" stroke="#191970" d="M1280.152,-315.9473C1284.8364,-314.9465 1289.4969,-313.9542 1294,-313 1352.4322,-300.6189 1418.6931,-286.8163 1467.8653,-276.6191"/>
<polygon fill="#191970" stroke="#191970" points="1468.6954,-280.0215 1477.7768,-274.5645 1467.2745,-273.1672 1468.6954,-280.0215"/>
</g>
-<!-- Node42->Node34 -->
+<!-- Node43->Node35 -->
<g id="edge71" class="edge">
-<title>Node42->Node34</title>
+<title>Node43->Node35</title>
<path fill="none" stroke="#191970" d="M1280.2906,-328.1376C1509.2957,-326.4353 2347.5551,-317.5117 2610,-277 2706.9451,-262.0353 2817.3172,-226.595 2872.9212,-207.3959"/>
<polygon fill="#191970" stroke="#191970" points="2874.3575,-210.602 2882.6508,-204.0086 2872.0559,-203.9912 2874.3575,-210.602"/>
</g>
-<!-- Node42->Node36 -->
+<!-- Node43->Node37 -->
<g id="edge83" class="edge">
-<title>Node42->Node36</title>
+<title>Node43->Node37</title>
<path fill="none" stroke="#191970" d="M1280.2772,-324.3594C1444.5362,-312.6888 1911.056,-279.5427 2092.0096,-266.686"/>
<polygon fill="#191970" stroke="#191970" points="2092.4941,-270.1605 2102.2208,-265.9604 2091.9979,-263.1781 2092.4941,-270.1605"/>
</g>
-<!-- Node42->Node41 -->
+<!-- Node43->Node42 -->
<g id="edge85" class="edge">
-<title>Node42->Node41</title>
+<title>Node43->Node42</title>
<path fill="none" stroke="#191970" d="M1280.1814,-314.9914C1335.6617,-301.9191 1412.6934,-283.1989 1426,-277 1447.3561,-267.0512 1447.5027,-255.6398 1469,-246 1531.43,-218.0052 1610.33,-204.669 1658.1652,-198.7474"/>
<polygon fill="#191970" stroke="#191970" points="1658.656,-202.2138 1668.1744,-197.5608 1657.8319,-195.2625 1658.656,-202.2138"/>
</g>
-<!-- Node43 -->
+<!-- Node44 -->
<g id="node29" class="node">
-<title>Node43</title>
+<title>Node44</title>
<g id="a_node29"><a xlink:href="map_8h.html" target="_top" xlink:title="Runtime Map container types. ">
<polygon fill="#ffffff" stroke="#ff0000" points="294,-246.5 294,-276.5 420,-276.5 420,-246.5 294,-246.5"/>
<text text-anchor="start" x="302" y="-264.5" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/container</text>
@@ -953,447 +953,447 @@
</a>
</g>
</g>
-<!-- Node42->Node43 -->
+<!-- Node43->Node44 -->
<g id="edge67" class="edge">
-<title>Node42->Node43</title>
+<title>Node43->Node44</title>
<path fill="none" stroke="#191970" d="M1163.9018,-323.9999C1011.1476,-312.1681 598.4592,-280.2026 430.0565,-267.1587"/>
<polygon fill="#191970" stroke="#191970" points="430.2865,-263.6661 420.046,-266.3833 429.7458,-270.6452 430.2865,-263.6661"/>
</g>
-<!-- Node44 -->
+<!-- Node45 -->
<g id="node30" class="node">
-<title>Node44</title>
+<title>Node45</title>
<g id="a_node30"><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="1122,-252 1122,-271 1248,-271 1248,-252 1122,-252"/>
<text text-anchor="middle" x="1185" y="-259" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/runtime/module.h</text>
</a>
</g>
</g>
-<!-- Node42->Node44 -->
+<!-- Node43->Node45 -->
<g id="edge73" class="edge">
-<title>Node42->Node44</title>
+<title>Node43->Node45</title>
<path fill="none" stroke="#191970" d="M1207.8433,-313.2967C1201.2979,-303.5672 1194.1544,-290.8675 1189.3292,-280.4784"/>
<polygon fill="#191970" stroke="#191970" points="1192.4745,-278.928 1185.3847,-271.055 1186.0174,-281.6309 1192.4745,-278.928"/>
</g>
-<!-- Node42->Node46 -->
+<!-- Node43->Node47 -->
<g id="edge86" class="edge">
-<title>Node42->Node46</title>
+<title>Node43->Node47</title>
<path fill="none" stroke="#191970" d="M1280.2552,-322.4952C1413.4287,-308.7681 1736.578,-275.4589 1839.7617,-264.823"/>
<polygon fill="#191970" stroke="#191970" points="1840.2834,-268.2879 1849.8718,-263.7809 1839.5656,-261.3248 1840.2834,-268.2879"/>
</g>
-<!-- Node47 -->
+<!-- Node48 -->
<g id="node32" class="node">
-<title>Node47</title>
+<title>Node48</title>
<polygon fill="#ffffff" stroke="#bfbfbf" points="1342.5,-252 1342.5,-271 1383.5,-271 1383.5,-252 1342.5,-252"/>
<text text-anchor="middle" x="1363" y="-259" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tuple</text>
</g>
-<!-- Node42->Node47 -->
+<!-- Node43->Node48 -->
<g id="edge89" class="edge">
-<title>Node42->Node47</title>
+<title>Node43->Node48</title>
<path fill="none" stroke="#191970" d="M1253.6432,-313.4639C1277.6208,-302.0702 1310.3193,-286.5327 1333.7017,-275.4219"/>
<polygon fill="#191970" stroke="#191970" points="1335.2588,-278.5571 1342.7888,-271.1039 1332.2544,-272.2346 1335.2588,-278.5571"/>
</g>
-<!-- Node43->Node18 -->
+<!-- Node44->Node18 -->
<g id="edge70" class="edge">
-<title>Node43->Node18</title>
+<title>Node44->Node18</title>
<path fill="none" stroke="#191970" d="M357.8859,-246.2067C360.0559,-219.3817 367.4182,-162.9112 392,-123 417.0773,-82.2843 462.5709,-48.7498 490.9317,-30.4809"/>
<polygon fill="#191970" stroke="#191970" points="492.8901,-33.3839 499.4902,-25.0961 489.1622,-27.4591 492.8901,-33.3839"/>
</g>
-<!-- Node43->Node26 -->
+<!-- Node44->Node26 -->
<g id="edge68" class="edge">
-<title>Node43->Node26</title>
+<title>Node44->Node26</title>
<path fill="none" stroke="#191970" d="M420.0916,-252.9615C537.9904,-237.1615 799.8232,-202.8072 1021,-179 1192.3778,-160.5531 1397.0039,-143.416 1483.7768,-136.3754"/>
<polygon fill="#191970" stroke="#191970" points="1484.0793,-139.8625 1493.7645,-135.5674 1483.5148,-132.8853 1484.0793,-139.8625"/>
</g>
-<!-- Node43->Node32 -->
+<!-- Node44->Node33 -->
<g id="edge69" class="edge">
-<title>Node43->Node32</title>
+<title>Node44->Node33</title>
<path fill="none" stroke="#191970" d="M367.7962,-246.2548C384.6524,-222.4524 417.1532,-176.5584 435.1762,-151.1084"/>
<polygon fill="#191970" stroke="#191970" points="438.1824,-152.9194 441.1054,-142.7358 432.4697,-148.8739 438.1824,-152.9194"/>
</g>
-<!-- Node44->Node10 -->
+<!-- Node45->Node10 -->
<g id="edge77" class="edge">
-<title>Node44->Node10</title>
+<title>Node45->Node10</title>
<path fill="none" stroke="#191970" d="M1223.0205,-251.9213C1249.7975,-243.8518 1285.4596,-230.3241 1312,-210 1350.5334,-180.4918 1335.1311,-145.7555 1378,-123 1445.8542,-86.9818 1965.2759,-79.1215 2156.0289,-77.4438"/>
<polygon fill="#191970" stroke="#191970" points="2156.2493,-80.9421 2166.2192,-77.3573 2156.1898,-73.9424 2156.2493,-80.9421"/>
</g>
-<!-- Node44->Node11 -->
+<!-- Node45->Node11 -->
<g id="edge74" class="edge">
-<title>Node44->Node11</title>
+<title>Node45->Node11</title>
<path fill="none" stroke="#191970" d="M1248.3549,-254.632C1274.5465,-251.8533 1305.2269,-248.6764 1333,-246 1514.6391,-228.4958 1561.0062,-233.2468 1742,-210 1953.6074,-182.8211 2596.39,-68.4155 2816.2918,-28.951"/>
<polygon fill="#191970" stroke="#191970" points="2817.2161,-32.3411 2826.4403,-27.129 2815.9791,-25.4512 2817.2161,-32.3411"/>
</g>
-<!-- Node44->Node16 -->
+<!-- Node45->Node16 -->
<g id="edge79" class="edge">
-<title>Node44->Node16</title>
+<title>Node45->Node16</title>
<path fill="none" stroke="#191970" d="M1220.6885,-251.9314C1252.4442,-242.8663 1299.7946,-227.9832 1339,-210 1407.6578,-178.5073 1412.6383,-144.6671 1485,-123 1657.3934,-71.3807 2114.2761,-96.1303 2294,-87 2779.2832,-62.3466 3370.0132,-26.335 3513.481,-17.508"/>
<polygon fill="#191970" stroke="#191970" points="3514.0076,-20.9823 3523.7735,-16.8741 3513.5772,-13.9955 3514.0076,-20.9823"/>
</g>
-<!-- Node44->Node20 -->
+<!-- Node45->Node20 -->
<g id="edge81" class="edge">
-<title>Node44->Node20</title>
+<title>Node45->Node20</title>
<path fill="none" stroke="#191970" d="M1248.019,-254.4001C1297.7624,-247.3042 1368.173,-233.9043 1426,-210 1449.4695,-200.2983 1450.2856,-188.0866 1474,-179 1520.02,-161.3666 1662.5777,-143.9524 1728.2487,-136.6199"/>
<polygon fill="#191970" stroke="#191970" points="1728.9164,-140.0675 1738.4716,-135.4907 1728.1478,-133.1098 1728.9164,-140.0675"/>
</g>
-<!-- Node44->Node28 -->
+<!-- Node45->Node28 -->
<g id="edge75" class="edge">
-<title>Node44->Node28</title>
+<title>Node45->Node28</title>
<path fill="none" stroke="#191970" d="M1171.8186,-251.9005C1158.8718,-242.4718 1138.6876,-227.7725 1121.9863,-215.6096"/>
<polygon fill="#191970" stroke="#191970" points="1123.7709,-212.5795 1113.6269,-209.5218 1119.65,-218.238 1123.7709,-212.5795"/>
</g>
-<!-- Node44->Node24 -->
+<!-- Node45->Node24 -->
<g id="edge76" class="edge">
-<title>Node44->Node24</title>
+<title>Node45->Node24</title>
<path fill="none" stroke="#191970" d="M1199.8096,-251.6651C1212.1085,-242.5276 1228.7237,-227.6884 1236,-210 1243.7424,-191.1785 1237.9344,-167.9535 1231.8055,-151.9613"/>
<polygon fill="#191970" stroke="#191970" points="1234.9769,-150.474 1227.8708,-142.6156 1228.5254,-153.1902 1234.9769,-150.474"/>
</g>
-<!-- Node44->Node31 -->
+<!-- Node45->Node31 -->
<g id="edge78" class="edge">
-<title>Node44->Node31</title>
+<title>Node45->Node31</title>
<path fill="none" stroke="#191970" d="M1121.9568,-256.9796C1072.7247,-251.317 1003.8641,-238.5862 950,-210 922.1633,-195.2268 897.0434,-168.204 882.3983,-150.4222"/>
<polygon fill="#191970" stroke="#191970" points="885.0643,-148.1519 876.0843,-142.5295 879.5981,-152.5247 885.0643,-148.1519"/>
</g>
-<!-- Node44->Node32 -->
+<!-- Node45->Node33 -->
<g id="edge80" class="edge">
-<title>Node44->Node32</title>
+<title>Node45->Node33</title>
<path fill="none" stroke="#191970" d="M1130.5073,-251.9989C994.8694,-228.3497 643.1093,-167.0184 504.9259,-142.9253"/>
<polygon fill="#191970" stroke="#191970" points="505.4123,-139.4574 494.9597,-141.1877 504.2099,-146.3534 505.4123,-139.4574"/>
</g>
-<!-- Node44->Node42 -->
+<!-- Node45->Node43 -->
<g id="edge82" class="edge">
-<title>Node44->Node42</title>
+<title>Node45->Node43</title>
<path fill="none" stroke="#191970" d="M1195.1686,-271.055C1201.5951,-279.6613 1209.3301,-292.6846 1215.0772,-304.1564"/>
<polygon fill="#191970" stroke="#191970" points="1211.9493,-305.7297 1219.3649,-313.2967 1218.2866,-302.7568 1211.9493,-305.7297"/>
</g>
-<!-- Node48->Node8 -->
+<!-- Node49->Node8 -->
<g id="edge97" class="edge">
-<title>Node48->Node8</title>
+<title>Node49->Node8</title>
<path fill="none" stroke="#191970" d="M3174.1958,-380.4054C3141.2874,-354.1985 3044.8289,-281.0293 2952,-246 2874.2542,-216.6624 2778.8298,-203.9349 2716.7777,-198.4812"/>
<polygon fill="#191970" stroke="#191970" points="2716.9916,-194.987 2706.7334,-197.6364 2716.4049,-201.9623 2716.9916,-194.987"/>
</g>
-<!-- Node49 -->
+<!-- Node50 -->
<g id="node34" class="node">
-<title>Node49</title>
+<title>Node50</title>
<polygon fill="#ffffff" stroke="#bfbfbf" points="3249,-319 3249,-338 3309,-338 3309,-319 3249,-319"/>
<text text-anchor="middle" x="3279" y="-326" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">iostream</text>
</g>
-<!-- Node48->Node49 -->
+<!-- Node49->Node50 -->
<g id="edge98" class="edge">
-<title>Node48->Node49</title>
+<title>Node49->Node50</title>
<path fill="none" stroke="#191970" d="M3200.5313,-380.3906C3215.4094,-370.5519 3238.7504,-355.1167 3256.1128,-343.6351"/>
<polygon fill="#191970" stroke="#191970" points="3258.1602,-346.4773 3264.5708,-338.0419 3254.299,-340.6385 3258.1602,-346.4773"/>
</g>
-<!-- Node50->Node4 -->
+<!-- Node51->Node4 -->
<g id="edge111" class="edge">
-<title>Node50->Node4</title>
+<title>Node51->Node4</title>
<path fill="none" stroke="#191970" d="M2438,-548.2455C2438,-540.9382 2438,-530.6944 2438,-521.7046"/>
<polygon fill="#191970" stroke="#191970" points="2441.5001,-521.6426 2438,-511.6427 2434.5001,-521.6427 2441.5001,-521.6426"/>
</g>
-<!-- Node50->Node5 -->
+<!-- Node51->Node5 -->
<g id="edge112" class="edge">
-<title>Node50->Node5</title>
+<title>Node51->Node5</title>
<path fill="none" stroke="#191970" d="M2451.2342,-548.208C2462.1653,-539.6055 2477.5323,-526.2643 2488,-512 2498.6236,-497.5233 2506.9916,-478.8269 2512.3422,-465.0313"/>
<polygon fill="#191970" stroke="#191970" points="2515.6485,-466.1812 2515.833,-455.588 2509.0827,-463.7541 2515.6485,-466.1812"/>
</g>
-<!-- Node50->Node10 -->
+<!-- Node51->Node10 -->
<g id="edge115" class="edge">
-<title>Node50->Node10</title>
+<title>Node51->Node10</title>
<path fill="none" stroke="#191970" d="M2397.6083,-551.5358C2343.9896,-540.3903 2256,-512.0395 2256,-446 2256,-446 2256,-446 2256,-390 2256,-325.6982 2277.6732,-309.8423 2270,-246 2263.2854,-190.1329 2254.8643,-177.2704 2240,-123 2237.6081,-114.2669 2234.7297,-104.6761 2232.2066,-96.5223"/>
<polygon fill="#191970" stroke="#191970" points="2235.4972,-95.3184 2229.165,-86.824 2228.8179,-97.4132 2235.4972,-95.3184"/>
</g>
-<!-- Node50->Node16 -->
+<!-- Node51->Node16 -->
<g id="edge116" class="edge">
-<title>Node50->Node16</title>
+<title>Node51->Node16</title>
<path fill="none" stroke="#191970" d="M2478.0524,-556.3802C2591.3481,-551.5489 2912.5228,-536.0644 3015,-512 3145.0798,-481.4537 3176.8806,-464.3185 3294,-400 3331.4064,-379.4575 3342.6196,-374.952 3372,-344 3464.6953,-246.3461 3521.8674,-89.6875 3539.8527,-35.0519"/>
<polygon fill="#191970" stroke="#191970" points="3543.2706,-35.8568 3543.0107,-25.2651 3536.6088,-33.7071 3543.2706,-35.8568"/>
</g>
-<!-- Node50->Node33 -->
+<!-- Node51->Node34 -->
<g id="edge113" class="edge">
-<title>Node50->Node33</title>
+<title>Node51->Node34</title>
<path fill="none" stroke="#191970" d="M2397.7951,-550.1934C2357.8815,-542.0632 2295.1438,-528.2552 2242,-512 2151.0682,-484.1866 2131.3528,-468.5308 2042,-436 1880.2108,-377.0972 1687.5577,-311.1702 1595.3876,-279.8864"/>
<polygon fill="#191970" stroke="#191970" points="1596.3132,-276.5046 1585.7189,-276.6067 1594.0646,-283.1336 1596.3132,-276.5046"/>
</g>
-<!-- Node50->Node34 -->
+<!-- Node51->Node35 -->
<g id="edge114" class="edge">
-<title>Node50->Node34</title>
+<title>Node51->Node35</title>
<path fill="none" stroke="#191970" d="M2478.2467,-556.7489C2629.3732,-551.1291 3159.5224,-522.8124 3261,-400 3307.9002,-343.2394 3045.9446,-242.9639 2946.2803,-207.4283"/>
<polygon fill="#191970" stroke="#191970" points="2947.1999,-204.041 2936.6052,-204.0005 2944.8623,-210.6391 2947.1999,-204.041"/>
</g>
-<!-- Node51->Node4 -->
+<!-- Node52->Node4 -->
<g id="edge134" class="edge">
-<title>Node51->Node4</title>
+<title>Node52->Node4</title>
<path fill="none" stroke="#191970" d="M3700.8694,-839.4526C3702.4958,-815.8358 3703.5029,-753.8573 3674,-716 3638.0693,-669.8949 3610.6989,-677.7293 3555,-660 3157.8057,-533.5707 2649.0134,-508.1548 2488.7004,-503.178"/>
<polygon fill="#191970" stroke="#191970" points="2488.7495,-499.678 2478.6497,-502.8789 2488.5412,-506.6749 2488.7495,-499.678"/>
</g>
-<!-- Node51->Node5 -->
+<!-- Node52->Node5 -->
<g id="edge135" class="edge">
-<title>Node51->Node5</title>
+<title>Node52->Node5</title>
<path fill="none" stroke="#191970" d="M3730.3547,-839.4735C3774.1852,-823.6031 3850,-787.2929 3850,-726 3850,-726 3850,-726 3850,-670 3850,-406.298 3543.3287,-539.8269 3284,-492 3148.9783,-467.0985 2734.4634,-452.4505 2578.8557,-447.7079"/>
<polygon fill="#191970" stroke="#191970" points="2578.6472,-444.2001 2568.5461,-447.3966 2578.4359,-451.1969 2578.6472,-444.2001"/>
</g>
-<!-- Node51->Node16 -->
+<!-- Node52->Node16 -->
<g id="edge162" class="edge">
-<title>Node51->Node16</title>
+<title>Node52->Node16</title>
<path fill="none" stroke="#191970" d="M3738.9541,-839.4814C3812.3901,-820.3935 3964,-775.0469 3964,-726 3964,-726 3964,-726 3964,-133 3964,-98.6999 3952.6781,-85.8165 3924,-67 3867.2941,-29.7937 3659.3886,-19.1573 3578.5019,-16.3912"/>
<polygon fill="#191970" stroke="#191970" points="3578.5082,-12.8896 3568.4003,-16.0645 3578.2819,-19.886 3578.5082,-12.8896"/>
</g>
-<!-- Node51->Node20 -->
+<!-- Node52->Node20 -->
<g id="edge163" class="edge">
-<title>Node51->Node20</title>
+<title>Node52->Node20</title>
<path fill="none" stroke="#191970" d="M3650.2393,-848.3634C3426.0136,-845.0078 2508.7385,-825.9335 2238,-736 1952.1711,-641.0538 1800.2229,-244.4079 1768.3692,-152.2067"/>
<polygon fill="#191970" stroke="#191970" points="1771.6073,-150.8572 1765.079,-142.5126 1764.9787,-153.107 1771.6073,-150.8572"/>
</g>
-<!-- Node52 -->
+<!-- Node53 -->
<g id="node37" class="node">
-<title>Node52</title>
+<title>Node53</title>
<g id="a_node37"><a xlink:href="tir_2expr_8h.html" target="_top" xlink:title="TIR expressions. ">
<polygon fill="#ffffff" stroke="#000000" points="2697.5,-778 2697.5,-797 2780.5,-797 2780.5,-778 2697.5,-778"/>
<text text-anchor="middle" x="2739" y="-785" font-family="Helvetica,sans-Serif" font-size="10.00" fill="#000000">tvm/tir/expr.h</text>
</a>
</g>
</g>
-<!-- Node51->Node52 -->
+<!-- Node52->Node53 -->
<g id="edge136" class="edge">
-<title>Node51->Node52</title>
+<title>Node52->Node53</title>
<path fill="none" stroke="#191970" d="M3650.3561,-845.823C3485.6525,-835.2827 2958.2327,-801.53 2790.6382,-790.8046"/>
<polygon fill="#191970" stroke="#191970" points="2790.7943,-787.3076 2780.5912,-790.1617 2790.3472,-794.2933 2790.7943,-787.3076"/>
</g>
-<!-- Node52->Node3 -->
+<!-- Node53->Node3 -->
<g id="edge137" class="edge">
-<title>Node52->Node3</title>
+<title>Node53->Node3</title>
<path fill="none" stroke="#191970" d="M2697.3054,-785.4518C2606.3819,-780.4769 2393.2009,-765.9062 2328,-736 2282.2382,-715.0101 2284.0191,-690.5481 2244,-660 2229.2017,-648.7039 2211.6076,-637.4274 2197.5161,-628.8664"/>
<polygon fill="#191970" stroke="#191970" points="2199.2033,-625.7972 2188.8263,-623.6599 2195.6056,-631.8019 2199.2033,-625.7972"/>
</g>
-<!-- Node52->Node5 -->
+<!-- Node53->Node5 -->
<g id="edge139" class="edge">
-<title>Node52->Node5</title>
+<title>Node53->Node5</title>
<path fill="none" stroke="#191970" d="M2732.6531,-777.6479C2702.2254,-730.4157 2571.1035,-526.8789 2530.8187,-464.3459"/>
<polygon fill="#191970" stroke="#191970" points="2533.5206,-462.077 2525.1625,-455.5659 2527.6359,-465.868 2533.5206,-462.077"/>
</g>
-<!-- Node52->Node8 -->
+<!-- Node53->Node8 -->
<g id="edge138" class="edge">
-<title>Node52->Node8</title>
+<title>Node53->Node8</title>
<path fill="none" stroke="#191970" d="M2780.7221,-785.7344C2897.1081,-780.51 3223.5688,-763.736 3327,-736 3383.8465,-720.7561 3415.3846,-729.6424 3447,-680 3474.238,-637.2309 3481.1188,-640.5206 3390,-548 3362.1009,-519.6717 2818.4912,-270.4956 2682.5899,-208.4465"/>
<polygon fill="#191970" stroke="#191970" points="2683.7181,-205.1141 2673.1676,-204.1461 2680.8116,-211.4822 2683.7181,-205.1141"/>
</g>
-<!-- Node52->Node11 -->
+<!-- Node53->Node11 -->
<g id="edge140" class="edge">
-<title>Node52->Node11</title>
+<title>Node53->Node11</title>
<path fill="none" stroke="#191970" d="M2780.5702,-787.2501C2930.9178,-786.0016 3445.8562,-778.6385 3604,-736 3747.8291,-697.221 3888,-706.9652 3888,-558 3888,-558 3888,-558 3888,-446 3888,-245.1304 3197.2675,-80.3944 2965.6686,-30.8048"/>
<polygon fill="#191970" stroke="#191970" points="2966.1576,-27.3305 2955.6478,-28.6698 2964.6989,-34.1769 2966.1576,-27.3305"/>
</g>
-<!-- Node52->Node16 -->
+<!-- Node53->Node16 -->
<g id="edge159" class="edge">
-<title>Node52->Node16</title>
+<title>Node53->Node16</title>
<path fill="none" stroke="#191970" d="M2780.5983,-786.4728C2936.6871,-782.4169 3487.3692,-766.2401 3660,-736 3800.1165,-711.4555 3926,-700.25 3926,-558 3926,-558 3926,-558 3926,-133 3926,-101.7875 3926.2141,-86.6952 3902,-67 3852.7221,-26.9184 3656.9565,-18.0178 3578.6353,-16.0526"/>
<polygon fill="#191970" stroke="#191970" points="3578.4681,-12.5479 3568.3908,-15.8188 3578.3083,-19.5461 3578.4681,-12.5479"/>
</g>
-<!-- Node52->Node18 -->
+<!-- Node53->Node18 -->
<g id="edge161" class="edge">
-<title>Node52->Node18</title>
+<title>Node53->Node18</title>
<path fill="none" stroke="#191970" d="M2697.1914,-787.0141C2390.094,-783.3908 493.5402,-760.0644 440,-736 375.944,-707.2092 114,-264.7287 114,-194.5 114,-194.5 114,-194.5 114,-133 114,-91.8295 142.3374,-85.7314 179,-67 231.8725,-39.9869 408.8625,-23.683 482.9474,-17.8942"/>
<polygon fill="#191970" stroke="#191970" points="483.4007,-21.3698 493.1043,-17.1164 482.8662,-14.3902 483.4007,-21.3698"/>
</g>
-<!-- Node52->Node28 -->
+<!-- Node53->Node28 -->
<g id="edge143" class="edge">
-<title>Node52->Node28</title>
+<title>Node53->Node28</title>
<path fill="none" stroke="#191970" d="M2697.469,-786.8302C2445.6462,-782.6988 1128.1129,-760.0667 950,-736 846.1668,-721.97 810.5485,-734.3988 721,-680 672.915,-650.7894 666.9645,-624.0427 662,-568 649.2083,-423.6003 603.2235,-339.5066 714,-246 736.8963,-226.6732 915.8999,-208.9991 1019.5064,-200.2542"/>
<polygon fill="#191970" stroke="#191970" points="1019.9236,-203.7316 1029.5972,-199.4106 1019.3403,-196.756 1019.9236,-203.7316"/>
</g>
-<!-- Node52->Node26 -->
+<!-- Node53->Node26 -->
<g id="edge156" class="edge">
-<title>Node52->Node26</title>
+<title>Node53->Node26</title>
<path fill="none" stroke="#191970" d="M2717.4,-777.8642C2647.9618,-746.2746 2425.9666,-640.4702 2271,-512 2180.5534,-437.0181 2190.8658,-383.6547 2097,-313 2068.1561,-291.2886 1835.6023,-189.2974 1801,-179 1720.583,-155.0686 1622.996,-142.4562 1568.3542,-136.806"/>
<polygon fill="#191970" stroke="#191970" points="1568.6406,-133.3172 1558.3406,-135.7991 1567.9403,-140.2821 1568.6406,-133.3172"/>
</g>
-<!-- Node52->Node32 -->
+<!-- Node53->Node33 -->
<g id="edge160" class="edge">
-<title>Node52->Node32</title>
+<title>Node53->Node33</title>
<path fill="none" stroke="#191970" d="M2697.3052,-787.435C2450.8808,-786.8936 1182.6135,-781.8863 796,-736 656.1161,-719.3975 486,-810.8658 486,-670 486,-670 486,-670 486,-558 486,-481.4471 448,-466.5529 448,-390 448,-390 448,-390 448,-261.5 448,-223.0012 448,-178.0145 448,-152.7812"/>
<polygon fill="#191970" stroke="#191970" points="451.5001,-152.6718 448,-142.6719 444.5001,-152.6719 451.5001,-152.6718"/>
</g>
-<!-- Node52->Node33 -->
+<!-- Node53->Node34 -->
<g id="edge141" class="edge">
-<title>Node52->Node33</title>
+<title>Node53->Node34</title>
<path fill="none" stroke="#191970" d="M2697.4285,-786.0913C2548.8337,-780.8847 2044.3795,-761.6923 1885,-736 1837.1061,-728.2794 1709.7068,-712.8398 1674,-680 1556.3115,-571.7609 1542.1903,-359.9768 1540.9301,-286.9612"/>
<polygon fill="#191970" stroke="#191970" points="1544.4288,-286.8095 1540.8255,-276.8463 1537.4292,-286.882 1544.4288,-286.8095"/>
</g>
-<!-- Node52->Node34 -->
+<!-- Node53->Node35 -->
<g id="edge144" class="edge">
-<title>Node52->Node34</title>
+<title>Node53->Node35</title>
<path fill="none" stroke="#191970" d="M2780.7288,-786.3089C2946.3558,-781.4075 3549.9213,-761.8312 3581,-736 3737.6603,-605.7906 3405.2632,-358.1933 3351,-313 3303.765,-273.6602 3288.1071,-266.0582 3230,-246 3150.4901,-218.5536 3054.7697,-205.6099 2988.657,-199.5848"/>
<polygon fill="#191970" stroke="#191970" points="2988.5753,-196.0642 2978.3075,-198.6761 2987.963,-203.0374 2988.5753,-196.0642"/>
</g>
-<!-- Node52->Node43 -->
+<!-- Node53->Node44 -->
<g id="edge142" class="edge">
-<title>Node52->Node43</title>
+<title>Node53->Node44</title>
<path fill="none" stroke="#191970" d="M2697.2261,-787.1227C2392.4464,-784.301 521.5957,-765.7849 472,-736 394.2095,-689.2826 372,-648.7407 372,-558 372,-558 372,-558 372,-502 372,-423.3248 363.905,-330.4527 359.6085,-286.7006"/>
<polygon fill="#191970" stroke="#191970" points="363.0836,-286.2759 358.6053,-276.6741 356.1184,-286.9729 363.0836,-286.2759"/>
</g>
-<!-- Node52->Node46 -->
+<!-- Node53->Node47 -->
<g id="edge158" class="edge">
-<title>Node52->Node46</title>
+<title>Node53->Node47</title>
<path fill="none" stroke="#191970" d="M2697.4556,-785.1418C2601.643,-779.3103 2367.7627,-762.72 2294,-736 2167.4404,-690.1547 2132.3806,-667.7519 2042,-568 1957.513,-474.7528 1898.5936,-332.159 1879.0641,-280.7849"/>
<polygon fill="#191970" stroke="#191970" points="1882.2818,-279.3969 1875.4996,-271.2574 1875.7257,-281.8498 1882.2818,-279.3969"/>
</g>
... 470418 lines suppressed ...