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 2023/03/17 09:32:09 UTC
[tvm-site] branch asf-site updated: deploying docs (apache/tvm@d1dea13762dca4a6285d7ae58542084fb34f5f3c)
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 270c18422a deploying docs (apache/tvm@d1dea13762dca4a6285d7ae58542084fb34f5f3c)
270c18422a is described below
commit 270c18422a9536bd678a57623b11e9e618052879
Author: tvm-bot <95...@users.noreply.github.com>
AuthorDate: Fri Mar 17 09:32:03 2023 +0000
deploying docs (apache/tvm@d1dea13762dca4a6285d7ae58542084fb34f5f3c)
---
docs/_images/sphx_glr_micro_train_001.png | Bin 364844 -> 343660 bytes
docs/_images/sphx_glr_micro_train_thumb.png | Bin 24068 -> 23847 bytes
.../how_to/compile_models/from_darknet.rst.txt | 2 +-
.../how_to/compile_models/from_keras.rst.txt | 2 +-
.../how_to/compile_models/from_mxnet.rst.txt | 2 +-
.../how_to/compile_models/from_oneflow.rst.txt | 2 +-
.../how_to/compile_models/from_pytorch.rst.txt | 2 +-
.../how_to/compile_models/from_tensorflow.rst.txt | 2 +-
.../compile_models/sg_execution_times.rst.txt | 22 +-
.../deploy_models/deploy_model_on_adreno.rst.txt | 2 +-
.../deploy_model_on_adreno_tvmc.rst.txt | 2 +-
.../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 | 22 +-
.../extend_tvm/bring_your_own_datatypes.rst.txt | 2 +-
.../how_to/extend_tvm/sg_execution_times.rst.txt | 10 +-
.../how_to/extend_tvm/use_pass_instrument.rst.txt | 16 +-
.../optimize_operators/opt_conv_cuda.rst.txt | 2 +-
.../optimize_operators/opt_conv_tensorcore.rst.txt | 2 +-
.../how_to/optimize_operators/opt_gemm.rst.txt | 16 +-
.../optimize_operators/sg_execution_times.rst.txt | 8 +-
.../sg_execution_times.rst.txt | 14 +-
.../tune_conv2d_layer_cuda.rst.txt | 1825 +++++++++++++++++---
.../tune_network_cuda.rst.txt | 4 +-
.../tune_network_x86.rst.txt | 4 +-
.../tune_sparse_x86.rst.txt | 78 +-
.../tune_with_autotvm/sg_execution_times.rst.txt | 4 +-
.../tune_with_autotvm/tune_conv2d_cuda.rst.txt | 708 +-------
.../work_with_microtvm/micro_autotune.rst.txt | 18 +-
.../work_with_microtvm/micro_pytorch.rst.txt | 4 +-
.../how_to/work_with_microtvm/micro_train.rst.txt | 18 +-
.../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 | 18 +-
.../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 | 11 +-
docs/_sources/tutorial/autotvm_matmul_x86.rst.txt | 20 +-
docs/_sources/tutorial/autotvm_relay_x86.rst.txt | 54 +-
.../tutorial/cross_compilation_and_rpc.rst.txt | 2 +-
docs/_sources/tutorial/intro_topi.rst.txt | 2 +-
docs/_sources/tutorial/sg_execution_times.rst.txt | 18 +-
.../tutorial/tensor_expr_get_started.rst.txt | 42 +-
docs/commit_hash | 2 +-
docs/how_to/compile_models/from_darknet.html | 2 +-
docs/how_to/compile_models/from_keras.html | 2 +-
docs/how_to/compile_models/from_mxnet.html | 2 +-
docs/how_to/compile_models/from_oneflow.html | 14 +-
docs/how_to/compile_models/from_pytorch.html | 9 +-
docs/how_to/compile_models/from_tensorflow.html | 2 +-
docs/how_to/compile_models/sg_execution_times.html | 22 +-
.../deploy_models/deploy_model_on_adreno.html | 2 +-
.../deploy_models/deploy_model_on_adreno_tvmc.html | 29 +-
.../deploy_models/deploy_model_on_android.html | 2 +-
.../deploy_object_detection_pytorch.html | 52 +-
docs/how_to/deploy_models/deploy_prequantized.html | 9 +-
.../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 | 38 +-
docs/how_to/deploy_models/sg_execution_times.html | 22 +-
.../extend_tvm/bring_your_own_datatypes.html | 2 +-
docs/how_to/extend_tvm/sg_execution_times.html | 10 +-
docs/how_to/extend_tvm/use_pass_instrument.html | 16 +-
docs/how_to/optimize_operators/opt_conv_cuda.html | 2 +-
.../optimize_operators/opt_conv_tensorcore.html | 2 +-
docs/how_to/optimize_operators/opt_gemm.html | 16 +-
.../optimize_operators/sg_execution_times.html | 8 +-
.../sg_execution_times.html | 14 +-
.../tune_conv2d_layer_cuda.html | 1825 +++++++++++++++++---
.../tune_with_autoscheduler/tune_network_cuda.html | 4 +-
.../tune_with_autoscheduler/tune_network_x86.html | 4 +-
.../tune_with_autoscheduler/tune_sparse_x86.html | 78 +-
.../tune_with_autotvm/sg_execution_times.html | 4 +-
.../how_to/tune_with_autotvm/tune_conv2d_cuda.html | 708 +-------
docs/how_to/work_with_microtvm/micro_autotune.html | 18 +-
docs/how_to/work_with_microtvm/micro_pytorch.html | 5 +-
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 | 18 +-
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 +-
docs/reference/api/typedoc/classes/instance.html | 58 +-
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 +-
.../api/typedoc/classes/runtimecontext.html | 22 +-
docs/reference/api/typedoc/classes/scalar.html | 6 +-
docs/reference/api/typedoc/classes/tvmarray.html | 16 +-
docs/reference/api/typedoc/classes/tvmobject.html | 12 +-
.../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 | 124 +-
.../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 | 8 +-
docs/tutorial/autotvm_matmul_x86.html | 20 +-
docs/tutorial/autotvm_relay_x86.html | 268 ++-
docs/tutorial/cross_compilation_and_rpc.html | 2 +-
docs/tutorial/intro_topi.html | 2 +-
docs/tutorial/sg_execution_times.html | 18 +-
docs/tutorial/tensor_expr_get_started.html | 42 +-
131 files changed, 4115 insertions(+), 2818 deletions(-)
diff --git a/docs/_images/sphx_glr_micro_train_001.png b/docs/_images/sphx_glr_micro_train_001.png
index 2b438e4497..4f3246f529 100644
Binary files a/docs/_images/sphx_glr_micro_train_001.png and b/docs/_images/sphx_glr_micro_train_001.png differ
diff --git a/docs/_images/sphx_glr_micro_train_thumb.png b/docs/_images/sphx_glr_micro_train_thumb.png
index d68a0ab65a..3ef5dcacc2 100644
Binary files a/docs/_images/sphx_glr_micro_train_thumb.png and b/docs/_images/sphx_glr_micro_train_thumb.png differ
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 4d7b730adc..80aaf6232e 100644
--- a/docs/_sources/how_to/compile_models/from_darknet.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_darknet.rst.txt
@@ -318,7 +318,7 @@ The process is no different from other examples.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 21.154 seconds)
+ **Total running time of the script:** ( 1 minutes 24.180 seconds)
.. _sphx_glr_download_how_to_compile_models_from_darknet.py:
diff --git a/docs/_sources/how_to/compile_models/from_keras.rst.txt b/docs/_sources/how_to/compile_models/from_keras.rst.txt
index 071e3e9832..e5a3cd3226 100644
--- a/docs/_sources/how_to/compile_models/from_keras.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_keras.rst.txt
@@ -232,7 +232,7 @@ Look up prediction top 1 index in 1000 class synset.
.. code-block:: none
Relay top-1 id: 285, class name: Egyptian cat
-
1/1 [==============================] - ETA: 0s
1/1 [==============================] - 1s 934ms/step
+
1/1 [==============================] - ETA: 0s
1/1 [==============================] - 1s 994ms/step
Keras top-1 id: 285, class name: Egyptian cat
diff --git a/docs/_sources/how_to/compile_models/from_mxnet.rst.txt b/docs/_sources/how_to/compile_models/from_mxnet.rst.txt
index 19a0081376..f988e5bd0d 100644
--- a/docs/_sources/how_to/compile_models/from_mxnet.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_mxnet.rst.txt
@@ -116,7 +116,7 @@ In this section, we download a pretrained imagenet model and classify an image.
.. code-block:: none
- Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zip7cfa7269-8f11-4505-8ca3-485484b15aee from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip...
+ Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zip6618f4c1-5f4e-4765-b2df-6c1622fe77fb 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 e0843acfd7..21f0bedf1b 100644
--- a/docs/_sources/how_to/compile_models/from_oneflow.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_oneflow.rst.txt
@@ -121,7 +121,7 @@ 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, 49.3MB/s]
27%|##6 | 11.0M/41.5M [00:00<00:00, 37.1MB/s]
35%|###5 | 14.7M/41.5M [00:00<00:00, 29.7MB/s]
42%|####2 | 17.6M/41.5M [00:00<00:00, 25.6MB/s]
58%|#####7 | 24.0M/41.5M [00:00<00:00, 25.1MB/s]
77%|#######7 | 32.0M/41.5M [00:01<00:00, 33.3MB/s]
96%|#########6| 40.0M/41.5M [00:01<00:00, 39.6MB/s]
100%|##########| 41.5M/41.5M [00:01<00:00, 35.8MB/s]
+
0%| | 0.00/41.5M [00:00<?, ?B/s]
19%|#9 | 7.99M/41.5M [00:00<00:00, 49.9MB/s]
39%|###8 | 16.0M/41.5M [00:00<00:00, 43.7MB/s]
58%|#####7 | 24.0M/41.5M [00:00<00:00, 40.7MB/s]
79%|#######8 | 32.7M/41.5M [00:00<00:00, 52.9MB/s]
96%|#########6| 40.0M/41.5M [00:00<00:00, 54.5MB/s]
100%|##########| 41.5M/41.5M [00:00<00:00, 51.9MB/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 e9f8174638..c185b360fc 100644
--- a/docs/_sources/how_to/compile_models/from_pytorch.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_pytorch.rst.txt
@@ -101,7 +101,7 @@ Load a pretrained PyTorch model
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing `weights=ResNet18_Weights.IMAGENET1K_V1`. You can also use `weights=ResNet18_Weights.DEFAULT` to get the most up-to-date weights.
warnings.warn(msg)
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]
18%|#7 | 7.99M/44.7M [00:00<00:00, 65.9MB/s]
46%|####6 | 20.7M/44.7M [00:00<00:00, 101MB/s]
69%|######8 | 30.7M/44.7M [00:00<00:00, 88.3MB/s]
90%|########9 | 40.0M/44.7M [00:00<00:00, 81.1MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 76.0MB/s]
+
0%| | 0.00/44.7M [00:00<?, ?B/s]
18%|#7 | 7.99M/44.7M [00:00<00:00, 76.1MB/s]
51%|##### | 22.7M/44.7M [00:00<00:00, 120MB/s]
77%|#######7 | 34.4M/44.7M [00:00<00:00, 105MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 81.4MB/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 bc911e4daa..cd2418dd65 100644
--- a/docs/_sources/how_to/compile_models/from_tensorflow.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_tensorflow.rst.txt
@@ -424,7 +424,7 @@ Run the corresponding model on tensorflow
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 23.708 seconds)
+ **Total running time of the script:** ( 1 minutes 26.994 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 8032a1d976..ed36a871b5 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
=================
-**06:42.364** total execution time for **how_to_compile_models** files:
+**06:56.986** total execution time for **how_to_compile_models** files:
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:23.708 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:26.994 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``) | 01:21.154 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``) | 01:24.180 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``) | 00:56.382 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``) | 00:59.830 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``) | 00:37.884 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``) | 00:39.149 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``) | 00:32.116 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``) | 00:33.675 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``) | 00:31.190 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``) | 00:31.903 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``) | 00:28.273 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``) | 00:28.568 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``) | 00:26.011 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``) | 00:27.021 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``) | 00:22.871 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``) | 00:22.868 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``) | 00:02.773 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``) | 00:02.800 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/deploy_models/deploy_model_on_adreno.rst.txt b/docs/_sources/how_to/deploy_models/deploy_model_on_adreno.rst.txt
index 13e8f6082e..ad43752198 100644
--- a/docs/_sources/how_to/deploy_models/deploy_model_on_adreno.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_model_on_adreno.rst.txt
@@ -637,7 +637,7 @@ well as provides information about the model's performance
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 2538.3758 2537.5783 2542.5970 2535.2333 2.4244
+ 2541.5621 2540.5103 2547.5712 2539.2064 2.7461
diff --git a/docs/_sources/how_to/deploy_models/deploy_model_on_adreno_tvmc.rst.txt b/docs/_sources/how_to/deploy_models/deploy_model_on_adreno_tvmc.rst.txt
index 921fdf418f..c1618f93ef 100644
--- a/docs/_sources/how_to/deploy_models/deploy_model_on_adreno_tvmc.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_model_on_adreno_tvmc.rst.txt
@@ -127,7 +127,7 @@ Make a Keras Resnet50 Model
.. code-block:: none
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels.h5
-
8192/102967424 [..............................] - ETA: 0s
5292032/102967424 [>.............................] - ETA: 0s
8470528/102967424 [=>............................] - ETA: 1s
12828672/102967424 [==>...........................] - ETA: 1s
16769024/102967424 [===>..........................] - ETA: 1s
22216704/102967424 [=====>........................] - ETA: 1s
25157632/102967424 [======>.......................] - ETA: 1s
27353088/102967424 [======>.......................] - ETA: 1s
33546240/102967424 [========>.....................] - ETA: 1s
35741696/102967424 [=========>....................] - ETA: 1s
41934848/102967424 [===========>..................] - ETA: 1s
46383104/102967424 [============>.................] - ETA: 1s
50323456/102967424 [=============>................] - ETA: 0s
50462720/102967424 [=============>................] - ETA: 1s
58712064/102967424 [================>.............] - ETA: 0s
67100672/102967424 [==================>...........] -
ETA: 0s
75489280/102967424 [====================>.........] - ETA: 0s
82124800/102967424 [======================>.......] - ETA: 0s
83877888/102967424 [=======================>......] - ETA: 0s
90497024/102967424 [=========================>....] - ETA: 0s
92266496/102967424 [=========================>....] - ETA: 0s
98910208/102967424 [===========================>..] - ETA: 0s
101621760/102967424 [============================>.] - ETA: 0s
102967424/102967424
[==============================] - 2s 0us/step
+
8192/102967424 [..............................] - ETA: 0s
8093696/102967424 [=>............................] - ETA: 0s
8839168/102967424 [=>............................] - ETA: 1s
16769024/102967424 [===>..........................] - ETA: 1s
23412736/102967424 [=====>........................] - ETA: 0s
25239552/102967424 [======>.......................] - ETA: 0s
33546240/102967424 [========>.....................] - ETA: 0s
35741696/102967424 [=========>....................] - ETA: 0s
40189952/102967424 [==========>...................] - ETA: 0s
41934848/102967424 [===========>..................] - ETA: 0s
50323456/102967424 [=============>................] - ETA: 0s
56967168/102967424 [===============>..............] - ETA: 0s
58712064/102967424 [================>.............] - ETA: 0s
64659456/102967424 [=================>............] - ETA: 0s
65355776/102967424 [==================>...........] - ETA: 0s
69083136/102967424 [===================>..........] -
ETA: 0s
72540160/102967424 [====================>.........] - ETA: 0s
75489280/102967424 [====================>.........] - ETA: 0s
82124800/102967424 [======================>.......] - ETA: 0s
83877888/102967424 [=======================>......] - ETA: 0s
92266496/102967424 [=========================>....] - ETA: 0s
98910208/102967424 [===========================>..] - ETA: 0s
102711296/102967424 [============================>.] - ETA: 0s
102875136/102967424
[============================>.] - ETA: 0s
102967424/102967424 [==============================] - 2s 0us/step
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 53a5a2bc71..85aa92cbf8 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
@@ -437,7 +437,7 @@ Execute on TVM
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 16.4630 16.3437 17.2113 15.9040 0.4360
+ 16.4419 16.2021 17.6050 16.1123 0.5233
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 67350ff48d..3cbd4a902a 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
@@ -130,7 +130,7 @@ Load pre-trained maskrcnn from torchvision and do tracing
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing `weights=MaskRCNN_ResNet50_FPN_Weights.COCO_V1`. You can also use `weights=MaskRCNN_ResNet50_FPN_Weights.DEFAULT` to get the most up-to-date weights.
warnings.warn(msg)
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]
5%|4 | 7.99M/170M [00:00<00:04, 41.5MB/s]
9%|9 | 16.0M/170M [00:00<00:05, 30.8MB/s]
14%|#4 | 24.0M/170M [00:00<00:03, 38.3MB/s]
19%|#8 | 32.0M/170M [00:00<00:03, 40.1MB/s]
24%|##3 | 40.0M/170M [00:01<00:03, 43.6MB/s]
28%|##8 | 48.0M/170M [00:01<00:02, 46.3MB/s]
33%|###2 | 56.0M/170M [00:01<00:02, 52.1MB/s]
38%|###7 | 64.0M/170M [00:01<00:02, 53.0MB/s]
42%|####2 | 72.0M/170M [00:01<00:01, 54.2MB/s]
47%|####7 | 80.0M/170M [00:01<00:01, 55.2MB/s]
52%|#####1 | 88.0M/170M [00:01<00:01, 54.4MB/s]
56%|#####5 | 94.3M/170M [00:02<00:01, 56.3MB/s]
59%|#####8 | 99.8M/170M [00:02<00:01, 55.4MB/s]
62%|######1 | 105M/170M [00:02<00:01, 47.0MB/s]
66%|######5 | 112M/170M [00:02<00:01, 45.4MB/s]
71%|####### | 120M/170M [00:02<00:01, 43.6MB/s]
75%|#######5 | 128M/170M [00:02<00:00, 45.4MB/s]
80%|######## | 136M/170M [00:03<00:00, 44.6MB/s]
85%|########4 | 144M/170M [00:03<00:00, 42.9MB/s]
88%|########8 | 150M/170M [00:03<00:00, 40.3MB/s]
91%|######### | 154M/170M [00:03<00:00, 37.7MB/s]
93%|#########3| 158M/170M [00:03<00:00, 38.3MB/s]
95%|#########5| 162M/170M [00:03<00:00, 38.7MB/s]
99%|#########8| 168M/170M [00:03<00:00, 41.4MB/s]
100%|##########| 170M/170M [00:03<00:00, 45.0MB/s]
+
0%| | 0.00/170M [00:00<?, ?B/s]
3%|2 | 4.85M/170M [00:00<00:03, 50.9MB/s]
6%|5 | 9.70M/170M [00:00<00:06, 28.0MB/s]
9%|9 | 16.0M/170M [00:00<00:04, 38.9MB/s]
14%|#4 | 24.0M/170M [00:00<00:03, 47.7MB/s]
19%|#8 | 31.6M/170M [00:00<00:02, 56.7MB/s]
22%|##2 | 37.5M/170M [00:00<00:02, 55.6MB/s]
27%|##6 | 45.2M/170M [00:00<00:02, 62.4MB/s]
30%|### | 51.5M/170M [00:01<00:02, 46.6MB/s]
33%|###3 | 56.6M/170M [00:01<00:02, 42.1MB/s]
38%|###7 | 64.0M/170M [00:01<00:02, 44.0MB/s]
45%|####5 | 77.0M/170M [00:01<00:01, 64.1MB/s]
50%|####9 | 84.1M/170M [00:01<00:01, 61.5MB/s]
53%|#####3 | 90.6M/170M [00:01<00:01, 50.9MB/s]
57%|#####6 | 96.2M/170M [00:02<00:02, 34.6MB/s]
60%|###### | 102M/170M [00:02<00:01, 37.6MB/s]
63%|######3 | 107M/170M [00:02<00:01, 40.0MB/s]
66%|######6 | 112M/170M [00:02<00:01, 42.5MB/s]
71%|####### | 120M/170M [00:02<00:01, 48.8MB/s]
74%|#######4 | 126M/170M [00:02<00:00, 47.5MB/s]
80%|######## | 136M/170M [00:03<00:00, 51.7MB/s]
86%|########5 | 146M/170M [00:03<00:00, 59.3MB/s]
89%|########9 | 152M/170M [00:03<00:00, 59.4MB/s]
94%|#########4| 160M/170M [00:03<00:00, 58.0MB/s]
99%|#########8| 168M/170M [00:03<00:00, 57.2MB/s]
100%|##########| 170M/170M [00:03<00:00, 50.1MB/s]
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/torch/nn/functional.py:3897: 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)
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/torchvision/models/detection/anchor_utils.py:124: 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').
@@ -299,7 +299,7 @@ Get boxes with score larger than 0.9
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 3 minutes 36.224 seconds)
+ **Total running time of the script:** ( 3 minutes 48.383 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 b1824c11d7..0127a98ca0 100644
--- a/docs/_sources/how_to/deploy_models/deploy_prequantized.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_prequantized.rst.txt
@@ -227,7 +227,7 @@ training. Other models require a full post training calibration.
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing `weights=MobileNet_V2_Weights.IMAGENET1K_V1`. You can also use `weights=MobileNet_V2_Weights.DEFAULT` to get the most up-to-date weights.
warnings.warn(msg)
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]
57%|#####6 | 7.68M/13.6M [00:00<00:00, 51.2MB/s]
93%|#########2| 12.6M/13.6M [00:00<00:00, 38.0MB/s]
100%|##########| 13.6M/13.6M [00:00<00:00, 42.3MB/s]
+
0%| | 0.00/13.6M [00:00<?, ?B/s]
59%|#####8 | 7.99M/13.6M [00:00<00:00, 57.8MB/s]
100%|##########| 13.6M/13.6M [00:00<00:00, 66.1MB/s]
@@ -409,7 +409,7 @@ 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)
- 90.1348 90.0830 94.1666 89.8210 0.4250
+ 90.3151 90.2691 90.8183 90.0883 0.1735
@@ -458,7 +458,7 @@ TODO
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 16.961 seconds)
+ **Total running time of the script:** ( 1 minutes 20.333 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 dd16530cd4..56c8249a0e 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
@@ -423,7 +423,7 @@ 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)
- 119.6617 119.8028 120.9767 117.9153 0.6340
+ 120.5423 120.4685 125.6759 119.5865 0.6641
@@ -460,7 +460,7 @@ 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:** ( 2 minutes 29.156 seconds)
+ **Total running time of the script:** ( 2 minutes 31.977 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 056387eda3..a362e7a08a 100644
--- a/docs/_sources/how_to/deploy_models/deploy_quantized.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_quantized.rst.txt
@@ -257,7 +257,7 @@ We create a Relay VM to build and execute the model.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 43.517 seconds)
+ **Total running time of the script:** ( 1 minutes 42.223 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 f13daca951..6e3917a874 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
@@ -170,7 +170,7 @@ Convert and compile model for CPU.
data: None
input_sym_arg_type = in_param.infer_type()[0]
Downloading /workspace/.mxnet/models/ssd_512_resnet50_v1_voc-9c8b225a.zip from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/ssd_512_resnet50_v1_voc-9c8b225a.zip...
-
0%| | 0/132723 [00:00<?, ?KB/s]
5%|4 | 6315/132723 [00:00<00:02, 63139.97KB/s]
11%|#1 | 14790/132723 [00:00<00:01, 75844.69KB/s]
17%|#6 | 22375/132723 [00:00<00:01, 68936.78KB/s]
22%|##2 | 29470/132723 [00:00<00:01, 60860.59KB/s]
27%|##6 | 35699/132723 [00:00<00:01, 50204.26KB/s]
33%|###3 | 44062/132723 [00:00<00:01, 59246.80KB/s]
38%|###7 | 50407/132723 [00:00<00:01, 58438.31KB/s]
44%|####4 | 58798/132723 [00:00<00:01, 65518.96KB/s]
50%|##### | 66934/132723 [00:01<00:00, 70021.38KB/s]
57%|#####6 | 75433/132723 [00:01<00:00, 74341.81KB/s]
63%|######2 | 83067/132723 [00:01<00:00, 66702.48KB/s]
69%|######8 | 91466/132723 [00:01<00:00, 71382.30KB/s]
75%|#######5 | 99853/132723 [00:01<00:00, 74863.44KB/s]
82%|########1 | 108184/132723 [00:01<00:00, 77268.05KB/s]
87%|########7 | 116066/132723 [00:01<00:00, 76981.59KB/s]
94%|#########
3| 124530/132723 [00:01<00:00, 79203.96KB/s]
100%|#########9| 132534/132723 [00:01<00:00, 72683.40KB/s]
100%|##########| 132723/132723 [00:01<00:00, 68954.83KB/s]
+
0%| | 0/132723 [00:00<?, ?KB/s]
3%|2 | 3682/132723 [00:00<00:03, 36817.58KB/s]
8%|8 | 10682/132723 [00:00<00:02, 56330.46KB/s]
14%|#3 | 18572/132723 [00:00<00:01, 66632.38KB/s]
20%|#9 | 26510/132723 [00:00<00:01, 71660.76KB/s]
25%|##5 | 33677/132723 [00:00<00:01, 67776.40KB/s]
31%|###1 | 41468/132723 [00:00<00:01, 71094.61KB/s]
37%|###7 | 49280/132723 [00:00<00:01, 71541.97KB/s]
43%|####3 | 57112/132723 [00:00<00:01, 73622.94KB/s]
49%|####9 | 65102/132723 [00:00<00:00, 75536.45KB/s]
55%|#####5 | 73005/132723 [00:01<00:00, 76594.96KB/s]
61%|######1 | 80985/132723 [00:01<00:00, 77562.88KB/s]
67%|######7 | 89008/132723 [00:01<00:00, 78363.15KB/s]
73%|#######3 | 97044/132723 [00:01<00:00, 78962.42KB/s]
79%|#######9 | 104947/132723 [00:01<00:00, 70184.36KB/s]
85%|########5 | 112902/132723 [00:01<00:00, 72770.75KB/s]
91%|#########
| 120572/132723 [00:01<00:00, 73881.25KB/s]
97%|#########6| 128556/132723 [00:01<00:00, 75593.83KB/s]
100%|##########| 132723/132723 [00:01<00:00, 73010.88KB/s]
@@ -246,7 +246,7 @@ Display result
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 3 minutes 45.898 seconds)
+ **Total running time of the script:** ( 3 minutes 56.461 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 569ed88eef..d4ad9f254b 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
@@ -5,28 +5,28 @@
Computation times
=================
-**16:16.748** total execution time for **how_to_deploy_models** files:
+**16:51.562** total execution time for **how_to_deploy_models** files:
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``) | 03:45.898 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``) | 03:56.461 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 03:36.224 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 03:48.383 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``) | 02:29.156 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``) | 02:31.977 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``) | 01:43.517 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``) | 01:42.223 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``) | 01:16.961 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``) | 01:20.333 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno.py` (``deploy_model_on_adreno.py``) | 00:54.886 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno.py` (``deploy_model_on_adreno.py``) | 00:56.111 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno_tvmc.py` (``deploy_model_on_adreno_tvmc.py``) | 00:51.589 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno_tvmc.py` (``deploy_model_on_adreno_tvmc.py``) | 00:53.265 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``) | 00:42.608 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``) | 00:44.676 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``) | 00:28.162 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``) | 00:29.423 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``) | 00:27.741 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``) | 00:28.704 | 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 665dc847c9..55a6a68c1c 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
@@ -463,7 +463,7 @@ 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.zip80284e7a-7990-4b90-baf4-58da2c8595f7 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
+ Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zipe820fe04-b70a-4359-a47c-6cf9420def47 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
diff --git a/docs/_sources/how_to/extend_tvm/sg_execution_times.rst.txt b/docs/_sources/how_to/extend_tvm/sg_execution_times.rst.txt
index 484b5ab9f1..8dabcd0372 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
=================
-**00:54.320** total execution time for **how_to_extend_tvm** files:
+**00:57.036** 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``) | 00:50.530 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_bring_your_own_datatypes.py` (``bring_your_own_datatypes.py``) | 00:53.001 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``) | 00:02.709 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``) | 00:02.890 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_infra.py` (``use_pass_infra.py``) | 00:01.074 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_infra.py` (``use_pass_infra.py``) | 00:01.137 | 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 944f32edf4..7f01376db5 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
@@ -220,10 +220,10 @@ profile the execution time of each passes.
.. code-block:: none
Printing results of timing profile...
- InferType: 22219us [22219us] (48.74%; 48.74%)
- FoldScaleAxis: 23364us [7us] (51.26%; 51.26%)
- FoldConstant: 23357us [1666us] (51.24%; 99.97%)
- InferType: 21691us [21691us] (47.59%; 92.87%)
+ InferType: 22452us [22452us] (47.72%; 47.72%)
+ FoldScaleAxis: 24594us [9us] (52.28%; 52.28%)
+ FoldConstant: 24585us [1726us] (52.26%; 99.96%)
+ InferType: 22859us [22859us] (48.59%; 92.98%)
@@ -262,10 +262,10 @@ Refer to following sections and :py:func:`tvm.instrument.pass_instrument` for th
.. code-block:: none
Printing results of timing profile...
- InferType: 21601us [21601us] (48.24%; 48.24%)
- FoldScaleAxis: 23175us [5us] (51.76%; 51.76%)
- FoldConstant: 23170us [1708us] (51.75%; 99.98%)
- InferType: 21463us [21463us] (47.93%; 92.63%)
+ InferType: 22534us [22534us] (48.18%; 48.18%)
+ FoldScaleAxis: 24235us [8us] (51.82%; 51.82%)
+ FoldConstant: 24227us [1791us] (51.80%; 99.97%)
+ InferType: 22436us [22436us] (47.97%; 92.61%)
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 14951d78d1..6150002491 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
@@ -331,7 +331,7 @@ latency of convolution.
.. code-block:: none
- Convolution: 54.182239 ms
+ Convolution: 54.190113 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 358e067eb7..5da5c1314a 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
@@ -598,7 +598,7 @@ be able to run on our build server
.. code-block:: none
- conv2d with tensor core: 13.373636 ms
+ conv2d with tensor core: 13.361943 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 f059823630..51774e3391 100644
--- a/docs/_sources/how_to/optimize_operators/opt_gemm.rst.txt
+++ b/docs/_sources/how_to/optimize_operators/opt_gemm.rst.txt
@@ -134,8 +134,8 @@ Then we write a baseline implementation, the simplest way to write a matrix mult
.. code-block:: none
- Numpy running time: 0.018695
- Baseline: 3.251712
+ Numpy running time: 0.019399
+ Baseline: 3.484613
@@ -227,7 +227,7 @@ fill 32 * 32 * sizeof(float) which is 4KB in the cache whose total size is 32KB
.. code-block:: none
- Opt1: 0.296724
+ Opt1: 0.331902
@@ -318,7 +318,7 @@ In this tutorial, we chose to vectorize the inner loop row data since it is cach
.. code-block:: none
- Opt2: 0.342606
+ Opt2: 0.350420
@@ -406,7 +406,7 @@ the access pattern for A matrix is more cache friendly.
.. code-block:: none
- Opt3: 0.114652
+ Opt3: 0.122769
@@ -523,7 +523,7 @@ flattening.
.. code-block:: none
- Opt4: 0.110320
+ Opt4: 0.109488
@@ -635,7 +635,7 @@ write to C when all the block results are ready.
.. code-block:: none
- Opt5: 0.111773
+ Opt5: 0.112409
@@ -748,7 +748,7 @@ Furthermore, we can also utilize multi-core processors to do the thread-level pa
.. code-block:: none
- Opt6: 0.147087
+ Opt6: 0.148717
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 ecb2880069..b44e13463d 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
=================
-**00:34.823** total execution time for **how_to_optimize_operators** files:
+**00:36.149** total execution time for **how_to_optimize_operators** files:
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``) | 00:31.905 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``) | 00:33.396 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.706 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.610 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``) | 00:01.212 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``) | 00:01.143 | 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 b8758bd72e..9066636f80 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
=================
-**09:58.198** total execution time for **how_to_tune_with_autoscheduler** files:
+**10:20.053** 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``) | 06:09.023 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_conv2d_layer_cuda.py` (``tune_conv2d_layer_cuda.py``) | 06:24.225 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``) | 01:41.602 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``) | 01:44.830 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``) | 01:07.297 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``) | 01:09.353 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``) | 00:32.629 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``) | 00:32.826 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``) | 00:14.156 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``) | 00:14.692 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``) | 00:13.491 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``) | 00:14.127 | 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 0f89241f54..4c062034a6 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
@@ -245,183 +245,840 @@ cooperative fetching, unrolling and operator fusion.
T.func_attr({"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True})
blockIdx_x = T.launch_thread("blockIdx.x", 32)
conv2d_nchw = T.allocate([14], "float32", "local")
- pad_temp_shared = T.allocate([2592], "float32", "shared")
- kernel_shared = T.allocate([4608], "float32", "shared")
+ pad_temp_shared = T.allocate([2016], "float32", "shared")
+ kernel_shared = T.allocate([1536], "float32", "shared")
threadIdx_x = T.launch_thread("threadIdx.x", 56)
- conv2d_nchw_1 = T.Buffer((14,), data=conv2d_nchw, scope="local", align=32)
+ conv2d_nchw_1 = T.Buffer((49,), data=conv2d_nchw, scope="local", align=16)
conv2d_nchw_1[0] = T.float32(0)
- conv2d_nchw_1[1] = T.float32(0)
- conv2d_nchw_1[2] = T.float32(0)
- conv2d_nchw_1[3] = T.float32(0)
- conv2d_nchw_1[4] = T.float32(0)
- conv2d_nchw_1[5] = T.float32(0)
- conv2d_nchw_1[6] = T.float32(0)
conv2d_nchw_1[7] = T.float32(0)
+ conv2d_nchw_1[1] = T.float32(0)
conv2d_nchw_1[8] = T.float32(0)
+ conv2d_nchw_1[2] = T.float32(0)
conv2d_nchw_1[9] = T.float32(0)
+ conv2d_nchw_1[3] = T.float32(0)
conv2d_nchw_1[10] = T.float32(0)
+ conv2d_nchw_1[4] = T.float32(0)
conv2d_nchw_1[11] = T.float32(0)
+ conv2d_nchw_1[5] = T.float32(0)
conv2d_nchw_1[12] = T.float32(0)
+ conv2d_nchw_1[6] = T.float32(0)
conv2d_nchw_1[13] = T.float32(0)
- for rc_outer_outer in range(16):
- cse_var_1: T.int32 = rc_outer_outer * 1568
+ for rc_outer_outer, rx_outer_outer in T.grid(16, 3):
+ cse_var_2: T.int32 = rc_outer_outer * 1568
+ cse_var_1: T.int32 = rc_outer_outer * 288
threadIdx_x_1 = T.env_thread("threadIdx.x")
- pad_temp_shared_1 = T.Buffer((2592,), data=pad_temp_shared, scope="shared")
+ pad_temp_shared_1 = T.Buffer((2016,), data=pad_temp_shared, scope="shared")
data_1 = T.Buffer((25088,), data=data.data)
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(9 <= threadIdx_x_1 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + threadIdx_x_1 // 9 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 56] = T.if_then_else(9 <= (threadIdx_x_1 + 56) % 81 and (threadIdx_x_1 + 56) % 81 < 72 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 56) // 81 * 49 + (threadIdx_x_1 + 56) % 81 // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 112] = T.if_then_else(9 <= (threadIdx_x_1 + 31) % 81 and (threadIdx_x_1 + 31) % 81 < 72 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 112) // 81 * 49 + (threadIdx_x_1 + 31) % 81 // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 168] = T.if_then_else(3 <= threadIdx_x_1 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 168) // 81 * 49 + (threadIdx_x_1 + 6) // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 224] = T.if_then_else(9 <= (threadIdx_x_1 + 62) % 81 and (threadIdx_x_1 + 62) % 81 < 72 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 224) // 81 * 49 + (threadIdx_x_1 + 62) % 81 // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 280] = T.if_then_else(9 <= (threadIdx_x_1 + 37) % 81 and (threadIdx_x_1 + 37) % 81 < 72 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 280) // 81 * 49 + (threadIdx_x_1 + 37) % 81 // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 336] = T.if_then_else(1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 336) // 81 * 49 + (threadIdx_x_1 + 12) // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 392] = T.if_then_else(9 <= (threadIdx_x_1 + 68) % 81 and (threadIdx_x_1 + 68) % 81 < 72 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 392) // 81 * 49 + (threadIdx_x_1 + 68) % 81 // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 448] = T.if_then_else(9 <= (threadIdx_x_1 + 43) % 81 and (threadIdx_x_1 + 43) % 81 < 72 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 448) // 81 * 49 + (threadIdx_x_1 + 43) % 81 // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 504] = T.if_then_else(threadIdx_x_1 < 54 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 504) // 81 * 49 + threadIdx_x_1 // 9 * 7 + threadIdx_x_1 % 9 + 6], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(7 <= threadIdx_x_1 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + threadIdx_x_1 + rx_outer_outer - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 560] = T.if_then_else(9 <= (threadIdx_x_1 + 74) % 81 and (threadIdx_x_1 + 74) % 81 < 72 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 560) // 81 * 49 + (threadIdx_x_1 + 74) % 81 // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 56] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 8) % 9 and (threadIdx_x_1 // 7 + 8) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 56) // 63 * 49 + (threadIdx_x_1 // 7 + 8) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 616] = T.if_then_else(9 <= (threadIdx_x_1 + 49) % 81 and (threadIdx_x_1 + 49) % 81 < 72 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 616) // 81 * 49 + (threadIdx_x_1 + 49) % 81 // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 112] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 7) % 9 and (threadIdx_x_1 // 7 + 7) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 112) // 63 * 49 + (threadIdx_x_1 // 7 + 7) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 672] = T.if_then_else(threadIdx_x_1 < 48 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 672) // 81 * 49 + (threadIdx_x_1 + 24) // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 168] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 6) % 9 and (threadIdx_x_1 // 7 + 6) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 168) // 63 * 49 + (threadIdx_x_1 // 7 + 6) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 728] = T.if_then_else(9 <= (threadIdx_x_1 + 80) % 81 and (threadIdx_x_1 + 80) % 81 < 72 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 728) // 81 * 49 + (threadIdx_x_1 + 80) % 81 // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 224] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 5) % 9 and (threadIdx_x_1 // 7 + 5) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 224) // 63 * 49 + (threadIdx_x_1 // 7 + 5) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 784] = T.if_then_else(9 <= (threadIdx_x_1 + 55) % 81 and (threadIdx_x_1 + 55) % 81 < 72 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 784) // 81 * 49 + (threadIdx_x_1 + 55) % 81 // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 280] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 4) % 9 and (threadIdx_x_1 // 7 + 4) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 280) // 63 * 49 + (threadIdx_x_1 // 7 + 4) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 840] = T.if_then_else(9 <= (threadIdx_x_1 + 30) % 81 and (threadIdx_x_1 + 30) % 81 < 72 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 840) // 81 * 49 + (threadIdx_x_1 + 30) % 81 // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 336] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 3) % 9 and (threadIdx_x_1 // 7 + 3) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 336) // 63 * 49 + (threadIdx_x_1 // 7 + 3) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 896] = T.if_then_else(4 <= threadIdx_x_1 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 896) // 81 * 49 + (threadIdx_x_1 + 5) // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 392] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 2) % 9 and (threadIdx_x_1 // 7 + 2) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 392) // 63 * 49 + (threadIdx_x_1 // 7 + 2) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 952] = T.if_then_else(9 <= (threadIdx_x_1 + 61) % 81 and (threadIdx_x_1 + 61) % 81 < 72 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 952) // 81 * 49 + (threadIdx_x_1 + 61) % 81 // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 448] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 448) // 63 * 49 + threadIdx_x_1 + rx_outer_outer - 1], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1008] = T.if_then_else(1 <= (threadIdx_x_1 // 9 + 4) % 9 and (threadIdx_x_1 + 36) % 81 < 72 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1008) // 81 * 49 + (threadIdx_x_1 // 9 + 4) % 9 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 504] = T.if_then_else(7 <= threadIdx_x_1 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + threadIdx_x_1 + rx_outer_outer + 384], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1064] = T.if_then_else(1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1064) // 81 * 49 + (threadIdx_x_1 + 11) // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 560] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 8) % 9 and (threadIdx_x_1 // 7 + 8) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 560) // 63 * 49 + (threadIdx_x_1 // 7 + 8) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1120] = T.if_then_else(9 <= (threadIdx_x_1 + 67) % 81 and (threadIdx_x_1 + 67) % 81 < 72 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1120) // 81 * 49 + (threadIdx_x_1 + 67) % 81 // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 616] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 7) % 9 and (threadIdx_x_1 // 7 + 7) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 616) // 63 * 49 + (threadIdx_x_1 // 7 + 7) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1176] = T.if_then_else(9 <= (threadIdx_x_1 + 42) % 81 and (threadIdx_x_1 + 42) % 81 < 72 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1176) // 81 * 49 + (threadIdx_x_1 + 42) % 81 // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 672] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 6) % 9 and (threadIdx_x_1 // 7 + 6) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 672) // 63 * 49 + (threadIdx_x_1 // 7 + 6) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1232] = T.if_then_else(threadIdx_x_1 < 55 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1232) // 81 * 49 + (threadIdx_x_1 + 17) // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 728] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 5) % 9 and (threadIdx_x_1 // 7 + 5) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 728) // 63 * 49 + (threadIdx_x_1 // 7 + 5) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1288] = T.if_then_else(9 <= (threadIdx_x_1 + 73) % 81 and (threadIdx_x_1 + 73) % 81 < 72 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1288) // 81 * 49 + (threadIdx_x_1 + 73) % 81 // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 784] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 4) % 9 and (threadIdx_x_1 // 7 + 4) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 784) // 63 * 49 + (threadIdx_x_1 // 7 + 4) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1344] = T.if_then_else(9 <= (threadIdx_x_1 + 48) % 81 and (threadIdx_x_1 + 48) % 81 < 72 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1344) // 81 * 49 + (threadIdx_x_1 + 48) % 81 // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 840] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 3) % 9 and (threadIdx_x_1 // 7 + 3) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 840) // 63 * 49 + (threadIdx_x_1 // 7 + 3) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1400] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1400) // 81 * 49 + (threadIdx_x_1 + 23) // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 896] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 2) % 9 and (threadIdx_x_1 // 7 + 2) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 896) // 63 * 49 + (threadIdx_x_1 // 7 + 2) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1456] = T.if_then_else(9 <= (threadIdx_x_1 + 79) % 81 and (threadIdx_x_1 + 79) % 81 < 72 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1456) // 81 * 49 + (threadIdx_x_1 + 79) % 81 // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 952] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 952) // 63 * 49 + threadIdx_x_1 + rx_outer_outer - 1], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1512] = T.if_then_else(1 <= (threadIdx_x_1 // 9 + 6) % 9 and (threadIdx_x_1 + 54) % 81 < 72 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1512) // 81 * 49 + (threadIdx_x_1 // 9 + 6) % 9 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1008] = T.if_then_else(7 <= threadIdx_x_1 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + threadIdx_x_1 + rx_outer_outer + 776], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1568] = T.if_then_else(9 <= (threadIdx_x_1 + 29) % 81 and (threadIdx_x_1 + 29) % 81 < 72 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1568) // 81 * 49 + (threadIdx_x_1 + 29) % 81 // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1064] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 8) % 9 and (threadIdx_x_1 // 7 + 8) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1064) // 63 * 49 + (threadIdx_x_1 // 7 + 8) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1624] = T.if_then_else(5 <= threadIdx_x_1 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1624) // 81 * 49 + (threadIdx_x_1 + 4) // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1120] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 7) % 9 and (threadIdx_x_1 // 7 + 7) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1120) // 63 * 49 + (threadIdx_x_1 // 7 + 7) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1680] = T.if_then_else(9 <= (threadIdx_x_1 + 60) % 81 and (threadIdx_x_1 + 60) % 81 < 72 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1680) // 81 * 49 + (threadIdx_x_1 + 60) % 81 // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1176] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 6) % 9 and (threadIdx_x_1 // 7 + 6) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1176) // 63 * 49 + (threadIdx_x_1 // 7 + 6) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1736] = T.if_then_else(9 <= (threadIdx_x_1 + 35) % 81 and (threadIdx_x_1 + 35) % 81 < 72 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1736) // 81 * 49 + (threadIdx_x_1 + 35) % 81 // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1232] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 5) % 9 and (threadIdx_x_1 // 7 + 5) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1232) // 63 * 49 + (threadIdx_x_1 // 7 + 5) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1792] = T.if_then_else(1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1792) // 81 * 49 + (threadIdx_x_1 + 10) // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1288] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 4) % 9 and (threadIdx_x_1 // 7 + 4) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1288) // 63 * 49 + (threadIdx_x_1 // 7 + 4) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1848] = T.if_then_else(9 <= (threadIdx_x_1 + 66) % 81 and (threadIdx_x_1 + 66) % 81 < 72 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1848) // 81 * 49 + (threadIdx_x_1 + 66) % 81 // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1344] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 3) % 9 and (threadIdx_x_1 // 7 + 3) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1344) // 63 * 49 + (threadIdx_x_1 // 7 + 3) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1904] = T.if_then_else(9 <= (threadIdx_x_1 + 41) % 81 and (threadIdx_x_1 + 41) % 81 < 72 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1904) // 81 * 49 + (threadIdx_x_1 + 41) % 81 // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1400] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 2) % 9 and (threadIdx_x_1 // 7 + 2) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1400) // 63 * 49 + (threadIdx_x_1 // 7 + 2) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1960] = T.if_then_else(1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1960) // 81 * 49 + (threadIdx_x_1 + 16) // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1456] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1456) // 63 * 49 + threadIdx_x_1 + rx_outer_outer - 1], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2016] = T.if_then_else(1 <= (threadIdx_x_1 // 9 + 8) % 9 and (threadIdx_x_1 + 72) % 81 < 72 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2016) // 81 * 49 + (threadIdx_x_1 // 9 + 8) % 9 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1512] = T.if_then_else(7 <= threadIdx_x_1 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + threadIdx_x_1 + rx_outer_outer + 1168], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2072] = T.if_then_else(9 <= (threadIdx_x_1 + 47) % 81 and (threadIdx_x_1 + 47) % 81 < 72 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2072) // 81 * 49 + (threadIdx_x_1 + 47) % 81 // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1568] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 8) % 9 and (threadIdx_x_1 // 7 + 8) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1568) // 63 * 49 + (threadIdx_x_1 // 7 + 8) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2128] = T.if_then_else(threadIdx_x_1 < 50 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2128) // 81 * 49 + (threadIdx_x_1 + 22) // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1624] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 7) % 9 and (threadIdx_x_1 // 7 + 7) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1624) // 63 * 49 + (threadIdx_x_1 // 7 + 7) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2184] = T.if_then_else(9 <= (threadIdx_x_1 + 78) % 81 and (threadIdx_x_1 + 78) % 81 < 72 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2184) // 81 * 49 + (threadIdx_x_1 + 78) % 81 // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1680] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 6) % 9 and (threadIdx_x_1 // 7 + 6) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1680) // 63 * 49 + (threadIdx_x_1 // 7 + 6) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2240] = T.if_then_else(9 <= (threadIdx_x_1 + 53) % 81 and (threadIdx_x_1 + 53) % 81 < 72 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2240) // 81 * 49 + (threadIdx_x_1 + 53) % 81 // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1736] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 5) % 9 and (threadIdx_x_1 // 7 + 5) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1736) // 63 * 49 + (threadIdx_x_1 // 7 + 5) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2296] = T.if_then_else(9 <= (threadIdx_x_1 + 28) % 81 and (threadIdx_x_1 + 28) % 81 < 72 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2296) // 81 * 49 + (threadIdx_x_1 + 28) % 81 // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1792] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 4) % 9 and (threadIdx_x_1 // 7 + 4) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1792) // 63 * 49 + (threadIdx_x_1 // 7 + 4) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2352] = T.if_then_else(6 <= threadIdx_x_1 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2352) // 81 * 49 + (threadIdx_x_1 + 3) // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1848] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 3) % 9 and (threadIdx_x_1 // 7 + 3) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1848) // 63 * 49 + (threadIdx_x_1 // 7 + 3) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2408] = T.if_then_else(9 <= (threadIdx_x_1 + 59) % 81 and (threadIdx_x_1 + 59) % 81 < 72 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2408) // 81 * 49 + (threadIdx_x_1 + 59) % 81 // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1904] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 2) % 9 and (threadIdx_x_1 // 7 + 2) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1904) // 63 * 49 + (threadIdx_x_1 // 7 + 2) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2464] = T.if_then_else(9 <= (threadIdx_x_1 + 34) % 81 and (threadIdx_x_1 + 34) % 81 < 72 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2464) // 81 * 49 + (threadIdx_x_1 + 34) % 81 // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2520] = T.if_then_else(1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2520) // 81 * 49 + threadIdx_x_1 // 9 * 7 + threadIdx_x_1 % 9 - 1], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- if T.likely(threadIdx_x_1 < 16):
- pad_temp_shared_1[threadIdx_x_1 + 2576] = T.if_then_else(threadIdx_x_1 < 7 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2576) // 81 * 49 + (threadIdx_x_1 + 65) // 9 * 7 + threadIdx_x_1 - 6], T.float32(0))
- kernel_shared_1 = T.Buffer((4608,), data=kernel_shared, scope="shared")
- for ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer in range(21):
- threadIdx_x_2 = T.launch_thread("threadIdx.x", 56)
- kernel_1 = T.Buffer((2359296,), data=kernel.data)
- if T.likely(ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8 < 144):
- kernel_shared_1[ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4] = kernel_1[blockIdx_x * 73728 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8) // 9 * 4608 + rc_outer_outer * 288 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4) % 288]
- if T.likely(ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8 < 144):
- kernel_shared_1[ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 1] = kernel_1[blockIdx_x * 73728 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8) // 9 * 4608 + rc_outer_outer * 288 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 1) % 288]
- if T.likely(ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8 < 144):
- kernel_shared_1[ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 2] = kernel_1[blockIdx_x * 73728 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8) // 9 * 4608 + rc_outer_outer * 288 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 2) % 288]
- if T.likely(ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8 < 144):
- kernel_shared_1[ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 3] = kernel_1[blockIdx_x * 73728 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8) // 9 * 4608 + rc_outer_outer * 288 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 3) % 288]
- for rc_outer_inner, ry_outer_inner, rc_inner in T.grid(8, 3, 4):
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 9] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 18] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 27] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 36] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 45] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 54] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 9] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 18] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 27] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 36] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 45] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 54] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 1] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 10] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 19] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 28] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 37] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 46] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 55] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 1] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 10] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 19] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 28] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 37] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 46] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 55] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 2] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 11] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 20] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 29] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 38] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 47] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 56] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 2] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 11] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 20] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 29] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 38] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 47] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 56] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- for i1_inner, i2_inner in T.grid(2, 7):
+ pad_temp_shared_1[threadIdx_x_1 + 1960] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1960) // 63 * 49 + threadIdx_x_1 + rx_outer_outer - 1], T.float32(0))
+ threadIdx_x_2 = T.env_thread("threadIdx.x")
+ kernel_shared_1 = T.Buffer((1536,), data=kernel_shared, scope="shared")
+ kernel_1 = T.Buffer((2359296,), data=kernel.data)
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2] = kernel_1[blockIdx_x * 73728 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 56] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 56) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 112] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 112) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 48]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 168] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 168) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 72) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 224] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 224) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 96]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 280] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 280) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 88) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 336] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 336) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 392] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 392) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 24]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 448] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 448) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 504] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 504) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 72]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 560] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 560) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 616] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 616) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 120]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 672] = kernel_1[blockIdx_x * 73728 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 32256]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 728] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 728) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 784] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 784) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 48]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 840] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 840) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 72) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 896] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 896) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 96]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 952] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 952) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 88) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1008] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1008) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1064] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1064) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 24]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1120] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1120) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1176] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1176) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 72]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1232] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1232) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1288] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1288) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 120]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1344] = kernel_1[blockIdx_x * 73728 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 64512]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1400] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1400) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1456] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1456) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 48]
+ with T.launch_thread(threadIdx_x_2, 56):
+ if T.likely(threadIdx_x_2 < 24):
+ kernel_shared_1[threadIdx_x_2 + 1512] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1512) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 216]
+ for rc_outer_inner in range(2):
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 1] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 1] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 2] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 2] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 3] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 3] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 4] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 4] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 5] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 5] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 6] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 6] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 63] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 63] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 64] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 64] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 65] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 65] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 66] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 66] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 67] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 67] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 68] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 68] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 69] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 69] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 126] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 126] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 127] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 127] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 128] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 128] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 129] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 129] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 130] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 130] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 131] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 131] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 132] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 132] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 189] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 189] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 190] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 190] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 191] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 191] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 192] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 192] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 193] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 193] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 194] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 194] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 195] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 195] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 252] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 252] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 253] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 253] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 254] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 254] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 255] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 255] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 256] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 256] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 257] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 257] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 258] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 258] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 315] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 315] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 316] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 316] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 317] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 317] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 318] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 318] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 319] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 319] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 320] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 320] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 321] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 321] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 378] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 378] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 379] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 379] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 380] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 380] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 381] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 381] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 382] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 382] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 383] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 383] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 384] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 384] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 441] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 441] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 442] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 442] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 443] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 443] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 444] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 444] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 445] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 445] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 446] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 446] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 447] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 447] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 504] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 504] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 505] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 505] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 506] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 506] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 507] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 507] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 508] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 508] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 509] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 509] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 510] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 510] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 567] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 567] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 568] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 568] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 569] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 569] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 570] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 570] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 571] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 571] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 572] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 572] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 573] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 573] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 630] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 630] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 631] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 631] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 632] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 632] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 633] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 633] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 634] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 634] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 635] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 635] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 636] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 636] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 693] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 693] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 694] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 694] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 695] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 695] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 696] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 696] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 697] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 697] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 698] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 698] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 699] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 699] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 756] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 756] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 757] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 757] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 758] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 758] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 759] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 759] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 760] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 760] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 761] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 761] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 762] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 762] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 819] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 819] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 820] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 820] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 821] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 821] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 822] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 822] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 823] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 823] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 824] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 824] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 825] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 825] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 882] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 882] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 883] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 883] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 884] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 884] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 885] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 885] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 886] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 886] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 887] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 887] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 888] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 888] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 945] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 945] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 946] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 946] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 947] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 947] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 948] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 948] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 949] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 949] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 950] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 950] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 951] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 951] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 7] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 7] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 8] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 8] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 9] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 9] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 10] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 10] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 11] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 11] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 12] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 12] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 13] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 13] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 70] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 70] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 71] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 71] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 72] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 72] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 73] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 73] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 74] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 74] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 75] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 75] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 76] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 76] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 133] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 133] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 134] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 134] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 135] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 135] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 136] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 136] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 137] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 137] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 138] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 138] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 139] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 139] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 196] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 196] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 197] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 197] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 198] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 198] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 199] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 199] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 200] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 200] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 201] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 201] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 202] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 202] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 259] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 259] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 260] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 260] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 261] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 261] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 262] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 262] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 263] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 263] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 264] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 264] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 265] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 265] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 322] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 322] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 323] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 323] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 324] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 324] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 325] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 325] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 326] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 326] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 327] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 327] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 328] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 328] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 385] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 385] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 386] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 386] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 387] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 387] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 388] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 388] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 389] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 389] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 390] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 390] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 391] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 391] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 448] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 448] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 449] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 449] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 450] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 450] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 451] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 451] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 452] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 452] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 453] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 453] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 454] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 454] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 511] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 511] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 512] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 512] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 513] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 513] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 514] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 514] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 515] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 515] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 516] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 516] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 517] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 517] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 574] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 574] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 575] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 575] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 576] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 576] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 577] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 577] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 578] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 578] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 579] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 579] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 580] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 580] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 637] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 637] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 638] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 638] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 639] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 639] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 640] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 640] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 641] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 641] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 642] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 642] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 643] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 643] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 700] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 700] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 701] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 701] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 702] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 702] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 703] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 703] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 704] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 704] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 705] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 705] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 706] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 706] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 763] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 763] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 764] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 764] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 765] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 765] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 766] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 766] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 767] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 767] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 768] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 768] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 769] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 769] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 826] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 826] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 827] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 827] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 828] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 828] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 829] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 829] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 830] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 830] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 831] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 831] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 832] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 832] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 889] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 889] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 890] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 890] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 891] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 891] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 892] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 892] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 893] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 893] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 894] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 894] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 895] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 895] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 952] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 952] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 953] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 953] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 954] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 954] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 955] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 955] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 956] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 956] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 957] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 957] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 958] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 958] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 14] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 14] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 15] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 15] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 16] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 16] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 17] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 17] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 18] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 18] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 19] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 19] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 20] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 20] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 77] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 77] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 78] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 78] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 79] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 79] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 80] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 80] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 81] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 81] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 82] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 82] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 83] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 83] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 140] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 140] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 141] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 141] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 142] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 142] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 143] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 143] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 144] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 144] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 145] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 145] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 146] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 146] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 203] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 203] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 204] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 204] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 205] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 205] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 206] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 206] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 207] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 207] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 208] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 208] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 209] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 209] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 266] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 266] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 267] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 267] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 268] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 268] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 269] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 269] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 270] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 270] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 271] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 271] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 272] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 272] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 329] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 329] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 330] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 330] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 331] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 331] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 332] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 332] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 333] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 333] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 334] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 334] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 335] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 335] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 392] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 392] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 393] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 393] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 394] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 394] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 395] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 395] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 396] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 396] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 397] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 397] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 398] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 398] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 455] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 455] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 456] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 456] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 457] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 457] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 458] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 458] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 459] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 459] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 460] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 460] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 461] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 461] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 518] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 518] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 519] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 519] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 520] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 520] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 521] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 521] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 522] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 522] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 523] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 523] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 524] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 524] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 581] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 581] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 582] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 582] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 583] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 583] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 584] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 584] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 585] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 585] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 586] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 586] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 587] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 587] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 644] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 644] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 645] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 645] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 646] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 646] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 647] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 647] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 648] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 648] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 649] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 649] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 650] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 650] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 707] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 707] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 708] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 708] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 709] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 709] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 710] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 710] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 711] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 711] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 712] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 712] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 713] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 713] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 770] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 770] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 771] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 771] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 772] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 772] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 773] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 773] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 774] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 774] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 775] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 775] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 776] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 776] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 833] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 833] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 834] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 834] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 835] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 835] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 836] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 836] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 837] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 837] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 838] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 838] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 839] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 839] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 896] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 896] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 897] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 897] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 898] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 898] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 899] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 899] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 900] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 900] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 901] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 901] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 902] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 902] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 959] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 959] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 960] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 960] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 961] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 961] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 962] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 962] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 963] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 963] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 964] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 964] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 965] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 965] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ for i3_inner in range(7):
compute_1 = T.Buffer((25088,), data=compute.data)
bias_1 = T.Buffer((512,), data=bias.data)
- compute_1[blockIdx_x * 784 + threadIdx_x // 7 * 98 + i1_inner * 49 + i2_inner * 7 + threadIdx_x % 7] = T.max(conv2d_nchw_1[i1_inner * 7 + i2_inner] + bias_1[blockIdx_x * 16 + threadIdx_x // 7 * 2 + i1_inner], T.float32(0))
+ compute_1[blockIdx_x * 784 + threadIdx_x * 7 + i3_inner] = T.max(conv2d_nchw_1[i3_inner] + bias_1[blockIdx_x * 16 + threadIdx_x // 7], T.float32(0))
+ compute_1[blockIdx_x * 784 + threadIdx_x * 7 + i3_inner + 392] = T.max(conv2d_nchw_1[i3_inner + 7] + bias_1[blockIdx_x * 16 + threadIdx_x // 7 + 8], T.float32(0))
@@ -471,7 +1128,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 0.275 ms
+ Execution time of this operator: 0.325 ms
@@ -519,36 +1176,36 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
conv2d_nchw_nn_o_o_i, conv2d_nchw_nn_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_i, factor=1)
conv2d_nchw_nn_o_o_o_i, conv2d_nchw_nn_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_i, factor=1)
conv2d_nchw_nn_o_o_o_o, conv2d_nchw_nn_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_o_i, factor=1)
- conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=2)
+ conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=1)
conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=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=8)
- 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=7)
+ conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=2)
+ conv2d_nchw_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=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=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_o_o_i, conv2d_nchw_xx_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_i, factor=7)
+ conv2d_nchw_xx_o_o_o_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=4)
- conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=8)
+ conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=16)
+ conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=2)
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_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=1)
conv2d_nchw_rx_o_o, conv2d_nchw_rx_o_i = s[conv2d_nchw].split(conv2d_nchw_rx_o_i, factor=1)
s[conv2d_nchw].reorder(conv2d_nchw_nn_o_o_o_o, conv2d_nchw_ff_o_o_o_o, conv2d_nchw_yy_o_o_o_o, conv2d_nchw_xx_o_o_o_o, conv2d_nchw_nn_o_o_o_i, conv2d_nchw_ff_o_o_o_i, conv2d_nchw_yy_o_o_o_i, conv2d_nchw_xx_o_o_o_i, conv2d_nchw_nn_o_o_i, conv2d_nchw_ff_o_o_i, conv2d_nchw_yy_o_o_i, conv2d_nchw_xx_o_o_i, conv2d_nchw_rc_o_o, conv2d_nchw_ry_o_o, conv2d_nchw_rx_o_o, conv2d_nchw_rc_o_i, conv2d_nchw_ry_o_i, conv2d_nchw_rx_o_i, conv2d_nchw_nn_o_i, conv2d_nchw_ff_o_i, conv2d_nchw_yy_o_i, conv2 [...]
compute_i0_o_i, compute_i0_i = s[compute].split(compute_i0, factor=1)
compute_i0_o_o_i, compute_i0_o_i = s[compute].split(compute_i0_o_i, factor=1)
compute_i0_o_o_o, compute_i0_o_o_i = s[compute].split(compute_i0_o_o_i, factor=1)
- compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=2)
+ 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=8)
- 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=7)
- compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=1)
+ compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=2)
+ compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
+ compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=7)
compute_i2_o_o_o, compute_i2_o_o_i = s[compute].split(compute_i2_o_o_i, factor=1)
- compute_i3_o_i, compute_i3_i = s[compute].split(compute_i3, factor=1)
- compute_i3_o_o_i, compute_i3_o_i = s[compute].split(compute_i3_o_i, factor=7)
+ compute_i3_o_i, compute_i3_i = s[compute].split(compute_i3, factor=7)
+ compute_i3_o_o_i, compute_i3_o_i = s[compute].split(compute_i3_o_i, factor=1)
compute_i3_o_o_o, compute_i3_o_o_i = s[compute].split(compute_i3_o_o_i, factor=1)
s[compute].reorder(compute_i0_o_o_o, compute_i1_o_o_o, compute_i2_o_o_o, compute_i3_o_o_o, compute_i0_o_o_i, compute_i1_o_o_i, compute_i2_o_o_i, compute_i3_o_o_i, compute_i0_o_i, compute_i1_o_i, compute_i2_o_i, compute_i3_o_i, compute_i0_i, compute_i1_i, compute_i2_i, compute_i3_i)
s[conv2d_nchw].compute_at(s[compute], compute_i3_o_i)
@@ -566,7 +1223,7 @@ 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=4)
+ 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)
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=56)
s[kernel_shared].bind(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
@@ -575,7 +1232,7 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
s[pad_temp_shared].vectorize(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=56)
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", 64)
+ s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 1024)
s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
@@ -595,141 +1252,771 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
#endif
extern "C" __global__ void __launch_bounds__(56) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
float conv2d_nchw[14];
- __shared__ float pad_temp_shared[2592];
- __shared__ float kernel_shared[4608];
+ __shared__ float pad_temp_shared[2016];
+ __shared__ float kernel_shared[1536];
conv2d_nchw[0] = 0.000000e+00f;
- conv2d_nchw[1] = 0.000000e+00f;
- conv2d_nchw[2] = 0.000000e+00f;
- conv2d_nchw[3] = 0.000000e+00f;
- conv2d_nchw[4] = 0.000000e+00f;
- conv2d_nchw[5] = 0.000000e+00f;
- conv2d_nchw[6] = 0.000000e+00f;
conv2d_nchw[7] = 0.000000e+00f;
+ conv2d_nchw[1] = 0.000000e+00f;
conv2d_nchw[8] = 0.000000e+00f;
+ conv2d_nchw[2] = 0.000000e+00f;
conv2d_nchw[9] = 0.000000e+00f;
+ conv2d_nchw[3] = 0.000000e+00f;
conv2d_nchw[10] = 0.000000e+00f;
+ conv2d_nchw[4] = 0.000000e+00f;
conv2d_nchw[11] = 0.000000e+00f;
+ conv2d_nchw[5] = 0.000000e+00f;
conv2d_nchw[12] = 0.000000e+00f;
+ conv2d_nchw[6] = 0.000000e+00f;
conv2d_nchw[13] = 0.000000e+00f;
for (int rc_outer_outer = 0; rc_outer_outer < 16; ++rc_outer_outer) {
- __syncthreads();
- pad_temp_shared[((int)threadIdx.x)] = ((((9 <= ((int)threadIdx.x)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((rc_outer_outer * 1568) + ((((int)threadIdx.x) / 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 56)] = (((((9 <= ((((int)threadIdx.x) + 56) % 81)) && (((((int)threadIdx.x) + 56) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 56) / 81) * 49)) + ((((((int)threadIdx.x) + 56) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 112)] = (((((9 <= ((((int)threadIdx.x) + 31) % 81)) && (((((int)threadIdx.x) + 31) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 112) / 81) * 49)) + ((((((int)threadIdx.x) + 31) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 168)] = ((((3 <= ((int)threadIdx.x)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 168) / 81) * 49)) + (((((int)threadIdx.x) + 6) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 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 * 1568) + (((((int)threadIdx.x) + 224) / 81) * 49)) + ((((((int)threadIdx.x) + 62) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 280)] = (((((9 <= ((((int)threadIdx.x) + 37) % 81)) && (((((int)threadIdx.x) + 37) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 280) / 81) * 49)) + ((((((int)threadIdx.x) + 37) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 336)] = (((1 <= ((((int)threadIdx.x) + 3) % 9)) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 336) / 81) * 49)) + (((((int)threadIdx.x) + 12) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 392)] = (((((9 <= ((((int)threadIdx.x) + 68) % 81)) && (((((int)threadIdx.x) + 68) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 392) / 81) * 49)) + ((((((int)threadIdx.x) + 68) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- 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 * 1568) + (((((int)threadIdx.x) + 448) / 81) * 49)) + ((((((int)threadIdx.x) + 43) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 504)] = ((((((int)threadIdx.x) < 54) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 504) / 81) * 49)) + ((((int)threadIdx.x) / 9) * 7)) + (((int)threadIdx.x) % 9)) + 6)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 560)] = (((((9 <= ((((int)threadIdx.x) + 74) % 81)) && (((((int)threadIdx.x) + 74) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 560) / 81) * 49)) + ((((((int)threadIdx.x) + 74) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 616)] = (((((9 <= ((((int)threadIdx.x) + 49) % 81)) && (((((int)threadIdx.x) + 49) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 616) / 81) * 49)) + ((((((int)threadIdx.x) + 49) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 672)] = ((((((int)threadIdx.x) < 48) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 672) / 81) * 49)) + (((((int)threadIdx.x) + 24) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 728)] = (((((9 <= ((((int)threadIdx.x) + 80) % 81)) && (((((int)threadIdx.x) + 80) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 728) / 81) * 49)) + ((((((int)threadIdx.x) + 80) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 784)] = (((((9 <= ((((int)threadIdx.x) + 55) % 81)) && (((((int)threadIdx.x) + 55) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 784) / 81) * 49)) + ((((((int)threadIdx.x) + 55) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 840)] = (((((9 <= ((((int)threadIdx.x) + 30) % 81)) && (((((int)threadIdx.x) + 30) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 840) / 81) * 49)) + ((((((int)threadIdx.x) + 30) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 896)] = ((((4 <= ((int)threadIdx.x)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 896) / 81) * 49)) + (((((int)threadIdx.x) + 5) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 952)] = (((((9 <= ((((int)threadIdx.x) + 61) % 81)) && (((((int)threadIdx.x) + 61) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 952) / 81) * 49)) + ((((((int)threadIdx.x) + 61) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1008)] = (((((1 <= (((((int)threadIdx.x) / 9) + 4) % 9)) && (((((int)threadIdx.x) + 36) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1008) / 81) * 49)) + ((((((int)threadIdx.x) / 9) + 4) % 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1064)] = (((1 <= ((((int)threadIdx.x) + 2) % 9)) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1064) / 81) * 49)) + (((((int)threadIdx.x) + 11) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1120)] = (((((9 <= ((((int)threadIdx.x) + 67) % 81)) && (((((int)threadIdx.x) + 67) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1120) / 81) * 49)) + ((((((int)threadIdx.x) + 67) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1176)] = (((((9 <= ((((int)threadIdx.x) + 42) % 81)) && (((((int)threadIdx.x) + 42) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1176) / 81) * 49)) + ((((((int)threadIdx.x) + 42) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1232)] = ((((((int)threadIdx.x) < 55) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1232) / 81) * 49)) + (((((int)threadIdx.x) + 17) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1288)] = (((((9 <= ((((int)threadIdx.x) + 73) % 81)) && (((((int)threadIdx.x) + 73) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1288) / 81) * 49)) + ((((((int)threadIdx.x) + 73) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1344)] = (((((9 <= ((((int)threadIdx.x) + 48) % 81)) && (((((int)threadIdx.x) + 48) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1344) / 81) * 49)) + ((((((int)threadIdx.x) + 48) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1400)] = ((((((int)threadIdx.x) < 49) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1400) / 81) * 49)) + (((((int)threadIdx.x) + 23) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1456)] = (((((9 <= ((((int)threadIdx.x) + 79) % 81)) && (((((int)threadIdx.x) + 79) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1456) / 81) * 49)) + ((((((int)threadIdx.x) + 79) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1512)] = (((((1 <= (((((int)threadIdx.x) / 9) + 6) % 9)) && (((((int)threadIdx.x) + 54) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1512) / 81) * 49)) + ((((((int)threadIdx.x) / 9) + 6) % 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1568)] = (((((9 <= ((((int)threadIdx.x) + 29) % 81)) && (((((int)threadIdx.x) + 29) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1568) / 81) * 49)) + ((((((int)threadIdx.x) + 29) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1624)] = ((((5 <= ((int)threadIdx.x)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1624) / 81) * 49)) + (((((int)threadIdx.x) + 4) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1680)] = (((((9 <= ((((int)threadIdx.x) + 60) % 81)) && (((((int)threadIdx.x) + 60) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1680) / 81) * 49)) + ((((((int)threadIdx.x) + 60) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1736)] = (((((9 <= ((((int)threadIdx.x) + 35) % 81)) && (((((int)threadIdx.x) + 35) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1736) / 81) * 49)) + ((((((int)threadIdx.x) + 35) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1792)] = (((1 <= ((((int)threadIdx.x) + 1) % 9)) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1792) / 81) * 49)) + (((((int)threadIdx.x) + 10) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1848)] = (((((9 <= ((((int)threadIdx.x) + 66) % 81)) && (((((int)threadIdx.x) + 66) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1848) / 81) * 49)) + ((((((int)threadIdx.x) + 66) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1904)] = (((((9 <= ((((int)threadIdx.x) + 41) % 81)) && (((((int)threadIdx.x) + 41) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1904) / 81) * 49)) + ((((((int)threadIdx.x) + 41) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1960)] = (((1 <= ((((int)threadIdx.x) + 7) % 9)) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1960) / 81) * 49)) + (((((int)threadIdx.x) + 16) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2016)] = (((((1 <= (((((int)threadIdx.x) / 9) + 8) % 9)) && (((((int)threadIdx.x) + 72) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2016) / 81) * 49)) + ((((((int)threadIdx.x) / 9) + 8) % 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2072)] = (((((9 <= ((((int)threadIdx.x) + 47) % 81)) && (((((int)threadIdx.x) + 47) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2072) / 81) * 49)) + ((((((int)threadIdx.x) + 47) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2128)] = ((((((int)threadIdx.x) < 50) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2128) / 81) * 49)) + (((((int)threadIdx.x) + 22) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2184)] = (((((9 <= ((((int)threadIdx.x) + 78) % 81)) && (((((int)threadIdx.x) + 78) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2184) / 81) * 49)) + ((((((int)threadIdx.x) + 78) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2240)] = (((((9 <= ((((int)threadIdx.x) + 53) % 81)) && (((((int)threadIdx.x) + 53) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2240) / 81) * 49)) + ((((((int)threadIdx.x) + 53) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2296)] = (((((9 <= ((((int)threadIdx.x) + 28) % 81)) && (((((int)threadIdx.x) + 28) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2296) / 81) * 49)) + ((((((int)threadIdx.x) + 28) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2352)] = ((((6 <= ((int)threadIdx.x)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2352) / 81) * 49)) + (((((int)threadIdx.x) + 3) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2408)] = (((((9 <= ((((int)threadIdx.x) + 59) % 81)) && (((((int)threadIdx.x) + 59) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2408) / 81) * 49)) + ((((((int)threadIdx.x) + 59) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2464)] = (((((9 <= ((((int)threadIdx.x) + 34) % 81)) && (((((int)threadIdx.x) + 34) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2464) / 81) * 49)) + ((((((int)threadIdx.x) + 34) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2520)] = (((1 <= (((int)threadIdx.x) % 9)) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2520) / 81) * 49)) + ((((int)threadIdx.x) / 9) * 7)) + (((int)threadIdx.x) % 9)) - 1)] : 0.000000e+00f);
- if (((int)threadIdx.x) < 16) {
- pad_temp_shared[(((int)threadIdx.x) + 2576)] = ((((((int)threadIdx.x) < 7) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2576) / 81) * 49)) + (((((int)threadIdx.x) + 65) / 9) * 7)) + ((int)threadIdx.x)) - 6)] : 0.000000e+00f);
- }
- for (int ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer = 0; ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer < 21; ++ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer) {
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) < 144) {
- kernel_shared[((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4))] = kernel[((((((int)blockIdx.x) * 73728) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) / 9) * 4608)) + (rc_outer_outer * 288)) + (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) % 288))];
+ for (int rx_outer_outer = 0; rx_outer_outer < 3; ++rx_outer_outer) {
+ __syncthreads();
+ pad_temp_shared[((int)threadIdx.x)] = ((((7 <= ((int)threadIdx.x)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((rc_outer_outer * 1568) + ((int)threadIdx.x)) + rx_outer_outer) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 56)] = (((((1 <= (((((int)threadIdx.x) / 7) + 8) % 9)) && ((((((int)threadIdx.x) / 7) + 8) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 56) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 8) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 112)] = (((((1 <= (((((int)threadIdx.x) / 7) + 7) % 9)) && ((((((int)threadIdx.x) / 7) + 7) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 112) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 7) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 168)] = (((((1 <= (((((int)threadIdx.x) / 7) + 6) % 9)) && ((((((int)threadIdx.x) / 7) + 6) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 168) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 6) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((1 <= (((((int)threadIdx.x) / 7) + 5) % 9)) && ((((((int)threadIdx.x) / 7) + 5) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 224) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 5) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 280)] = (((((1 <= (((((int)threadIdx.x) / 7) + 4) % 9)) && ((((((int)threadIdx.x) / 7) + 4) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 280) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 4) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 336)] = (((((1 <= (((((int)threadIdx.x) / 7) + 3) % 9)) && ((((((int)threadIdx.x) / 7) + 3) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 336) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 3) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 392)] = (((((1 <= (((((int)threadIdx.x) / 7) + 2) % 9)) && ((((((int)threadIdx.x) / 7) + 2) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 392) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 2) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 448)] = ((((((int)threadIdx.x) < 49) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 448) / 63) * 49)) + ((int)threadIdx.x)) + rx_outer_outer) - 1)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 504)] = ((((7 <= ((int)threadIdx.x)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((rc_outer_outer * 1568) + ((int)threadIdx.x)) + rx_outer_outer) + 384)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 560)] = (((((1 <= (((((int)threadIdx.x) / 7) + 8) % 9)) && ((((((int)threadIdx.x) / 7) + 8) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 560) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 8) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 616)] = (((((1 <= (((((int)threadIdx.x) / 7) + 7) % 9)) && ((((((int)threadIdx.x) / 7) + 7) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 616) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 7) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 672)] = (((((1 <= (((((int)threadIdx.x) / 7) + 6) % 9)) && ((((((int)threadIdx.x) / 7) + 6) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 672) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 6) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 728)] = (((((1 <= (((((int)threadIdx.x) / 7) + 5) % 9)) && ((((((int)threadIdx.x) / 7) + 5) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 728) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 5) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = (((((1 <= (((((int)threadIdx.x) / 7) + 4) % 9)) && ((((((int)threadIdx.x) / 7) + 4) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 784) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 4) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 840)] = (((((1 <= (((((int)threadIdx.x) / 7) + 3) % 9)) && ((((((int)threadIdx.x) / 7) + 3) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 840) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 3) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 896)] = (((((1 <= (((((int)threadIdx.x) / 7) + 2) % 9)) && ((((((int)threadIdx.x) / 7) + 2) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 896) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 2) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 952)] = ((((((int)threadIdx.x) < 49) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 952) / 63) * 49)) + ((int)threadIdx.x)) + rx_outer_outer) - 1)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1008)] = ((((7 <= ((int)threadIdx.x)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((rc_outer_outer * 1568) + ((int)threadIdx.x)) + rx_outer_outer) + 776)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1064)] = (((((1 <= (((((int)threadIdx.x) / 7) + 8) % 9)) && ((((((int)threadIdx.x) / 7) + 8) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1064) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 8) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1120)] = (((((1 <= (((((int)threadIdx.x) / 7) + 7) % 9)) && ((((((int)threadIdx.x) / 7) + 7) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1120) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 7) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1176)] = (((((1 <= (((((int)threadIdx.x) / 7) + 6) % 9)) && ((((((int)threadIdx.x) / 7) + 6) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1176) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 6) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1232)] = (((((1 <= (((((int)threadIdx.x) / 7) + 5) % 9)) && ((((((int)threadIdx.x) / 7) + 5) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1232) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 5) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1288)] = (((((1 <= (((((int)threadIdx.x) / 7) + 4) % 9)) && ((((((int)threadIdx.x) / 7) + 4) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1288) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 4) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1344)] = (((((1 <= (((((int)threadIdx.x) / 7) + 3) % 9)) && ((((((int)threadIdx.x) / 7) + 3) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1344) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 3) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1400)] = (((((1 <= (((((int)threadIdx.x) / 7) + 2) % 9)) && ((((((int)threadIdx.x) / 7) + 2) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1400) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 2) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1456)] = ((((((int)threadIdx.x) < 49) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1456) / 63) * 49)) + ((int)threadIdx.x)) + rx_outer_outer) - 1)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1512)] = ((((7 <= ((int)threadIdx.x)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((rc_outer_outer * 1568) + ((int)threadIdx.x)) + rx_outer_outer) + 1168)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1568)] = (((((1 <= (((((int)threadIdx.x) / 7) + 8) % 9)) && ((((((int)threadIdx.x) / 7) + 8) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1568) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 8) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1624)] = (((((1 <= (((((int)threadIdx.x) / 7) + 7) % 9)) && ((((((int)threadIdx.x) / 7) + 7) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1624) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 7) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1680)] = (((((1 <= (((((int)threadIdx.x) / 7) + 6) % 9)) && ((((((int)threadIdx.x) / 7) + 6) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1680) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 6) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1736)] = (((((1 <= (((((int)threadIdx.x) / 7) + 5) % 9)) && ((((((int)threadIdx.x) / 7) + 5) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1736) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 5) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1792)] = (((((1 <= (((((int)threadIdx.x) / 7) + 4) % 9)) && ((((((int)threadIdx.x) / 7) + 4) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1792) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 4) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1848)] = (((((1 <= (((((int)threadIdx.x) / 7) + 3) % 9)) && ((((((int)threadIdx.x) / 7) + 3) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1848) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 3) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1904)] = (((((1 <= (((((int)threadIdx.x) / 7) + 2) % 9)) && ((((((int)threadIdx.x) / 7) + 2) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1904) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 2) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1960)] = ((((((int)threadIdx.x) < 49) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1960) / 63) * 49)) + ((int)threadIdx.x)) + rx_outer_outer) - 1)] : 0.000000e+00f);
+ kernel_shared[((int)threadIdx.x)] = kernel[((((((int)blockIdx.x) * 73728) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 56)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 56) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 56) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 112)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 112) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 48)];
+ kernel_shared[(((int)threadIdx.x) + 168)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 168) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 72) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 224)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 224) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 96)];
+ kernel_shared[(((int)threadIdx.x) + 280)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 280) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 88) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 336)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 336) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 48) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 392)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 392) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 24)];
+ kernel_shared[(((int)threadIdx.x) + 448)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 448) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 64) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 504)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 504) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 72)];
+ kernel_shared[(((int)threadIdx.x) + 560)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 560) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 80) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 616)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 616) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 120)];
+ kernel_shared[(((int)threadIdx.x) + 672)] = kernel[(((((((int)blockIdx.x) * 73728) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 32256)];
+ kernel_shared[(((int)threadIdx.x) + 728)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 728) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 56) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 784)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 784) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 48)];
+ kernel_shared[(((int)threadIdx.x) + 840)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 840) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 72) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 896)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 896) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 96)];
+ kernel_shared[(((int)threadIdx.x) + 952)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 952) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 88) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1008)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1008) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 48) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1064)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1064) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 24)];
+ kernel_shared[(((int)threadIdx.x) + 1120)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1120) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 64) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1176)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1176) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 72)];
+ kernel_shared[(((int)threadIdx.x) + 1232)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1232) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 80) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1288)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1288) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 120)];
+ kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[(((((((int)blockIdx.x) * 73728) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 64512)];
+ kernel_shared[(((int)threadIdx.x) + 1400)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1400) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 56) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1456)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1456) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 48)];
+ if (((int)threadIdx.x) < 24) {
+ kernel_shared[(((int)threadIdx.x) + 1512)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1512) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 216)];
}
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) < 144) {
- kernel_shared[(((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 1)] = kernel[((((((int)blockIdx.x) * 73728) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) / 9) * 4608)) + (rc_outer_outer * 288)) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 1) % 288))];
- }
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) < 144) {
- kernel_shared[(((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 2)] = kernel[((((((int)blockIdx.x) * 73728) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) / 9) * 4608)) + (rc_outer_outer * 288)) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 2) % 288))];
- }
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) < 144) {
- kernel_shared[(((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 3)] = kernel[((((((int)blockIdx.x) * 73728) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) / 9) * 4608)) + (rc_outer_outer * 288)) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 3) % 288))];
- }
- }
- __syncthreads();
- for (int rc_outer_inner = 0; rc_outer_inner < 8; ++rc_outer_inner) {
- for (int ry_outer_inner = 0; ry_outer_inner < 3; ++ry_outer_inner) {
- for (int rc_inner = 0; rc_inner < 4; ++rc_inner) {
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 9)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 18)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 27)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 36)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 45)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 54)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7))] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 9)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 18)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 27)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 36)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 45)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 54)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 1)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 10)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 19)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 28)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 37)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 46)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 55)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 1)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 10)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 19)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 28)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 37)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 46)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 55)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 2)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 11)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 20)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 29)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 38)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 47)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 56)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 2)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 11)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 20)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 29)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 38)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 47)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 56)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- }
+ __syncthreads();
+ for (int rc_outer_inner = 0; rc_outer_inner < 2; ++rc_outer_inner) {
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7))] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7))] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 1)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 2)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 3)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 4)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 5)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 6)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 63)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 63)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 64)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 64)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 65)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 65)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 66)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 66)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 67)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 67)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 68)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 68)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 69)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 69)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 126)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 126)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 127)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 127)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 128)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 128)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 129)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 129)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 130)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 130)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 131)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 131)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 132)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 132)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 189)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 189)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 190)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 190)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 191)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 191)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 192)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 192)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 193)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 193)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 194)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 194)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 195)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 195)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 252)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 252)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 253)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 253)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 254)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 254)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 255)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 255)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 256)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 256)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 257)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 257)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 258)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 258)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 315)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 315)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 316)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 316)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 317)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 317)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 318)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 318)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 319)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 319)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 320)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 320)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 321)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 321)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 378)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 378)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 379)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 379)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 380)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 380)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 381)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 381)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 382)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 382)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 383)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 383)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 384)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 384)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 441)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 441)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 442)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 442)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 443)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 443)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 444)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 444)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 445)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 445)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 446)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 446)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 447)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 447)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 504)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 504)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 505)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 505)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 506)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 506)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 507)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 507)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 508)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 508)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 509)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 509)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 510)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 510)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 567)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 567)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 568)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 568)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 569)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 569)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 570)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 570)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 571)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 571)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 572)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 572)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 573)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 573)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 630)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 630)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 631)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 631)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 632)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 632)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 633)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 633)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 634)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 634)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 635)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 635)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 636)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 636)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 693)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 693)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 694)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 694)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 695)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 695)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 696)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 696)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 697)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 697)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 698)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 698)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 699)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 699)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 756)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 756)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 757)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 757)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 758)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 758)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 759)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 759)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 760)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 760)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 761)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 761)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 762)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 762)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 819)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 819)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 820)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 820)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 821)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 821)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 822)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 822)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 823)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 823)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 824)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 824)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 825)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 825)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 882)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 882)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 883)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 883)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 884)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 884)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 885)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 885)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 886)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 886)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 887)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 887)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 888)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 888)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 945)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 945)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 946)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 946)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 947)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 947)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 948)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 948)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 949)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 949)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 950)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 950)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 951)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 951)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 7)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 7)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 8)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 8)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 9)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 9)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 10)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 10)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 11)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 11)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 12)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 12)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 13)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 13)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 70)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 70)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 71)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 71)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 72)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 72)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 73)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 73)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 74)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 74)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 75)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 75)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 76)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 76)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 133)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 133)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 134)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 134)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 135)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 135)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 136)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 136)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 137)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 137)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 138)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 138)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 139)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 139)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 196)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 196)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 197)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 197)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 198)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 198)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 199)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 199)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 200)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 200)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 201)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 201)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 202)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 202)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 259)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 259)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 260)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 260)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 261)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 261)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 262)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 262)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 263)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 263)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 264)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 264)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 265)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 265)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 322)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 322)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 323)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 323)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 324)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 324)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 325)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 325)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 326)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 326)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 327)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 327)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 328)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 328)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 385)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 385)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 386)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 386)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 387)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 387)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 388)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 388)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 389)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 389)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 390)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 390)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 391)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 391)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 448)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 448)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 449)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 449)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 450)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 450)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 451)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 451)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 452)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 452)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 453)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 453)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 454)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 454)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 511)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 511)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 512)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 512)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 513)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 513)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 514)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 514)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 515)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 515)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 516)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 516)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 517)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 517)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 574)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 574)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 575)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 575)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 576)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 576)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 577)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 577)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 578)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 578)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 579)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 579)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 580)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 580)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 637)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 637)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 638)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 638)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 639)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 639)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 640)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 640)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 641)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 641)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 642)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 642)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 643)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 643)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 700)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 700)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 701)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 701)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 702)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 702)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 703)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 703)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 704)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 704)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 705)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 705)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 706)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 706)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 763)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 763)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 764)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 764)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 765)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 765)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 766)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 766)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 767)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 767)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 768)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 768)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 769)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 769)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 826)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 826)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 827)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 827)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 828)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 828)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 829)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 829)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 830)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 830)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 831)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 831)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 832)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 832)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 889)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 889)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 890)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 890)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 891)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 891)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 892)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 892)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 893)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 893)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 894)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 894)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 895)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 895)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 952)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 952)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 953)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 953)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 954)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 954)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 955)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 955)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 956)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 956)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 957)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 957)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 958)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 958)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 14)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 14)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 15)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 15)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 16)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 16)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 17)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 17)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 18)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 18)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 19)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 19)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 20)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 20)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 77)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 77)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 78)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 78)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 79)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 79)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 80)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 80)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 81)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 81)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 82)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 82)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 83)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 83)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 140)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 140)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 141)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 141)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 142)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 142)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 143)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 143)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 144)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 144)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 145)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 145)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 146)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 146)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 203)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 203)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 204)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 204)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 205)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 205)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 206)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 206)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 207)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 207)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 208)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 208)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 209)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 209)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 266)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 266)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 267)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 267)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 268)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 268)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 269)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 269)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 270)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 270)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 271)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 271)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 272)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 272)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 329)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 329)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 330)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 330)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 331)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 331)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 332)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 332)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 333)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 333)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 334)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 334)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 335)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 335)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 392)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 392)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 393)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 393)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 394)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 394)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 395)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 395)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 396)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 396)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 397)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 397)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 398)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 398)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 455)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 455)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 456)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 456)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 457)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 457)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 458)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 458)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 459)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 459)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 460)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 460)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 461)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 461)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 518)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 518)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 519)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 519)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 520)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 520)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 521)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 521)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 522)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 522)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 523)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 523)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 524)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 524)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 581)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 581)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 582)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 582)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 583)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 583)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 584)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 584)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 585)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 585)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 586)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 586)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 587)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 587)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 644)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 644)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 645)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 645)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 646)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 646)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 647)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 647)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 648)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 648)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 649)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 649)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 650)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 650)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 707)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 707)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 708)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 708)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 709)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 709)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 710)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 710)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 711)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 711)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 712)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 712)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 713)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 713)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 770)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 770)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 771)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 771)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 772)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 772)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 773)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 773)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 774)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 774)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 775)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 775)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 776)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 776)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 833)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 833)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 834)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 834)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 835)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 835)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 836)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 836)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 837)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 837)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 838)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 838)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 839)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 839)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 896)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 896)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 897)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 897)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 898)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 898)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 899)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 899)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 900)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 900)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 901)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 901)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 902)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 902)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 959)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 959)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 960)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 960)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 961)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 961)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 962)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 962)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 963)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 963)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 964)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 964)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 965)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 965)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
}
}
}
- for (int i1_inner = 0; i1_inner < 2; ++i1_inner) {
- for (int i2_inner = 0; i2_inner < 7; ++i2_inner) {
- compute[(((((((int)blockIdx.x) * 784) + ((((int)threadIdx.x) / 7) * 98)) + (i1_inner * 49)) + (i2_inner * 7)) + (((int)threadIdx.x) % 7))] = max((conv2d_nchw[((i1_inner * 7) + i2_inner)] + bias[(((((int)blockIdx.x) * 16) + ((((int)threadIdx.x) / 7) * 2)) + i1_inner)]), 0.000000e+00f);
- }
+ for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
+ compute[(((((int)blockIdx.x) * 784) + (((int)threadIdx.x) * 7)) + i3_inner)] = max((conv2d_nchw[i3_inner] + bias[((((int)blockIdx.x) * 16) + (((int)threadIdx.x) / 7))]), 0.000000e+00f);
+ compute[((((((int)blockIdx.x) * 784) + (((int)threadIdx.x) * 7)) + i3_inner) + 392)] = max((conv2d_nchw[(i3_inner + 7)] + bias[(((((int)blockIdx.x) * 16) + (((int)threadIdx.x) / 7)) + 8)]), 0.000000e+00f);
}
}
@@ -789,7 +2076,7 @@ In the example below we resume the status and do more 5 trials.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 6 minutes 9.023 seconds)
+ **Total running time of the script:** ( 6 minutes 24.225 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 3a475e74ec..b18199fc2b 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)
- 7.8896 7.8893 7.8964 7.8830 0.0054
+ 7.9159 7.9132 7.9271 7.9074 0.0082
@@ -675,7 +675,7 @@ Other Tips
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 7.297 seconds)
+ **Total running time of the script:** ( 1 minutes 9.353 seconds)
.. _sphx_glr_download_how_to_tune_with_autoscheduler_tune_network_cuda.py:
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 90570541b9..2578fb7b99 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)
- 747.7804 748.3530 750.0463 744.9418 2.1229
+ 757.4250 755.7490 760.8145 755.7114 2.3968
@@ -694,7 +694,7 @@ Other Tips
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 41.602 seconds)
+ **Total running time of the script:** ( 1 minutes 44.830 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 54abf8575c..9f0265536d 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
@@ -392,9 +392,9 @@ layout transformation, parallelization, vectorization, unrolling, and operator f
for i0_outer_i1_outer_fused in T.parallel(32):
compute_1 = T.allocate([2048], "float32", "global")
compute_2 = T.Buffer((2048,), data=compute_1)
- for i_outer_inner, nb_j_inner in T.grid(2, 2):
- for i_inner_init in range(32):
- cse_var_1: T.int32 = i_outer_inner * 1024 + i_inner_init * 32 + nb_j_inner * 16
+ for nb_j_inner in range(2):
+ for i_inner_init in range(64):
+ cse_var_1: T.int32 = i_inner_init * 32 + nb_j_inner * 16
compute_2[cse_var_1] = T.float32(0)
compute_2[cse_var_1 + 1] = T.float32(0)
compute_2[cse_var_1 + 2] = T.float32(0)
@@ -411,47 +411,47 @@ layout transformation, parallelization, vectorization, unrolling, and operator f
compute_2[cse_var_1 + 13] = T.float32(0)
compute_2[cse_var_1 + 14] = T.float32(0)
compute_2[cse_var_1 + 15] = T.float32(0)
- for elem_idx, i_inner in T.grid(T.Let(placeholder_5[cse_var_2 + 1] - placeholder_5[cse_var_2], where={cse_var_2: i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner}), 32):
+ for elem_idx, i_inner in T.grid(T.Let(placeholder_5[cse_var_2 + 1] - placeholder_5[cse_var_2], where={cse_var_2: i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner}), 64):
cse_var_2 = T.int32()
placeholder_5 = T.Buffer((33,), "int32", data=placeholder_3.data)
cse_var_21: T.int32 = elem_idx * 16
- cse_var_20: T.int32 = i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner
- cse_var_19: T.int32 = i_outer_inner * 1024 + i_inner * 32 + nb_j_inner * 16
- cse_var_18: T.int32 = i0_outer_i1_outer_fused // 16 * 16384 + i_outer_inner * 8192 + i_inner * 256
- cse_var_17: T.int32 = cse_var_19 + 9
- cse_var_16: T.int32 = cse_var_19 + 8
- cse_var_15: T.int32 = cse_var_19 + 7
- cse_var_14: T.int32 = cse_var_19 + 6
- cse_var_13: T.int32 = cse_var_19 + 5
- cse_var_12: T.int32 = cse_var_19 + 4
- cse_var_11: T.int32 = cse_var_19 + 3
- cse_var_10: T.int32 = cse_var_19 + 2
- cse_var_9: T.int32 = cse_var_19 + 15
- cse_var_8: T.int32 = cse_var_19 + 14
- cse_var_7: T.int32 = cse_var_19 + 13
- cse_var_6: T.int32 = cse_var_19 + 12
- cse_var_5: T.int32 = cse_var_19 + 11
- cse_var_4: T.int32 = cse_var_19 + 10
- cse_var_3: T.int32 = cse_var_19 + 1
+ cse_var_20: T.int32 = i_inner * 32 + nb_j_inner * 16
+ cse_var_19: T.int32 = i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner
+ cse_var_18: T.int32 = i0_outer_i1_outer_fused // 16 * 16384 + i_inner * 256
+ cse_var_17: T.int32 = cse_var_20 + 9
+ cse_var_16: T.int32 = cse_var_20 + 8
+ cse_var_15: T.int32 = cse_var_20 + 7
+ cse_var_14: T.int32 = cse_var_20 + 6
+ cse_var_13: T.int32 = cse_var_20 + 5
+ cse_var_12: T.int32 = cse_var_20 + 4
+ cse_var_11: T.int32 = cse_var_20 + 3
+ cse_var_10: T.int32 = cse_var_20 + 2
+ cse_var_9: T.int32 = cse_var_20 + 15
+ cse_var_8: T.int32 = cse_var_20 + 14
+ cse_var_7: T.int32 = cse_var_20 + 13
+ cse_var_6: T.int32 = cse_var_20 + 12
+ cse_var_5: T.int32 = cse_var_20 + 11
+ cse_var_4: T.int32 = cse_var_20 + 10
+ cse_var_3: T.int32 = cse_var_20 + 1
placeholder_6 = T.Buffer((78656,), data=placeholder_1.data)
placeholder_7 = T.Buffer((32768,), data=placeholder.data)
placeholder_8 = T.Buffer((4916,), "int32", data=placeholder_2.data)
- compute_2[cse_var_19] = compute_2[cse_var_19] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_3] = compute_2[cse_var_3] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 1] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_10] = compute_2[cse_var_10] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 2] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_11] = compute_2[cse_var_11] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 3] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_12] = compute_2[cse_var_12] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 4] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_13] = compute_2[cse_var_13] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 5] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_14] = compute_2[cse_var_14] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 6] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_15] = compute_2[cse_var_15] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 7] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_16] = compute_2[cse_var_16] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 8] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_17] = compute_2[cse_var_17] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 9] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_4] = compute_2[cse_var_4] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 10] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_5] = compute_2[cse_var_5] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 11] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_6] = compute_2[cse_var_6] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 12] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_7] = compute_2[cse_var_7] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 13] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_8] = compute_2[cse_var_8] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 14] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_9] = compute_2[cse_var_9] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 15] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
+ compute_2[cse_var_20] = compute_2[cse_var_20] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_3] = compute_2[cse_var_3] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 1] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_10] = compute_2[cse_var_10] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 2] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_11] = compute_2[cse_var_11] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 3] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_12] = compute_2[cse_var_12] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 4] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_13] = compute_2[cse_var_13] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 5] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_14] = compute_2[cse_var_14] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 6] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_15] = compute_2[cse_var_15] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 7] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_16] = compute_2[cse_var_16] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 8] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_17] = compute_2[cse_var_17] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 9] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_4] = compute_2[cse_var_4] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 10] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_5] = compute_2[cse_var_5] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 11] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_6] = compute_2[cse_var_6] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 12] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_7] = compute_2[cse_var_7] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 13] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_8] = compute_2[cse_var_8] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 14] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_9] = compute_2[cse_var_9] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 15] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
for i0_inner in range(64):
cse_var_22: T.int32 = i0_outer_i1_outer_fused // 16 * 32768 + i0_inner * 512 + i0_outer_i1_outer_fused % 16 * 32
compute_3 = T.Buffer((65536,), data=compute.data)
@@ -506,7 +506,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 1.728 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 d4a2f2875b..e7faae0c99 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
@@ -5,10 +5,10 @@
Computation times
=================
-**00:33.815** total execution time for **how_to_tune_with_autotvm** files:
+**00:43.909** 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``) | 00:33.781 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_conv2d_cuda.py` (``tune_conv2d_cuda.py``) | 00:43.874 | 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 |
+--------------------------------------------------------------------------------------------------+-----------+--------+
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 465ccab66e..f38c1bdcd6 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
@@ -390,655 +390,8 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5021072
- No: 2 GFLOPS: 90.55/90.55 result: MeasureResult(costs=(0.002556609098360656,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.1473910808563232, timestamp=1679006697.4090955) [('tile_f', [-1, 1, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8254840
- No: 3 GFLOPS: 0.00/90.55 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 742, in __call__
- yield remote, remote.load_module(os.path.split(build_result.filename)[1])
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 706, in run_through_rpc
- costs = time_f(*args).results
- File "/workspace/python/tvm/runtime/module.py", line 357, in evaluator
- blob = feval(*args)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 262, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./packed_func.pxi", line 251, in tvm._ffi._cy3.core.FuncCall3
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
- tvm._ffi.base.TVMError: Traceback (most recent call last):
- 4: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 3: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 2: tvm::runtime::RPCWrappedFunc::operator()(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../src/runtime/rpc/rpc_module.cc:129
- 1: tvm::runtime::RPCClientSession::CallFunc(void*, TVMValue const*, int const*, int, std::function<void (tvm::runtime::TVMArgs)> const&)
- at ../src/runtime/rpc/rpc_endpoint.cc:1012
- 0: tvm::runtime::RPCEndpoint::CallFunc(void*, TVMValue const*, int const*, int, std::function<void (tvm::runtime::TVMArgs)>)
- at ../src/runtime/rpc/rpc_endpoint.cc:804
- File "../src/runtime/rpc/rpc_endpoint.cc", line 804
- TVMError:
- ---------------------------------------------------------------
- An error occurred during the execution of TVM.
- For more information, please see: https://tvm.apache.org/docs/errors.html
- ---------------------------------------------------------------
- Check failed: (code == RPCCode::kReturn) is false: code=kShutdown
-
- During handling of the above exception, another exception occurred:
-
- Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 706, in run_through_rpc
- costs = time_f(*args).results
- File "/usr/lib/python3.7/contextlib.py", line 130, in __exit__
- self.gen.throw(type, value, traceback)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 746, in __call__
- remote.remove(build_result.filename)
- File "/workspace/python/tvm/rpc/client.py", line 144, in remove
- self._remote_funcs["remove"] = self.get_function("tvm.rpc.server.remove")
- File "/workspace/python/tvm/rpc/client.py", line 72, in get_function
- return self._sess.get_function(name)
- File "/workspace/python/tvm/runtime/module.py", line 171, in get_function
- self.handle, c_str(name), ctypes.c_int(query_imports), ctypes.byref(ret_handle)
- File "/workspace/python/tvm/_ffi/base.py", line 348, in check_call
- raise get_last_ffi_error()
- tvm._ffi.base.TVMError: Traceback (most recent call last):
- 52: 0xffffffffffffffff
- 51: _start
- 50: __libc_start_main
- 49: _Py_UnixMain
- 48: 0x0000000000650da0
- 47: 0x0000000000650afa
- 46: _PyFunction_FastCallDict
- 45: _PyEval_EvalCodeWithName
- 44: _PyEval_EvalFrameDefault
- 43: _PyFunction_FastCallKeywords
- 42: _PyEval_EvalCodeWithName
- 41: _PyEval_EvalFrameDefault
- 40: _PyMethodDef_RawFastCallKeywords
- 39: 0x0000000000546369
- 38: _PyEval_EvalCodeWithName
- 37: _PyEval_EvalFrameDefault
- 36: _PyFunction_FastCallKeywords
- 35: _PyEval_EvalCodeWithName
- 34: _PyEval_EvalFrameDefault
- 33: _PyFunction_FastCallDict
- 32: _PyEval_EvalCodeWithName
- 31: _PyEval_EvalFrameDefault
- 30: _PyObject_FastCallDict
- 29: 0x00000000004c06e1
- 28: _PyFunction_FastCallDict
- 27: _PyEval_EvalFrameDefault
- 26: _PyMethodDescr_FastCallKeywords
- 25: 0x00000000005dcb58
- 24: 0x00000000005dc83f
- 23: 0x00000000004ba127
- 22: _PyEval_EvalFrameDefault
- 21: _PyFunction_FastCallKeywords
- 20: _PyEval_EvalFrameDefault
- 19: _PyFunction_FastCallKeywords
- 18: _PyEval_EvalFrameDefault
- 17: _PyFunction_FastCallKeywords
- 16: _PyEval_EvalCodeWithName
- 15: _PyEval_EvalFrameDefault
- 14: 0x0000000000537c30
- 13: _PyObject_FastCallKeywords
- 12: 0x00007f972211dfa2
- 11: _ctypes_callproc
- 10: ffi_call
- 9: ffi_call_unix64
- 8: TVMModGetFunction
- at ../src/runtime/c_runtime_api.cc:408
- 7: tvm::runtime::ModuleNode::GetFunction(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, bool)
- at ../src/runtime/module.cc:66
- 6: tvm::runtime::RPCModuleNode::GetFunction(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, tvm::runtime::ObjectPtr<tvm::runtime::Object> const&)
- at ../src/runtime/rpc/rpc_module.cc:185
- 5: tvm::runtime::RPCClientSession::GetFunction(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)
- at ../src/runtime/rpc/rpc_endpoint.cc:1007
- 4: tvm::runtime::TVMRetValue tvm::runtime::RPCEndpoint::SysCallRemote<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&>(tvm::runtime::RPCCode, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)
- at ../src/runtime/rpc/rpc_endpoint.h:223
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&>(int&&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/rpc/rpc_endpoint.cc:684
- File "../src/runtime/rpc/rpc_endpoint.cc", line 684
- TVMError:
- ---------------------------------------------------------------
- An error occurred during the execution of TVM.
- For more information, please see: https://tvm.apache.org/docs/errors.html
- ---------------------------------------------------------------
- Check failed: (code == RPCCode::kReturn) is false: code=1
-
- Traceback (most recent call last):
- 52: 0xffffffffffffffff
- 51: _start
- 50: __libc_start_main
- 49: _Py_UnixMain
- 48: 0x0000000000650da0
- 47: 0x0000000000650afa
- 46: _PyFunction_FastCallDict
- 45: _PyEval_EvalCodeWithName
- 44: _PyEval_EvalFrameDefault
- 43: _PyFunction_FastCallKeywords
- 42: _PyEval_EvalCodeWithName
- 41: _PyEval_EvalFrameDefault
- 40: _PyMethodDef_RawFastCallKeywords
- 39: 0x0000000000546369
- 38: _PyEval_EvalCodeWithName
- 37: _PyEval_EvalFrameDefault
- 36: _PyFunction_FastCallKeywords
- 35: _PyEval_EvalCodeWithName
- 34: _PyEval_EvalFrameDefault
- 33: _PyFunction_FastCallDict
- 32: _PyEval_EvalCodeWithName
- 31: _PyEval_EvalFrameDefault
- 30: _PyObject_FastCallDict
- 29: 0x00000000004c06e1
- 28: _PyFunction_FastCallDict
- 27: _PyEval_EvalFrameDefault
- 26: _PyMethodDescr_FastCallKeywords
- 25: 0x00000000005dcb58
- 24: 0x00000000005dc83f
- 23: 0x00000000004ba127
- 22: _PyEval_EvalFrameDefault
- 21: _PyFunction_FastCallKeywords
- 20: _PyEval_EvalFrameDefault
- 19: _PyFunction_FastCall [('tile_f', [-1, 64, 1, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1590266
- No: 4 GFLOPS: 0.00/90.55 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
- func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
- func = build(s, args, target=target, runtime=runtime)
- File "/workspace/python/tvm/driver/build_module.py", line 227, in build
- input_mod = lower(inputs, args, name=name, binds=binds)
- File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
- return ffi.lower_schedule(inp, args, name, binds, simple_mode)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
- tvm._ffi.base.TVMError: Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
- raise InstantiationError("Skipped because of invalid gpu kernel")
- tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
-
- Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 32]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4604973
- No: 5 GFLOPS: 0.00/90.55 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
- func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
- func = build(s, args, target=target, runtime=runtime)
- File "/workspace/python/tvm/driver/build_module.py", line 227, in build
- input_mod = lower(inputs, args, name=name, binds=binds)
- File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
- return ffi.lower_schedule(inp, args, name, binds, simple_mode)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
- tvm._ffi.base.TVMError: Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
- raise InstantiationError("Skipped because of invalid gpu kernel")
- tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
-
- Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 128, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4284533
- No: 6 GFLOPS: 0.00/90.55 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
- func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
- func = build(s, args, target=target, runtime=runtime)
- File "/workspace/python/tvm/driver/build_module.py", line 227, in build
- input_mod = lower(inputs, args, name=name, binds=binds)
- File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
- return ffi.lower_schedule(inp, args, name, binds, simple_mode)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
- tvm._ffi.base.TVMError: Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
- raise InstantiationError("Skipped because of invalid gpu kernel")
- tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
-
- Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 8, 8, 8]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4163217
- No: 7 GFLOPS: 222.61/222.61 result: MeasureResult(costs=(0.0010399281666666667,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.479459047317505, timestamp=1679006708.9954762) [('tile_f', [-1, 4, 4, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4304154
- No: 8 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
- func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
- func = build(s, args, target=target, runtime=runtime)
- File "/workspace/python/tvm/driver/build_module.py", line 227, in build
- input_mod = lower(inputs, args, name=name, binds=binds)
- File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
- return ffi.lower_schedule(inp, args, name, binds, simple_mode)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
- tvm._ffi.base.TVMError: Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
- raise InstantiationError("Skipped because of invalid gpu kernel")
- tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
-
- Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 32, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 256, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7420641
- No: 9 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 16, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 512, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7003199
+ No: 2 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1160,8 +513,8 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 1, 128]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 8, 64]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9851811
- No: 10 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 128, 1]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 256, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8195490
+ No: 3 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1283,8 +636,8 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 1, 512]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 32]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,1887379
- No: 11 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 4, 32]), ('tile_y', [-1, 7, 1, 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', 0), ('unroll_explicit', 1)],None,5966815
+ No: 4 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1406,8 +759,9 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 4, 64]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 16]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8060787
- No: 12 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 1, 128]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 64, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,3768371
+ No: 5 GFLOPS: 64.38/64.38 result: MeasureResult(costs=(0.003595801525,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.958209753036499, timestamp=1679043027.2978892) [('tile_f', [-1, 8, 8, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4103962
+ No: 6 GFLOPS: 0.00/64.38 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1529,8 +883,8 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 8, 4]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 8, 16]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6133285
- No: 13 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 16, 4, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 16, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9538473
+ No: 7 GFLOPS: 0.00/64.38 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1652,8 +1006,10 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 256, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4634399
- No: 14 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 2, 32]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2696950
+ No: 8 GFLOPS: 413.24/413.24 result: MeasureResult(costs=(0.0005602043184357541,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.21183180809021, timestamp=1679043028.17347) [('tile_f', [-1, 8, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9141003
+ No: 9 GFLOPS: 130.28/413.24 result: MeasureResult(costs=(0.0017769945714285714,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7461109161376953, timestamp=1679043034.4908655) [('tile_f', [-1, 2, 32, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5491991
+ No: 10 GFLOPS: 0.00/413.24 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1775,8 +1131,9 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 32, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8529752
- No: 15 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 128, 2, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 512]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,385731
+ No: 11 GFLOPS: 31.12/413.24 result: MeasureResult(costs=(0.0074379817142857145,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.8672966957092285, timestamp=1679043035.3957975) [('tile_f', [-1, 8, 4, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7647352
+ No: 12 GFLOPS: 0.00/413.24 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1898,8 +1255,8 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 32, 4]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 512]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,190431
- No: 16 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 1, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8262046
+ No: 13 GFLOPS: 0.00/413.24 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2021,8 +1378,8 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 8, 8, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 32, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1374150
- No: 17 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 2, 256]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 256, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,3744838
+ No: 14 GFLOPS: 0.00/413.24 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2144,8 +1501,10 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 16, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,3274074
- No: 18 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 64, 2, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 64]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4040756
+ No: 15 GFLOPS: 817.74/817.74 result: MeasureResult(costs=(0.00028309882245430813,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.182854175567627, timestamp=1679043039.8110738) [('tile_f', [-1, 1, 2, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 8]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7843944
+ No: 16 GFLOPS: 374.30/817.74 result: MeasureResult(costs=(0.0006184928658536585,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2676329612731934, timestamp=1679043040.6763678) [('tile_f', [-1, 1, 32, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,3721170
+ No: 17 GFLOPS: 0.00/817.74 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2267,8 +1626,9 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 4, 64]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 256]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,3091648
- No: 19 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 4, 128]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7338315
+ No: 18 GFLOPS: 4.06/817.74 result: MeasureResult(costs=(0.0570745505,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.9982800483703613, timestamp=1679043042.8743176) [('tile_f', [-1, 4, 4, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5425817
+ No: 19 GFLOPS: 0.00/817.74 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2390,8 +1750,8 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 256, 1]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,3293892
- No: 20 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 1, 8]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 64]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6750177
+ No: 20 GFLOPS: 0.00/817.74 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2513,7 +1873,7 @@ for this template
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 2, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 8, 64]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6171504
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 8, 1, 32]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 8, 32]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4994188
@@ -2568,9 +1928,9 @@ and measure running time.
Finish loading 20 records
Best config:
- [('tile_f', [-1, 4, 4, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4304154
+ [('tile_f', [-1, 1, 2, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 8]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7843944
Finish loading 20 records
- Time cost of this operator: 0.001350
+ Time cost of this operator: 0.000578
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 f41cea59f3..cd908930f5 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
@@ -360,10 +360,10 @@ 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 315.9 98.731 (1, 2, 10, 10, 3) 2 1 [315.9]
- tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.104 0.97 (1, 6, 10, 10) 1 1 [3.104]
- tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.955 0.299 (1, 1, 10, 10, 3) 1 1 [0.955]
- Total_time - 319.959 - - - - -
+ tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 316.2 98.746 (1, 2, 10, 10, 3) 2 1 [316.2]
+ tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.042 0.95 (1, 6, 10, 10) 1 1 [3.042]
+ tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.974 0.304 (1, 1, 10, 10, 3) 1 1 [0.974]
+ Total_time - 320.215 - - - - -
@@ -428,10 +428,10 @@ 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 103.8 97.509 (1, 6, 10, 10, 1) 2 1 [103.8]
- tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.802 1.692 (1, 6, 10, 10) 1 1 [1.802]
- tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.85 0.798 (1, 3, 10, 10, 1) 1 1 [0.85]
- Total_time - 106.451 - - - - -
+ tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 103.4 97.512 (1, 6, 10, 10, 1) 2 1 [103.4]
+ tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.781 1.679 (1, 6, 10, 10) 1 1 [1.781]
+ tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.857 0.808 (1, 3, 10, 10, 1) 1 1 [0.857]
+ Total_time - 106.038 - - - - -
@@ -439,7 +439,7 @@ Timing the tuned program
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 20.489 seconds)
+ **Total running time of the script:** ( 1 minutes 24.983 seconds)
.. _sphx_glr_download_how_to_work_with_microtvm_micro_autotune.py:
diff --git a/docs/_sources/how_to/work_with_microtvm/micro_pytorch.rst.txt b/docs/_sources/how_to/work_with_microtvm/micro_pytorch.rst.txt
index 0849cff02a..f4a4d1d6e0 100644
--- a/docs/_sources/how_to/work_with_microtvm/micro_pytorch.rst.txt
+++ b/docs/_sources/how_to/work_with_microtvm/micro_pytorch.rst.txt
@@ -118,7 +118,7 @@ download a cat image and preprocess it to use as the model input.
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/torch/ao/quantization/utils.py:281: UserWarning: must run observer before calling calculate_qparams. Returning default values.
"must run observer before calling calculate_qparams. " +
Downloading: "https://download.pytorch.org/models/quantized/mobilenet_v2_qnnpack_37f702c5.pth" to /workspace/.cache/torch/hub/checkpoints/mobilenet_v2_qnnpack_37f702c5.pth
-
0%| | 0.00/3.42M [00:00<?, ?B/s]
61%|###### | 2.09M/3.42M [00:00<00:00, 8.07MB/s]
100%|##########| 3.42M/3.42M [00:00<00:00, 13.0MB/s]
+
0%| | 0.00/3.42M [00:00<?, ?B/s]
100%|##########| 3.42M/3.42M [00:00<00:00, 50.5MB/s]
/workspace/python/tvm/relay/frontend/pytorch_utils.py:47: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
return LooseVersion(torch_ver) > ver
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/setuptools/_distutils/version.py:346: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
@@ -324,7 +324,7 @@ Look up prediction top 1 index in 1000 class synset.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 18.097 seconds)
+ **Total running time of the script:** ( 1 minutes 22.427 seconds)
.. _sphx_glr_download_how_to_work_with_microtvm_micro_pytorch.py:
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 6876a8c1ce..bea806723a 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
@@ -218,7 +218,7 @@ take about **2 minutes** to download the Stanford Cars, while COCO 2017 validati
.. code-block:: none
- '/tmp/tmpisplwp6d/images/random'
+ '/tmp/tmpwvh5kmcm/images/random'
@@ -309,7 +309,7 @@ objects to other stuff? We can display some examples from our datasets using ``m
.. image-sg:: /how_to/work_with_microtvm/images/sphx_glr_micro_train_001.png
- :alt: [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0]
+ :alt: [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0]
:srcset: /how_to/work_with_microtvm/images/sphx_glr_micro_train_001.png
:class: sphx-glr-single-img
@@ -318,8 +318,8 @@ objects to other stuff? We can display some examples from our datasets using ``m
.. code-block:: none
- /tmp/tmpisplwp6d/images/target contains 8144 images
- /tmp/tmpisplwp6d/images/random contains 5000 images
+ /tmp/tmpwvh5kmcm/images/target contains 8144 images
+ /tmp/tmpwvh5kmcm/images/random contains 5000 images
@@ -494,13 +494,13 @@ the time on our validation set).
.. code-block:: none
Epoch 1/3
- 328/328 - 48s - loss: 0.2149 - accuracy: 0.9244 - val_loss: 0.1410 - val_accuracy: 0.9532 - 48s/epoch - 146ms/step
+ 328/328 - 48s - loss: 0.2170 - accuracy: 0.9255 - val_loss: 0.1736 - val_accuracy: 0.9452 - 48s/epoch - 146ms/step
Epoch 2/3
- 328/328 - 44s - loss: 0.1080 - accuracy: 0.9608 - val_loss: 0.1752 - val_accuracy: 0.9392 - 44s/epoch - 133ms/step
+ 328/328 - 44s - loss: 0.0959 - accuracy: 0.9645 - val_loss: 0.1144 - val_accuracy: 0.9603 - 44s/epoch - 133ms/step
Epoch 3/3
- 328/328 - 43s - loss: 0.0610 - accuracy: 0.9776 - val_loss: 0.1257 - val_accuracy: 0.9535 - 43s/epoch - 132ms/step
+ 328/328 - 47s - loss: 0.0731 - accuracy: 0.9731 - val_loss: 0.1512 - val_accuracy: 0.9551 - 47s/epoch - 143ms/step
- <keras.callbacks.History object at 0x7fa6d4bbb950>
+ <keras.callbacks.History object at 0x7fed646bcf90>
@@ -861,7 +861,7 @@ Arduino tutorial for how to do that `on GitHub <https://github.com/guberti/tvm-a
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 4 minutes 42.118 seconds)
+ **Total running time of the script:** ( 4 minutes 51.956 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 5a373b193b..4041bbeb87 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
=================
-**07:45.881** total execution time for **how_to_work_with_microtvm** files:
+**08:06.016** total execution time for **how_to_work_with_microtvm** files:
+-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``) | 04:42.118 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``) | 04:51.956 | 0.0 MB |
+-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``) | 01:20.489 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``) | 01:24.983 | 0.0 MB |
+-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_pytorch.py` (``micro_pytorch.py``) | 01:18.097 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_pytorch.py` (``micro_pytorch.py``) | 01:22.427 | 0.0 MB |
+-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``) | 00:10.158 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``) | 00:10.690 | 0.0 MB |
+-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_custom_ide.py` (``micro_custom_ide.py``) | 00:07.843 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_custom_ide.py` (``micro_custom_ide.py``) | 00:08.286 | 0.0 MB |
+-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``) | 00:07.177 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``) | 00:07.673 | 0.0 MB |
+-----------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_microtvm_micro_ethosu.py` (``micro_ethosu.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 b5b8190f46..884a87c9c3 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
=================
-**00:46.294** total execution time for **how_to_work_with_relay** files:
+**00:48.171** 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``) | 00:34.066 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_pipeline_executor.py` (``using_pipeline_executor.py``) | 00:35.830 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``) | 00:10.478 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``) | 00:10.696 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``) | 00:01.744 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``) | 00:01.639 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_relay_using_relay_viz.py` (``using_relay_viz.py``) | 00:00.006 | 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 4e1d6d441c..d088b25af1 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
@@ -264,7 +264,7 @@ The following example customizes CUDA lowering rule for :code:`exp`.
.. code-block:: none
- <function my_cuda_math_rule at 0x7fa587119e60>
+ <function my_cuda_math_rule at 0x7fec19fa0a70>
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 bd9c97f0a6..6cbc31bd78 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
=================
-**00:08.367** total execution time for **how_to_work_with_schedules** files:
+**00:07.806** total execution time for **how_to_work_with_schedules** files:
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``) | 00:05.758 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``) | 00:05.107 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``) | 00:01.214 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``) | 00:01.227 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``) | 00:00.579 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``) | 00:00.615 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``) | 00:00.560 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``) | 00:00.589 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_extern_op.py` (``extern_op.py``) | 00:00.116 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_extern_op.py` (``extern_op.py``) | 00:00.123 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tedd.py` (``tedd.py``) | 00:00.060 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tedd.py` (``tedd.py``) | 00:00.062 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_schedule_primitives.py` (``schedule_primitives.py``) | 00:00.052 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_schedule_primitives.py` (``schedule_primitives.py``) | 00:00.056 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tuple_inputs.py` (``tuple_inputs.py``) | 00:00.026 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tuple_inputs.py` (``tuple_inputs.py``) | 00:00.028 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
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 562f509539..ff309b5575 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
=================
-**00:31.074** total execution time for **topic_vta_tutorials_autotvm** files:
+**00:32.764** total execution time for **topic_vta_tutorials_autotvm** files:
+---------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:31.068 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:32.757 | 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 294fea343d..e0a25906a0 100644
--- a/docs/_sources/topic/vta/tutorials/frontend/deploy_classification.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/frontend/deploy_classification.rst.txt
@@ -293,7 +293,7 @@ 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 33.11s!
+ resnet18_v1 inference graph built in 35.70s!
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 eef2fbcda8..185d2e4136 100644
--- a/docs/_sources/topic/vta/tutorials/frontend/deploy_detection.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/frontend/deploy_detection.rst.txt
@@ -337,7 +337,7 @@ The compilation steps are:
/workspace/python/tvm/relay/build_module.py:348: DeprecationWarning: Please use input parameter mod (tvm.IRModule) instead of deprecated parameter mod (tvm.relay.function.Function)
DeprecationWarning,
- yolov3-tiny inference graph built in 22.65s!
+ yolov3-tiny inference graph built in 23.97s!
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 9fb1b6d350..2343d6c8de 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
=================
-**01:39.612** total execution time for **topic_vta_tutorials_frontend** files:
+**01:43.772** total execution time for **topic_vta_tutorials_frontend** files:
+------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:49.832 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:52.694 | 0.0 MB |
+------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``) | 00:49.780 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``) | 00:51.078 | 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 63c270d98c..09c21deedc 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
=================
-**00:03.170** total execution time for **topic_vta_tutorials_optimize** files:
+**00:03.254** total execution time for **topic_vta_tutorials_optimize** files:
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``) | 00:02.704 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``) | 00:02.767 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.466 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.487 | 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 7557c23f06..90dbbaeab1 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
=================
-**00:00.763** total execution time for **topic_vta_tutorials** files:
+**00:00.827** total execution time for **topic_vta_tutorials** files:
+---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.395 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.429 | 0.0 MB |
+---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.367 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.398 | 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 093b20a5c2..13fba55cde 100644
--- a/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
+++ b/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
@@ -207,6 +207,13 @@ trials, we can load the best schedule from the log file and apply it.
+.. rst-class:: sphx-glr-script-out
+
+ .. code-block:: none
+
+ *E
+ *E
+
@@ -318,7 +325,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 93.580 ms
+ Execution time of this operator: 97.836 ms
@@ -434,7 +441,7 @@ operations.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 26.914 seconds)
+ **Total running time of the script:** ( 1 minutes 53.216 seconds)
.. _sphx_glr_download_tutorial_auto_scheduler_matmul_x86.py:
diff --git a/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt b/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt
index e6e7e71c30..f4b4245c7b 100644
--- a/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt
+++ b/docs/_sources/tutorial/autotvm_matmul_x86.rst.txt
@@ -454,16 +454,16 @@ reduce variance, we take 5 measurements and average them.
waiting for device...
device available
Get devices for measurement successfully!
- No: 1 GFLOPS: 10.60/10.60 result: MeasureResult(costs=(0.0253205662,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6898982524871826, timestamp=1679005063.0570695) [('tile_y', [-1, 8]), ('tile_x', [-1, 32])],None,53
- No: 2 GFLOPS: 1.91/10.60 result: MeasureResult(costs=(0.1405874398,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.487851619720459, timestamp=1679005066.8030055) [('tile_y', [-1, 4]), ('tile_x', [-1, 1])],None,2
- No: 3 GFLOPS: 3.96/10.60 result: MeasureResult(costs=(0.0677222502,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.3592486381530762, timestamp=1679005068.146178) [('tile_y', [-1, 64]), ('tile_x', [-1, 16])],None,46
- No: 4 GFLOPS: 12.58/12.58 result: MeasureResult(costs=(0.0213350322,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5824542045593262, timestamp=1679005069.9907389) [('tile_y', [-1, 32]), ('tile_x', [-1, 128])],None,75
- No: 5 GFLOPS: 1.86/12.58 result: MeasureResult(costs=(0.1444032902,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.5449867248535156, timestamp=1679005072.6575294) [('tile_y', [-1, 8]), ('tile_x', [-1, 1])],None,3
- No: 6 GFLOPS: 8.58/12.58 result: MeasureResult(costs=(0.0312723126,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.749152421951294, timestamp=1679005074.6644077) [('tile_y', [-1, 512]), ('tile_x', [-1, 32])],None,59
- No: 7 GFLOPS: 2.57/12.58 result: MeasureResult(costs=(0.10442363319999999,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.910839557647705, timestamp=1679005076.5930924) [('tile_y', [-1, 512]), ('tile_x', [-1, 8])],None,39
- No: 8 GFLOPS: 11.49/12.58 result: MeasureResult(costs=(0.023366896,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.621152400970459, timestamp=1679005077.2288888) [('tile_y', [-1, 128]), ('tile_x', [-1, 512])],None,97
- No: 9 GFLOPS: 10.43/12.58 result: MeasureResult(costs=(0.025730791599999997,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6473701000213623, timestamp=1679005077.9926288) [('tile_y', [-1, 4]), ('tile_x', [-1, 64])],None,62
- No: 10 GFLOPS: 3.68/12.58 result: MeasureResult(costs=(0.072924058,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.3931035995483398, timestamp=1679005079.418432) [('tile_y', [-1, 128]), ('tile_x', [-1, 16])],None,47
+ No: 1 GFLOPS: 3.65/3.65 result: MeasureResult(costs=(0.0736246602,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.446716547012329, timestamp=1679041304.1515517) [('tile_y', [-1, 128]), ('tile_x', [-1, 16])],None,47
+ No: 2 GFLOPS: 3.90/3.90 result: MeasureResult(costs=(0.06877848140000001,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.3615875244140625, timestamp=1679041305.5136292) [('tile_y', [-1, 16]), ('tile_x', [-1, 8])],None,34
+ No: 3 GFLOPS: 9.94/9.94 result: MeasureResult(costs=(0.0270128804,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6660201549530029, timestamp=1679041307.5284252) [('tile_y', [-1, 256]), ('tile_x', [-1, 512])],None,98
+ No: 4 GFLOPS: 4.37/9.94 result: MeasureResult(costs=(0.061452820799999995,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2496914863586426, timestamp=1679041310.0647764) [('tile_y', [-1, 8]), ('tile_x', [-1, 16])],None,43
+ No: 5 GFLOPS: 1.86/9.94 result: MeasureResult(costs=(0.1447007278,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.568298101425171, timestamp=1679041312.7782369) [('tile_y', [-1, 512]), ('tile_x', [-1, 8])],None,39
+ No: 6 GFLOPS: 2.09/9.94 result: MeasureResult(costs=(0.1282642068,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.3175342082977295, timestamp=1679041316.414036) [('tile_y', [-1, 128]), ('tile_x', [-1, 4])],None,27
+ No: 7 GFLOPS: 1.55/9.94 result: MeasureResult(costs=(0.1728739302,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.013164758682251, timestamp=1679041319.4385624) [('tile_y', [-1, 16]), ('tile_x', [-1, 1])],None,4
+ No: 8 GFLOPS: 11.54/11.54 result: MeasureResult(costs=(0.023261152799999997,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6456809043884277, timestamp=1679041320.0788038) [('tile_y', [-1, 8]), ('tile_x', [-1, 256])],None,83
+ No: 9 GFLOPS: 3.47/11.54 result: MeasureResult(costs=(0.07725540519999999,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.4549603462219238, timestamp=1679041321.6509154) [('tile_y', [-1, 2]), ('tile_x', [-1, 8])],None,31
+ No: 10 GFLOPS: 2.55/11.54 result: MeasureResult(costs=(0.1054291776,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.9046876430511475, timestamp=1679041323.6011984) [('tile_y', [-1, 4]), ('tile_x', [-1, 2])],None,12
diff --git a/docs/_sources/tutorial/autotvm_relay_x86.rst.txt b/docs/_sources/tutorial/autotvm_relay_x86.rst.txt
index d6380d68e4..80ef1c9341 100644
--- a/docs/_sources/tutorial/autotvm_relay_x86.rst.txt
+++ b/docs/_sources/tutorial/autotvm_relay_x86.rst.txt
@@ -311,7 +311,7 @@ standard deviation.
.. code-block:: none
- {'mean': 515.8791062600017, 'median': 516.126306450002, 'std': 1.534800583169039}
+ {'mean': 523.4406105400001, 'median': 522.6159517000042, 'std': 2.747173856786589}
@@ -545,31 +545,29 @@ the tuning data to.
.. code-block:: none
-
[Task 1/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 1/25] Current/Best: 17.60/ 21.58 GFLOPS | Progress: (4/20) | 12.42 s
[Task 1/25] Current/Best: 9.05/ 21.58 GFLOPS | Progress: (8/20) | 17.43 s
[Task 1/25] Current/Best: 12.17/ 22.97 GFLOPS | Progress: (12/20) | 20.35 s
[Task 1/25] Current/Best: 7.76/ 22.97 GFLOPS | Progress: (16/20) | 26.31 s
[Task 1/25] Current/Best: 9.48/ 23.04 GFLOPS | Progress: (20/20) | 30.01 s Done.
-
[Task 2/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 2/25] Current/Best: 20.89/ 20.89 GFLOPS | Progress: (4/20) | 4.46 s
[Task 2/25] Current/Best: 5.81/ 20.89 GFLOPS | Progress: (8/20) | 6.68 s
[Task 2/25] Current/Best: 23.05/ 23.05 GFLOPS | Progress: (12/20) | 8.34 s
[Task 2/25] Current/Best: 16.94/ 23.05 GFLOPS | Progress: (16/20) | 9.71 s
[Task 2/25] Current/Best: 5.27/ 23.05 GFLOPS | Progress: (20/20) | 11.67 s Done.
-
[Task 3/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 3/25] Current/Best: 15.79/ 15.79 GFLOPS | Progress: (4/20) | 5.05 s
[Task 3/25] Current/Best: 9.05/ 18.37 GFLOPS | Progress: (8/20) | 8.02 s
[Task 3/25] Current/Best: 16.35/ 18.37 GFLOPS | Progress: (12/20) | 11.11 s
[Task 3/25] Current/Best: 14.75/ 18.75 GFLOPS | Progress: (16/20) | 13.48 s
[Task 3/25] Current/Best: 11.28/ 18.75 GFLOPS | Progress: (20/20) | 16.01 s Done.
-
[Task 4/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 4/25] Current/Best: 9.68/ 20.90 GFLOPS | Progress: (4/20) | 5.63 s
[Task 4/25] Current/Best: 5.82/ 20.90 GFLOPS | Progress: (8/20) | 7.67 s
[Task 4/25] Current/Best: 18.53/ 20.90 GFLOPS | Progress: (12/20) | 10.10 s
[Task 4/25] Current/Best: 12.39/ 21.38 GFLOPS | Progress: (16/20) | 12.92 s
[Task 4/25] Current/Best: 15.73/ 21.38 GFLOPS | Progress: (20/20) | 15.23 s Done.
-
[Task 5/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 5/25] Current/Best: 6.02/ 19.05 GFLOPS | Progress: (4/20) | 4.93 s
[Task 5/25] Current/Best: 17.69/ 19.60 GFLOPS | Progress: (8/20) | 6.64 s
[Task 5/25] Current/Best: 9.75/ 19.60 GFLOPS | Progress: (12/20) | 10.32 s
[Task 5/25] Current/Best: 14.12/ 19.60 GFLOPS | Progress: (16/20) | 12.33 s
[Task 5/25] Current/Best: 13.86/ 19.60 GFLOPS | Progress: (20/20) | 14.66 s Done.
-
[Task 6/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 6/25] Current/Best: 4.08/ 20.36 GFLOPS | Progress: (4/20) | 9.32 s
[Task 6/25] Current/Best: 8.01/ 20.36 GFLOPS | Progress: (8/20) | 13.62 s
[Task 6/25] Current/Best: 19.55/ 20.36 GFLOPS | Progress: (12/20) | 16.17 s
[Task 6/25] Current/Best: 16.43/ 20.36 GFLOPS | Progress: (16/20) | 18.82 s
[Task 6/25] Current/Best: 13.05/ 20.36 GFLOPS | Progress: (20/20) | 22.19 s Done.
-
[Task 7/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 7/25] Current/Best: 11.53/ 14.17 GFLOPS | Progress: (4/20) | 5.57 s
[Task 7/25] Current/Best: 7.99/ 19.42 GFLOPS | Progress: (8/20) | 7.79 s
[Task 7/25] Current/Best: 8.10/ 19.42 GFLOPS | Progress: (12/20) | 11.38 s
[Task 7/25] Current/Best: 12.92/ 19.42 GFLOPS | Progress: (16/20) | 13.77 s
[Task 7/25] Current/Best: 11.68/ 19.42 GFLOPS | Progress: (20/20) | 16.62 s Done.
-
[Task 8/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 8/25] Current/Best: 10.51/ 12.96 GFLOPS | Progress: (4/20) | 12.40 s
[Task 8/25] Current/Best: 16.07/ 16.07 GFLOPS | Progress: (8/20) | 16.64 s
[Task 8/25] Current/Best: 13.25/ 16.07 GFLOPS | Progress: (12/20) | 28.19 s
[Task 8/25] Current/Best: 14.10/ 16.07 GFLOPS | Progress: (16/20) | 33.01 s
[Task 8/25] Current/Best: 4.33/ 16.07 GFLOPS | Progress: (20/20) | 39.05 s
[Task 9/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 9/25] Current/Best: 7.64/ 20.45 GFLOPS | Progress: (4/20) | 6.08 s Done.
-
[Task 9/25] Current/Best: 16.01/ 20.45 GFLOPS | Progress: (8/20) | 9.80 s
[Task 9/25] Current/Best: 5.85/ 20.45 GFLOPS | Progress: (12/20) | 11.78 s
[Task 9/25] Current/Best: 12.95/ 20.45 GFLOPS | Progress: (16/20) | 13.79 s
[Task 9/25] Current/Best: 12.90/ 23.11 GFLOPS | Progress: (20/20) | 15.54 s Done.
-
[Task 10/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 10/25] Current/Best: 11.41/ 13.33 GFLOPS | Progress: (4/20) | 4.90 s
[Task 10/25] Current/Best: 6.73/ 13.87 GFLOPS | Progress: (8/20) | 7.82 s
[Task 10/25] Current/Best: 18.25/ 18.52 GFLOPS | Progress: (12/20) | 9.48 s
[Task 10/25] Current/Best: 18.98/ 18.98 GFLOPS | Progress: (16/20) | 11.28 s
[Task 10/25] Current/Best: 2.58/ 18.98 GFLOPS | Progress: (20/20) | 13.43 s Done.
-
[Task 11/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 11/25] Current/Best: 13.11/ 19.35 GFLOPS | Progress: (4/20) | 4.75 s
[Task 11/25] Current/Best: 14.03/ 19.35 GFLOPS | Progress: (8/20) | 7.21 s
[Task 11/25] Current/Best: 12.59/ 19.35 GFLOPS | Progress: (12/20) | 9.49 s
[Task 11/25] Current/Best: 15.12/ 19.85 GFLOPS | Progress: (16/20) | 11.80 s
[Task 11/25] Current/Best: 12.34/ 19.85 GFLOPS | Progress: (20/20) | 14.75 s Done.
-
[Task 12/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 12/25] Current/Best: 16.64/ 20.59 GFLOPS | Progress: (4/20) | 4.73 s
[Task 12/25] Current/Best: 16.69/ 20.59 GFLOPS | Progress: (8/20) | 7.74 s
[Task 12/25] Current/Best: 15.90/ 20.59 GFLOPS | Progress: (12/20) | 9.84 s
[Task 12/25] Current/Best: 17.55/ 20.59 GFLOPS | Progress: (16/20) | 12.11 s
[Task 12/25] Current/Best: 9.21/ 20.59 GFLOPS | Progress: (20/20) | 16.76 s Done.
-
[Task 13/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 13/25] Current/Best: 19.94/ 19.94 GFLOPS | Progress: (4/20) | 5.13 s
[Task 13/25] Current/Best: 9.79/ 19.94 GFLOPS | Progress: (8/20) | 8.43 s
[Task 13/25] Current/Best: 12.23/ 19.94 GFLOPS | Progress: (12/20) | 11.43 s
[Task 13/25] Current/Best: 17.58/ 19.94 GFLOPS | Progress: (16/20) | 14.98 s
[Task 13/25] Current/Best: 11.78/ 19.94 GFLOPS | Progress: (20/20) | 17.43 s Done.
-
[Task 14/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 14/25] Current/Best: 8.63/ 21.09 GFLOPS | Progress: (4/20) | 7.35 s
[Task 14/25] Current/Best: 13.16/ 21.09 GFLOPS | Progress: (8/20) | 9.80 s
[Task 14/25] Current/Best: 15.48/ 21.09 GFLOPS | Progress: (12/20) | 12.80 s
[Task 14/25] Current/Best: 15.72/ 21.09 GFLOPS | Progress: (16/20) | 15.50 s
[Task 14/25] Current/Best: 13.10/ 21.09 GFLOPS | Progress: (20/20) | 22.09 s
[Task 15/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 15/25] Current/Best: 12.41/ 17.51 GFLOPS | Progress: (4/20) | 5.52 s
[Task 15/25] Current/Best: 3.10/ 22.08 GFLOPS | Progress: (8/20) | 7.32 s
[Task 15/25] Current/Best: 21.43/ 22.08 GFLOPS | Progress: (12/20) | 9.30 s
[Task 15/25] Current/Best: 11.43/ 22.08 GFLOPS | Progress: (16/20) | 11.20 s Done.
-
[Task 15/25] Current/Best: 11.28/ 22.08 GFLOPS | Progress: (20/20) | 14.78 s Done.
-
[Task 16/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 16/25] Current/Best: 4.45/ 15.49 GFLOPS | Progress: (4/20) | 5.05 s
[Task 16/25] Current/Best: 7.54/ 15.49 GFLOPS | Progress: (8/20) | 8.49 s
[Task 16/25] Current/Best: 17.17/ 17.17 GFLOPS | Progress: (12/20) | 11.60 s
[Task 16/25] Current/Best: 16.81/ 17.17 GFLOPS | Progress: (16/20) | 13.16 s
[Task 16/25] Current/Best: 17.05/ 17.20 GFLOPS | Progress: (20/20) | 15.27 s Done.
-
[Task 17/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 17/25] Current/Best: 6.81/ 21.75 GFLOPS | Progress: (4/20) | 5.22 s
[Task 17/25] Current/Best: 12.29/ 22.05 GFLOPS | Progress: (8/20) | 7.45 s
[Task 17/25] Current/Best: 15.37/ 22.05 GFLOPS | Progress: (12/20) | 10.16 s
[Task 17/25] Current/Best: 19.80/ 22.05 GFLOPS | Progress: (16/20) | 12.42 s
[Task 17/25] Current/Best: 17.14/ 22.05 GFLOPS | Progress: (20/20) | 14.91 s Done.
-
[Task 18/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 18/25] Current/Best: 11.41/ 17.21 GFLOPS | Progress: (4/20) | 8.55 s
[Task 18/25] Current/Best: 18.69/ 18.69 GFLOPS | Progress: (8/20) | 11.06 s
[Task 18/25] Current/Best: 9.86/ 18.69 GFLOPS | Progress: (12/20) | 15.00 s
[Task 18/25] Current/Best: 15.35/ 18.69 GFLOPS | Progress: (16/20) | 17.58 s
[Task 18/25] Current/Best: 20.82/ 20.82 GFLOPS | Progress: (20/20) | 19.43 s Done.
-
[Task 19/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 19/25] Current/Best: 4.55/ 18.49 GFLOPS | Progress: (4/20) | 5.73 s
[Task 19/25] Current/Best: 18.23/ 20.69 GFLOPS | Progress: (8/20) | 8.09 s
[Task 19/25] Current/Best: 9.90/ 20.69 GFLOPS | Progress: (12/20) | 11.48 s
[Task 19/25] Current/Best: 4.36/ 20.69 GFLOPS | Progress: (16/20) | 17.41 s
[Task 19/25] Current/Best: 16.11/ 20.85 GFLOPS | Progress: (20/20) | 20.82 s Done.
-
[Task 20/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 20/25] Current/Best: 6.78/ 15.70 GFLOPS | Progress: (4/20) | 6.15 s
[Task 20/25] Current/Best: 2.70/ 15.70 GFLOPS | Progress: (8/20) | 12.33 s
[Task 20/25] Current/Best: 16.60/ 17.41 GFLOPS | Progress: (12/20) | 14.70 s
[Task 20/25] Current/Best: 5.49/ 17.41 GFLOPS | Progress: (16/20) | 17.32 s
[Task 20/25] Current/Best: 6.23/ 17.41 GFLOPS | Progress: (20/20) | 20.84 s
[Task 21/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 21/25] Current/Best: 5.18/ 18.60 GFLOPS | Progress: (4/20) | 5.40 s
[Task 21/25] Current/Best: 8.38/ 20.28 GFLOPS | Progress: (8/20) | 8.72 s
[Task 21/25] Current/Best: 16.59/ 20.28 GFLOPS | Progress: (12/20) | 11.56 s
[Task 21/25] Current/Best: 5.23/ 20.28 GFLOPS | Progress: (16/20) | 12.96 s
[Task 21/25] Current/Best: 11.13/ 20.28 GFLOPS | Progress: (20/2
0) | 16.39 s
[Task 22/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s Done.
+
[Task 1/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 1/25] Current/Best: 14.91/ 14.91 GFLOPS | Progress: (4/20) | 12.30 s
[Task 1/25] Current/Best: 10.18/ 23.66 GFLOPS | Progress: (8/20) | 15.91 s
[Task 1/25] Current/Best: 16.27/ 23.66 GFLOPS | Progress: (12/20) | 19.05 s
[Task 1/25] Current/Best: 6.03/ 23.66 GFLOPS | Progress: (16/20) | 22.93 s
[Task 1/25] Current/Best: 5.42/ 23.66 GFLOPS | Progress: (20/20) | 25.65 s Done.
+
[Task 2/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 2/25] Current/Best: 19.31/ 19.31 GFLOPS | Progress: (4/20) | 4.39 s
[Task 2/25] Current/Best: 15.15/ 19.31 GFLOPS | Progress: (8/20) | 6.39 s
[Task 2/25] Current/Best: 3.57/ 19.31 GFLOPS | Progress: (12/20) | 8.39 s
[Task 2/25] Current/Best: 12.59/ 19.31 GFLOPS | Progress: (16/20) | 9.87 s
[Task 2/25] Current/Best: 16.24/ 19.31 GFLOPS | Progress: (20/20) | 12.06 s Done.
+
[Task 3/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 3/25] Current/Best: 16.02/ 19.87 GFLOPS | Progress: (4/20) | 4.79 s
[Task 3/25] Current/Best: 19.92/ 19.92 GFLOPS | Progress: (8/20) | 6.86 s
[Task 3/25] Current/Best: 5.72/ 19.92 GFLOPS | Progress: (12/20) | 9.66 s
[Task 3/25] Current/Best: 17.73/ 19.92 GFLOPS | Progress: (16/20) | 11.99 s
[Task 3/25] Current/Best: 19.26/ 19.92 GFLOPS | Progress: (20/20) | 14.07 s Done.
+
[Task 4/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 4/25] Current/Best: 13.80/ 17.15 GFLOPS | Progress: (4/20) | 5.68 s
[Task 4/25] Current/Best: 11.84/ 17.15 GFLOPS | Progress: (8/20) | 7.85 s
[Task 4/25] Current/Best: 8.87/ 17.15 GFLOPS | Progress: (12/20) | 16.22 s
[Task 4/25] Current/Best: 17.28/ 17.28 GFLOPS | Progress: (16/20) | 20.77 s
[Task 4/25] Current/Best: 6.26/ 17.28 GFLOPS | Progress: (20/20) | 23.37 s Done.
+
[Task 5/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 5/25] Current/Best: 13.80/ 15.88 GFLOPS | Progress: (4/20) | 4.73 s
[Task 5/25] Current/Best: 8.81/ 20.67 GFLOPS | Progress: (8/20) | 6.74 s
[Task 5/25] Current/Best: 5.44/ 20.67 GFLOPS | Progress: (12/20) | 10.46 s
[Task 5/25] Current/Best: 12.16/ 20.67 GFLOPS | Progress: (16/20) | 12.57 s
[Task 5/25] Current/Best: 3.96/ 20.67 GFLOPS | Progress: (20/20) | 15.14 s Done.
+
[Task 6/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 6/25] Current/Best: 10.87/ 12.90 GFLOPS | Progress: (4/20) | 8.53 s
[Task 6/25] Current/Best: 5.70/ 18.27 GFLOPS | Progress: (8/20) | 11.32 s
[Task 6/25] Current/Best: 11.14/ 19.02 GFLOPS | Progress: (12/20) | 13.95 s
[Task 6/25] Current/Best: 5.56/ 19.02 GFLOPS | Progress: (16/20) | 16.40 s
[Task 6/25] Current/Best: 9.19/ 19.02 GFLOPS | Progress: (20/20) | 20.16 s Done.
+
[Task 7/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 7/25] Current/Best: 17.10/ 20.32 GFLOPS | Progress: (4/20) | 5.57 s
[Task 7/25] Current/Best: 19.02/ 20.32 GFLOPS | Progress: (8/20) | 7.98 s
[Task 7/25] Current/Best: 15.15/ 20.32 GFLOPS | Progress: (12/20) | 10.14 s
[Task 7/25] Current/Best: 14.09/ 20.32 GFLOPS | Progress: (16/20) | 12.37 s
[Task 7/25] Current/Best: 19.59/ 20.32 GFLOPS | Progress: (20/20) | 15.30 s Done.
+
[Task 8/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 8/25] Current/Best: 7.44/ 17.69 GFLOPS | Progress: (4/20) | 5.36 s
[Task 8/25] Current/Best: 9.51/ 17.69 GFLOPS | Progress: (8/20) | 18.53 s
[Task 8/25] Current/Best: 12.08/ 17.69 GFLOPS | Progress: (12/20) | 27.74 s
[Task 8/25] Current/Best: 9.68/ 18.00 GFLOPS | Progress: (16/20) | 29.97 s
[Task 8/25] Current/Best: 8.41/ 18.00 GFLOPS | Progress: (20/20) | 32.73 s Done.
+
[Task 9/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 9/25] Current/Best: 21.90/ 21.90 GFLOPS | Progress: (4/20) | 10.16 s
[Task 9/25] Current/Best: 16.25/ 21.90 GFLOPS | Progress: (8/20) | 21.05 s
[Task 9/25] Current/Best: 11.77/ 21.90 GFLOPS | Progress: (12/20) | 26.20 s
[Task 9/25] Current/Best: 12.20/ 22.46 GFLOPS | Progress: (16/20) | 31.49 s
[Task 9/25] Current/Best: 17.01/ 22.46 GFLOPS | Progress: (20/20) | 33.09 s Done.
+
[Task 10/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 10/25] Current/Best: 8.14/ 16.01 GFLOPS | Progress: (4/20) | 6.26 s
[Task 10/25] Current/Best: 13.90/ 16.01 GFLOPS | Progress: (8/20) | 8.34 s
[Task 10/25] Current/Best: 12.03/ 21.14 GFLOPS | Progress: (12/20) | 10.31 s
[Task 10/25] Current/Best: 10.85/ 21.14 GFLOPS | Progress: (16/20) | 12.40 s
[Task 10/25] Current/Best: 15.25/ 21.14 GFLOPS | Progress: (20/20) | 14.74 s Done.
+
[Task 11/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 11/25] Current/Best: 20.76/ 20.76 GFLOPS | Progress: (4/20) | 5.87 s
[Task 11/25] Current/Best: 13.37/ 20.76 GFLOPS | Progress: (8/20) | 8.96 s
[Task 11/25] Current/Best: 5.95/ 20.76 GFLOPS | Progress: (12/20) | 11.60 s
[Task 11/25] Current/Best: 12.48/ 21.74 GFLOPS | Progress: (16/20) | 14.46 s
[Task 11/25] Current/Best: 17.86/ 21.74 GFLOPS | Progress: (20/20) | 17.16 s Done.
+
[Task 12/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 12/25] Current/Best: 5.48/ 18.26 GFLOPS | Progress: (4/20) | 5.28 s
[Task 12/25] Current/Best: 5.74/ 18.26 GFLOPS | Progress: (8/20) | 12.23 s
[Task 12/25] Current/Best: 15.16/ 18.26 GFLOPS | Progress: (12/20) | 14.53 s
[Task 12/25] Current/Best: 9.79/ 18.26 GFLOPS | Progress: (16/20) | 19.84 s
[Task 12/25] Current/Best: 18.46/ 20.76 GFLOPS | Progress: (20/20) | 22.36 s Done.
+
[Task 13/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 13/25] Current/Best: 20.01/ 20.01 GFLOPS | Progress: (4/20) | 5.47 s
[Task 13/25] Current/Best: 16.76/ 20.01 GFLOPS | Progress: (8/20) | 8.85 s
[Task 13/25] Current/Best: 16.15/ 20.01 GFLOPS | Progress: (12/20) | 11.99 s
[Task 13/25] Current/Best: 17.35/ 20.01 GFLOPS | Progress: (16/20) | 15.33 s
[Task 13/25] Current/Best: 18.31/ 20.01 GFLOPS | Progress: (20/20) | 17.45 s Done.
+
[Task 14/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 14/25] Current/Best: 9.00/ 19.90 GFLOPS | Progress: (4/20) | 4.42 s
[Task 14/25] Current/Best: 8.07/ 19.90 GFLOPS | Progress: (8/20) | 8.87 s
[Task 14/25] Current/Best: 17.05/ 21.26 GFLOPS | Progress: (12/20) | 11.45 s
[Task 14/25] Current/Best: 5.79/ 21.26 GFLOPS | Progress: (16/20) | 13.67 s
[Task 14/25] Current/Best: 10.95/ 21.26 GFLOPS | Progress: (20/20) | 18.56 s
[Task 15/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 15/25] Current/Best: 6.14/ 16.94 GFLOPS | Progress: (4/20) | 5.98 s
[Task 15/25] Current/Best: 15.81/ 19.36 GFLOPS | Progress: (8/20) | 8.30 s
[Task 15/25] Current/Best: 11.34/ 19.36 GFLOPS | Progress: (12/20) | 13.29 s
[Task 15/25] Current/Best: 9.68/ 20.21 GFLOPS | Progress: (16/20) | 15.33 s
[Task 15/25] Current/Best: 16.50/ 20.21 GFLOPS | Progress: (20/20
) | 17.37 s
[Task 16/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 16/25] Current/Best: 18.86/ 19.65 GFLOPS | Progress: (4/20) | 5.72 s
[Task 16/25] Current/Best: 11.09/ 19.65 GFLOPS | Progress: (8/20) | 7.73 s
[Task 16/25] Current/Best: 18.27/ 19.65 GFLOPS | Progress: (12/20) | 9.49 s
[Task 16/25] Current/Best: 10.63/ 19.65 GFLOPS | Progress: (16/20) | 12.79 s
[Task 16/25] Current/Best: 8.24/ 19.65 GFLOPS | Progress: (20/20) | 14.83 s Done.
+
[Task 17/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 17/25] Current/Best: 18.03/ 19.63 GFLOPS | Progress: (4/20) | 5.44 s
[Task 17/25] Current/Best: 9.81/ 22.97 GFLOPS | Progress: (8/20) | 8.95 s
[Task 17/25] Current/Best: 4.68/ 22.97 GFLOPS | Progress: (12/20) | 12.73 s
[Task 17/25] Current/Best: 15.25/ 22.97 GFLOPS | Progress: (16/20) | 15.16 s
[Task 17/25] Current/Best: 17.57/ 22.97 GFLOPS | Progress: (20/20) | 17.42 s Done.
+
[Task 18/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 18/25] Current/Best: 7.05/ 13.56 GFLOPS | Progress: (4/20) | 5.67 s
[Task 18/25] Current/Best: 18.64/ 18.64 GFLOPS | Progress: (8/20) | 9.41 s
[Task 18/25] Current/Best: 14.41/ 19.45 GFLOPS | Progress: (12/20) | 11.42 s
[Task 18/25] Current/Best: 16.60/ 19.45 GFLOPS | Progress: (16/20) | 17.34 s
[Task 18/25] Current/Best: 7.73/ 20.57 GFLOPS | Progress: (20/20) | 21.06 s Done.
+
[Task 19/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 19/25] Current/Best: 18.13/ 18.32 GFLOPS | Progress: (4/20) | 6.89 s
[Task 19/25] Current/Best: 17.66/ 18.32 GFLOPS | Progress: (8/20) | 10.13 s
[Task 19/25] Current/Best: 16.41/ 18.61 GFLOPS | Progress: (12/20) | 13.97 s
[Task 19/25] Current/Best: 10.70/ 18.61 GFLOPS | Progress: (16/20) | 17.74 s
[Task 19/25] Current/Best: 16.38/ 18.61 GFLOPS | Progress: (20/20) | 20.75 s Done.
+
[Task 20/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 20/25] Current/Best: 3.08/ 13.82 GFLOPS | Progress: (4/20) | 6.02 s
[Task 20/25] Current/Best: 10.59/ 13.82 GFLOPS | Progress: (8/20) | 8.93 s
[Task 20/25] Current/Best: 9.50/ 13.82 GFLOPS | Progress: (12/20) | 11.20 s
[Task 20/25] Current/Best: 7.63/ 13.82 GFLOPS | Progress: (16/20) | 16.05 s
[Task 20/25] Current/Best: 19.38/ 19.38 GFLOPS | Progress: (20/20) | 17.98 s
[Task 21/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 21/25] Current/Best: 11.00/ 12.12 GFLOPS | Progress: (4/20) | 8.16 s Done.
Done.
-
[Task 22/25] Current/Best: 19.66/ 19.66 GFLOPS | Progress: (4/20) | 5.66 s
[Task 22/25] Current/Best: 20.06/ 20.06 GFLOPS | Progress: (8/20) | 8.21 s
[Task 22/25] Current/Best: 10.11/ 20.06 GFLOPS | Progress: (12/20) | 11.65 s
[Task 22/25] Current/Best: 16.41/ 20.06 GFLOPS | Progress: (16/20) | 13.42 s
[Task 22/25] Current/Best: 19.68/ 20.06 GFLOPS | Progress: (20/20) | 15.61 s Done.
-
[Task 23/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 23/25] Current/Best: 11.70/ 11.70 GFLOPS | Progress: (4/20) | 6.88 s
[Task 23/25] Current/Best: 2.70/ 18.42 GFLOPS | Progress: (8/20) | 10.52 s
[Task 23/25] Current/Best: 12.48/ 20.28 GFLOPS | Progress: (12/20) | 13.87 s
[Task 23/25] Current/Best: 10.69/ 22.35 GFLOPS | Progress: (16/20) | 17.13 s
[Task 23/25] Current/Best: 18.66/ 22.35 GFLOPS | Progress: (20/20) | 20.70 s Done.
-
[Task 24/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 24/25] Current/Best: 3.59/ 10.17 GFLOPS | Progress: (4/20) | 13.42 s
[Task 24/25] Current/Best: 3.02/ 10.17 GFLOPS | Progress: (8/20) | 26.07 s
[Task 24/25] Current/Best: 1.68/ 10.17 GFLOPS | Progress: (12/20) | 33.24 s
[Task 24/25] Current/Best: 0.74/ 10.17 GFLOPS | Progress: (16/20) | 43.94 s
[Task 24/25] Current/Best: 2.06/ 10.17 GFLOPS | Progress: (20/20) | 47.20 s
[Task 25/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s Done.
-
[Task 25/25] Current/Best: 8.89/ 9.53 GFLOPS | Progress: (4/20) | 13.77 s
[Task 25/25] Current/Best: 3.32/ 9.53 GFLOPS | Progress: (8/20) | 19.89 s
[Task 25/25] Current/Best: 8.85/ 9.53 GFLOPS | Progress: (12/20) | 24.83 s
[Task 25/25] Current/Best: 5.85/ 9.53 GFLOPS | Progress: (16/20) | 35.51 s
[Task 25/25] Current/Best: 1.55/ 9.53 GFLOPS | Progress: (20/20) | 38.45 s
+ Done.
+
[Task 21/25] Current/Best: 5.24/ 19.75 GFLOPS | Progress: (8/20) | 11.11 s
[Task 21/25] Current/Best: 19.82/ 19.82 GFLOPS | Progress: (12/20) | 16.07 s
[Task 21/25] Current/Best: 6.48/ 19.82 GFLOPS | Progress: (16/20) | 19.05 s
[Task 21/25] Current/Best: 19.01/ 19.82 GFLOPS | Progress: (20/20) | 22.01 s
[Task 22/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 22/25] Current/Best: 4.17/ 13.51 GFLOPS | Progress: (4/20) | 5.03 s
[Task 22/25] Current/Best: 10.79/ 17.22 GFLOPS | Progress: (8/20) | 7.92 s
[Task 22/25] Current/Best: 10.53/ 20.85 GFLOPS | Progress: (12/20) | 9.59 s
[Task 22/25] Current/Best: 11.60/ 20.85 GFLOPS | Progress: (16/20) | 11.53 s
[Task 22/25] Current/Best: 2.49/ 20.85 GFLOPS | Progress: (20/20) | 13.73 s Done.
+
[Task 23/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 23/25] Current/Best: 10.47/ 22.06 GFLOPS | Progress: (4/20) | 5.56 s
[Task 23/25] Current/Best: 23.06/ 23.06 GFLOPS | Progress: (8/20) | 8.43 s
[Task 23/25] Current/Best: 17.76/ 23.06 GFLOPS | Progress: (12/20) | 16.72 s
[Task 23/25] Current/Best: 8.09/ 23.06 GFLOPS | Progress: (16/20) | 21.12 s
[Task 23/25] Current/Best: 6.76/ 23.06 GFLOPS | Progress: (20/20) | 25.14 s Done.
+
[Task 24/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 24/25] Current/Best: 5.73/ 5.73 GFLOPS | Progress: (4/20) | 13.57 s
[Task 24/25] Current/Best: 8.08/ 9.99 GFLOPS | Progress: (8/20) | 24.54 s
[Task 24/25] Current/Best: 1.12/ 9.99 GFLOPS | Progress: (12/20) | 35.57 s
[Task 24/25] Current/Best: 8.22/ 9.99 GFLOPS | Progress: (16/20) | 38.74 s
[Task 24/25] Current/Best: 1.78/ 9.99 GFLOPS | Progress: (20/20) | 49.76 s
[Task 25/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 25/25] Current/Best: 5.56/ 5.56 GFLOPS | Progress: (4/20) | 13.92 s
[Task 25/25] Current/Best: 4.13/ 5.61 GFLOPS | Progress: (8/20) | 24.60 s
[Task 25/25] Current/Best: 8.30/ 8.30 GFLOPS | Progress: (12/20) | 27.37 s
[Task 25/25] Current/Best: 7.33/ 8.30 GFLOPS | Progress: (16/20) | 38.35 s
[Task 25/25] Current/Best: 8.87/ 8.89 GFLOPS | Progress: (2
0/20) | 43.21 s
@@ -723,8 +721,8 @@ improvement in comparing the optimized model to the unoptimized model.
.. code-block:: none
- optimized: {'mean': 436.5033648499957, 'median': 436.46932330000254, 'std': 3.5534706870024184}
- unoptimized: {'mean': 515.8791062600017, 'median': 516.126306450002, 'std': 1.534800583169039}
+ optimized: {'mean': 421.37053469999955, 'median': 423.0767754999988, 'std': 5.286717630721996}
+ unoptimized: {'mean': 523.4406105400001, 'median': 522.6159517000042, 'std': 2.747173856786589}
@@ -747,7 +745,7 @@ profiling/benchmarking.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 12 minutes 24.027 seconds)
+ **Total running time of the script:** ( 13 minutes 3.834 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 19c4b7d391..7f2680eeab 100644
--- a/docs/_sources/tutorial/cross_compilation_and_rpc.rst.txt
+++ b/docs/_sources/tutorial/cross_compilation_and_rpc.rst.txt
@@ -274,7 +274,7 @@ device and returns the measured cost. Network overhead is excluded.
.. code-block:: none
- 1.24e-07 secs/op
+ 1.217e-07 secs/op
diff --git a/docs/_sources/tutorial/intro_topi.rst.txt b/docs/_sources/tutorial/intro_topi.rst.txt
index 79b1f5e2e5..26b253f771 100644
--- a/docs/_sources/tutorial/intro_topi.rst.txt
+++ b/docs/_sources/tutorial/intro_topi.rst.txt
@@ -270,7 +270,7 @@ As you can see, scheduled stages of computation have been accumulated and we can
.. code-block:: none
- [stage(a, placeholder(a, 0x13a0e880)), stage(b, placeholder(b, 0x22e0b3f0)), stage(T_add, compute(T_add, body=[a[ax0, ax1, ax2] + b[ax1, ax2]], axis=[T.iter_var(ax0, T.Range(0, 100), "DataPar", ""), T.iter_var(ax1, T.Range(0, 10), "DataPar", ""), T.iter_var(ax2, T.Range(0, 10), "DataPar", "")], reduce_axis=[], tag=broadcast, attrs={})), stage(T_multiply, compute(T_multiply, body=[a[ax0, ax1, ax2] * b[ax1, ax2]], axis=[T.iter_var(ax0, T.Range(0, 100), "DataPar", ""), T.iter_var(ax1, T [...]
+ [stage(a, placeholder(a, 0x25956810)), stage(b, placeholder(b, 0x25954850)), stage(T_add, compute(T_add, body=[a[ax0, ax1, ax2] + b[ax1, ax2]], axis=[T.iter_var(ax0, T.Range(0, 100), "DataPar", ""), T.iter_var(ax1, T.Range(0, 10), "DataPar", ""), T.iter_var(ax2, T.Range(0, 10), "DataPar", "")], reduce_axis=[], tag=broadcast, attrs={})), stage(T_multiply, compute(T_multiply, body=[a[ax0, ax1, ax2] * b[ax1, ax2]], axis=[T.iter_var(ax0, T.Range(0, 100), "DataPar", ""), T.iter_var(ax1, T [...]
diff --git a/docs/_sources/tutorial/sg_execution_times.rst.txt b/docs/_sources/tutorial/sg_execution_times.rst.txt
index 59073e6ba0..b467b3e372 100644
--- a/docs/_sources/tutorial/sg_execution_times.rst.txt
+++ b/docs/_sources/tutorial/sg_execution_times.rst.txt
@@ -5,24 +5,24 @@
Computation times
=================
-**15:54.013** total execution time for **tutorial** files:
+**17:06.283** total execution time for **tutorial** files:
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``) | 12:24.027 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``) | 13:03.834 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 01:26.914 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 01:53.216 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``) | 01:00.742 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``) | 01:01.207 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``) | 00:36.341 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``) | 00:37.784 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``) | 00:23.469 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``) | 00:27.678 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``) | 00:01.487 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``) | 00:01.481 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``) | 00:00.856 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``) | 00:00.881 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.176 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.201 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_tutorial_uma.py` (``uma.py``) | 00:00.000 | 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 585d27590a..e75a180c67 100644
--- a/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
+++ b/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
@@ -285,7 +285,7 @@ helper function to run a profile of the TVM generated code.
.. code-block:: none
- Numpy running time: 0.000007
+ Numpy running time: 0.000008
naive: 0.000007
@@ -498,10 +498,10 @@ We can now compare the different schedules
.. code-block:: none
Operator Timing Performance
- numpy 7.1122700001069465e-06 1.0
- naive 6.6776e-06 0.9388844911539621
- parallel 6.9934e-06 0.9832866299922305
- vector 2.4634900000000004e-05 3.4637183346005664
+ numpy 7.853959998556092e-06 1.0
+ naive 6.6807e-06 0.8506154858476757
+ parallel 7.3841e-06 0.9401754021356776
+ vector 2.4500200000000003e-05 3.1194709426205676
@@ -922,7 +922,7 @@ matrix multiplication.
.. code-block:: none
- Numpy running time: 0.017754
+ Numpy running time: 0.019062
@@ -980,7 +980,7 @@ optimizations.
.. code-block:: none
- none: 3.397279
+ none: 3.369341
@@ -1080,7 +1080,7 @@ schedule.
.. code-block:: none
- blocking: 0.301418
+ blocking: 0.326556
@@ -1164,7 +1164,7 @@ already cache friendly from our previous optimizations.
.. code-block:: none
- vectorization: 0.335739
+ vectorization: 0.352336
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1230,7 +1230,7 @@ more cache friendly.
.. code-block:: none
- loop permutation: 0.114749
+ loop permutation: 0.123854
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1321,7 +1321,7 @@ optimized schedule.
.. code-block:: none
- array packing: 0.107613
+ array packing: 0.108130
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1404,7 +1404,7 @@ to `C` when all the block results are ready.
.. code-block:: none
- block caching: 0.110977
+ block caching: 0.111080
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1478,7 +1478,7 @@ of thread-level parallelization.
.. code-block:: none
- parallelization: 0.147000
+ parallelization: 0.147349
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1548,13 +1548,13 @@ working, we can compare the results.
.. code-block:: none
Operator Timing Performance
- none 3.3972791391 1.0
- blocking 0.3014175217 0.08872321330058586
- vectorization 0.3357389099 0.0988258238882729
- loop permutation 0.114748958 0.03377672346064534
- array packing 0.1076130525 0.03167624681217941
- block caching 0.11097699319999998 0.03266643353580883
- parallelization 0.1470003962 0.04327003763339359
+ none 3.3693411657 1.0
+ blocking 0.32655604720000003 0.09691985202458896
+ vectorization 0.3523355961 0.10457106560973628
+ loop permutation 0.1238537977 0.03675905514135393
+ array packing 0.108130419 0.03209245181247038
+ block caching 0.11108006919999999 0.03296789008213197
+ parallelization 0.1473494413 0.04373241950088699
@@ -1596,7 +1596,7 @@ the computation for specific platforms.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 0.742 seconds)
+ **Total running time of the script:** ( 1 minutes 1.207 seconds)
.. _sphx_glr_download_tutorial_tensor_expr_get_started.py:
diff --git a/docs/commit_hash b/docs/commit_hash
index bc556b0867..a1b31f30a2 100644
--- a/docs/commit_hash
+++ b/docs/commit_hash
@@ -1 +1 @@
-32e500b7f7b48d7427a3ba7367d12b69baf782de
+d1dea13762dca4a6285d7ae58542084fb34f5f3c
diff --git a/docs/how_to/compile_models/from_darknet.html b/docs/how_to/compile_models/from_darknet.html
index 8f30cd610d..a583cc4755 100644
--- a/docs/how_to/compile_models/from_darknet.html
+++ b/docs/how_to/compile_models/from_darknet.html
@@ -590,7 +590,7 @@ 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 21.154 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 24.180 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-compile-models-from-darknet-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/7716f96385bd5abb6e822041e285be54/from_darknet.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">from_darknet.py</span></code></a></p>
diff --git a/docs/how_to/compile_models/from_keras.html b/docs/how_to/compile_models/from_keras.html
index 66fce2cc9e..40ffa90b02 100644
--- a/docs/how_to/compile_models/from_keras.html
+++ b/docs/how_to/compile_models/from_keras.html
@@ -511,7 +511,7 @@ Tensorflow is also required since it’s used as the default backend of keras.</
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Relay top-1 id: 285, class name: Egyptian cat
1/1 [==============================] - ETA: 0s
-1/1 [==============================] - 1s 934ms/step
+1/1 [==============================] - 1s 994ms/step
Keras top-1 id: 285, class name: Egyptian cat
</pre></div>
</div>
diff --git a/docs/how_to/compile_models/from_mxnet.html b/docs/how_to/compile_models/from_mxnet.html
index 09c46f8318..aaa7a54d14 100644
--- a/docs/how_to/compile_models/from_mxnet.html
+++ b/docs/how_to/compile_models/from_mxnet.html
@@ -444,7 +444,7 @@
<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.zip7cfa7269-8f11-4505-8ca3-485484b15aee 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.zip6618f4c1-5f4e-4765-b2df-6c1622fe77fb 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 5dcc7d5355..52dfb88586 100644
--- a/docs/how_to/compile_models/from_oneflow.html
+++ b/docs/how_to/compile_models/from_oneflow.html
@@ -454,14 +454,12 @@ Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdo
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading: "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/ResNet/resnet18.zip" to /workspace/.oneflow/flowvision_cache/resnet18.zip
0%| | 0.00/41.5M [00:00<?, ?B/s]
- 15%|#5 | 6.33M/41.5M [00:00<00:00, 49.3MB/s]
- 27%|##6 | 11.0M/41.5M [00:00<00:00, 37.1MB/s]
- 35%|###5 | 14.7M/41.5M [00:00<00:00, 29.7MB/s]
- 42%|####2 | 17.6M/41.5M [00:00<00:00, 25.6MB/s]
- 58%|#####7 | 24.0M/41.5M [00:00<00:00, 25.1MB/s]
- 77%|#######7 | 32.0M/41.5M [00:01<00:00, 33.3MB/s]
- 96%|#########6| 40.0M/41.5M [00:01<00:00, 39.6MB/s]
-100%|##########| 41.5M/41.5M [00:01<00:00, 35.8MB/s]
+ 19%|#9 | 7.99M/41.5M [00:00<00:00, 49.9MB/s]
+ 39%|###8 | 16.0M/41.5M [00:00<00:00, 43.7MB/s]
+ 58%|#####7 | 24.0M/41.5M [00:00<00:00, 40.7MB/s]
+ 79%|#######8 | 32.7M/41.5M [00:00<00:00, 52.9MB/s]
+ 96%|#########6| 40.0M/41.5M [00:00<00:00, 54.5MB/s]
+100%|##########| 41.5M/41.5M [00:00<00:00, 51.9MB/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 3c0797282c..f13ba10e8e 100644
--- a/docs/how_to/compile_models/from_pytorch.html
+++ b/docs/how_to/compile_models/from_pytorch.html
@@ -437,11 +437,10 @@ be unstable.</p>
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]
- 18%|#7 | 7.99M/44.7M [00:00<00:00, 65.9MB/s]
- 46%|####6 | 20.7M/44.7M [00:00<00:00, 101MB/s]
- 69%|######8 | 30.7M/44.7M [00:00<00:00, 88.3MB/s]
- 90%|########9 | 40.0M/44.7M [00:00<00:00, 81.1MB/s]
-100%|##########| 44.7M/44.7M [00:00<00:00, 76.0MB/s]
+ 18%|#7 | 7.99M/44.7M [00:00<00:00, 76.1MB/s]
+ 51%|##### | 22.7M/44.7M [00:00<00:00, 120MB/s]
+ 77%|#######7 | 34.4M/44.7M [00:00<00:00, 105MB/s]
+100%|##########| 44.7M/44.7M [00:00<00:00, 81.4MB/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 58eb75bb65..561b707145 100644
--- a/docs/how_to/compile_models/from_tensorflow.html
+++ b/docs/how_to/compile_models/from_tensorflow.html
@@ -654,7 +654,7 @@ 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 23.708 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 26.994 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 df3b16d066..fc8e5e0ad2 100644
--- a/docs/how_to/compile_models/sg_execution_times.html
+++ b/docs/how_to/compile_models/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>06:42.364</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
+<p><strong>06:56.986</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 81%" />
@@ -354,43 +354,43 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="from_tensorflow.html#sphx-glr-how-to-compile-models-from-tensorflow-py"><span class="std std-ref">Compile Tensorflow Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_tensorflow.py</span></code>)</p></td>
-<td><p>01:23.708</p></td>
+<td><p>01:26.994</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="from_darknet.html#sphx-glr-how-to-compile-models-from-darknet-py"><span class="std std-ref">Compile YOLO-V2 and YOLO-V3 in DarkNet Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_darknet.py</span></code>)</p></td>
-<td><p>01:21.154</p></td>
+<td><p>01:24.180</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>00:56.382</p></td>
+<td><p>00:59.830</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>00:37.884</p></td>
+<td><p>00:39.149</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>00:32.116</p></td>
+<td><p>00:33.675</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="from_mxnet.html#sphx-glr-how-to-compile-models-from-mxnet-py"><span class="std std-ref">Compile MXNet Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_mxnet.py</span></code>)</p></td>
-<td><p>00:31.190</p></td>
+<td><p>00:31.903</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="from_tflite.html#sphx-glr-how-to-compile-models-from-tflite-py"><span class="std std-ref">Compile TFLite Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_tflite.py</span></code>)</p></td>
-<td><p>00:28.273</p></td>
+<td><p>00:28.568</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>00:26.011</p></td>
+<td><p>00:27.021</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>00:22.871</p></td>
+<td><p>00:22.868</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>00:02.773</p></td>
+<td><p>00:02.800</p></td>
<td><p>0.0 MB</p></td>
</tr>
</tbody>
diff --git a/docs/how_to/deploy_models/deploy_model_on_adreno.html b/docs/how_to/deploy_models/deploy_model_on_adreno.html
index d1bc2ebdbc..6600860d6d 100644
--- a/docs/how_to/deploy_models/deploy_model_on_adreno.html
+++ b/docs/how_to/deploy_models/deploy_model_on_adreno.html
@@ -799,7 +799,7 @@ Top5 predictions:
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 2538.3758 2537.5783 2542.5970 2535.2333 2.4244
+ 2541.5621 2540.5103 2547.5712 2539.2064 2.7461
</pre></div>
</div>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-deploy-models-deploy-model-on-adreno-py">
diff --git a/docs/how_to/deploy_models/deploy_model_on_adreno_tvmc.html b/docs/how_to/deploy_models/deploy_model_on_adreno_tvmc.html
index 5d5fb90c1c..f590d6c4bb 100644
--- a/docs/how_to/deploy_models/deploy_model_on_adreno_tvmc.html
+++ b/docs/how_to/deploy_models/deploy_model_on_adreno_tvmc.html
@@ -443,28 +443,29 @@ to run this tutorial with a real device over rpc.</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels.h5
8192/102967424 [..............................] - ETA: 0s
- 5292032/102967424 [>.............................] - ETA: 0s
- 8470528/102967424 [=>............................] - ETA: 1s
- 12828672/102967424 [==>...........................] - ETA: 1s
+ 8093696/102967424 [=>............................] - ETA: 0s
+ 8839168/102967424 [=>............................] - ETA: 1s
16769024/102967424 [===>..........................] - ETA: 1s
- 22216704/102967424 [=====>........................] - ETA: 1s
- 25157632/102967424 [======>.......................] - ETA: 1s
- 27353088/102967424 [======>.......................] - ETA: 1s
- 33546240/102967424 [========>.....................] - ETA: 1s
- 35741696/102967424 [=========>....................] - ETA: 1s
- 41934848/102967424 [===========>..................] - ETA: 1s
- 46383104/102967424 [============>.................] - ETA: 1s
+ 23412736/102967424 [=====>........................] - ETA: 0s
+ 25239552/102967424 [======>.......................] - ETA: 0s
+ 33546240/102967424 [========>.....................] - ETA: 0s
+ 35741696/102967424 [=========>....................] - ETA: 0s
+ 40189952/102967424 [==========>...................] - ETA: 0s
+ 41934848/102967424 [===========>..................] - ETA: 0s
50323456/102967424 [=============>................] - ETA: 0s
- 50462720/102967424 [=============>................] - ETA: 1s
+ 56967168/102967424 [===============>..............] - ETA: 0s
58712064/102967424 [================>.............] - ETA: 0s
- 67100672/102967424 [==================>...........] - ETA: 0s
+ 64659456/102967424 [=================>............] - ETA: 0s
+ 65355776/102967424 [==================>...........] - ETA: 0s
+ 69083136/102967424 [===================>..........] - ETA: 0s
+ 72540160/102967424 [====================>.........] - ETA: 0s
75489280/102967424 [====================>.........] - ETA: 0s
82124800/102967424 [======================>.......] - ETA: 0s
83877888/102967424 [=======================>......] - ETA: 0s
- 90497024/102967424 [=========================>....] - ETA: 0s
92266496/102967424 [=========================>....] - ETA: 0s
98910208/102967424 [===========================>..] - ETA: 0s
-101621760/102967424 [============================>.] - ETA: 0s
+102711296/102967424 [============================>.] - ETA: 0s
+102875136/102967424 [============================>.] - ETA: 0s
102967424/102967424 [==============================] - 2s 0us/step
</pre></div>
</div>
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 9fc07368ed..6e9cf4d82b 100644
--- a/docs/how_to/deploy_models/deploy_model_on_android.html
+++ b/docs/how_to/deploy_models/deploy_model_on_android.html
@@ -667,7 +667,7 @@ to the remote android device.</p>
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 16.4630 16.3437 17.2113 15.9040 0.4360
+ 16.4419 16.2021 17.6050 16.1123 0.5233
</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 87e142cabb..e07772c620 100644
--- a/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
+++ b/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
@@ -459,31 +459,31 @@ be unstable.</p>
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]
- 5%|4 | 7.99M/170M [00:00<00:04, 41.5MB/s]
- 9%|9 | 16.0M/170M [00:00<00:05, 30.8MB/s]
- 14%|#4 | 24.0M/170M [00:00<00:03, 38.3MB/s]
- 19%|#8 | 32.0M/170M [00:00<00:03, 40.1MB/s]
- 24%|##3 | 40.0M/170M [00:01<00:03, 43.6MB/s]
- 28%|##8 | 48.0M/170M [00:01<00:02, 46.3MB/s]
- 33%|###2 | 56.0M/170M [00:01<00:02, 52.1MB/s]
- 38%|###7 | 64.0M/170M [00:01<00:02, 53.0MB/s]
- 42%|####2 | 72.0M/170M [00:01<00:01, 54.2MB/s]
- 47%|####7 | 80.0M/170M [00:01<00:01, 55.2MB/s]
- 52%|#####1 | 88.0M/170M [00:01<00:01, 54.4MB/s]
- 56%|#####5 | 94.3M/170M [00:02<00:01, 56.3MB/s]
- 59%|#####8 | 99.8M/170M [00:02<00:01, 55.4MB/s]
- 62%|######1 | 105M/170M [00:02<00:01, 47.0MB/s]
- 66%|######5 | 112M/170M [00:02<00:01, 45.4MB/s]
- 71%|####### | 120M/170M [00:02<00:01, 43.6MB/s]
- 75%|#######5 | 128M/170M [00:02<00:00, 45.4MB/s]
- 80%|######## | 136M/170M [00:03<00:00, 44.6MB/s]
- 85%|########4 | 144M/170M [00:03<00:00, 42.9MB/s]
- 88%|########8 | 150M/170M [00:03<00:00, 40.3MB/s]
- 91%|######### | 154M/170M [00:03<00:00, 37.7MB/s]
- 93%|#########3| 158M/170M [00:03<00:00, 38.3MB/s]
- 95%|#########5| 162M/170M [00:03<00:00, 38.7MB/s]
- 99%|#########8| 168M/170M [00:03<00:00, 41.4MB/s]
-100%|##########| 170M/170M [00:03<00:00, 45.0MB/s]
+ 3%|2 | 4.85M/170M [00:00<00:03, 50.9MB/s]
+ 6%|5 | 9.70M/170M [00:00<00:06, 28.0MB/s]
+ 9%|9 | 16.0M/170M [00:00<00:04, 38.9MB/s]
+ 14%|#4 | 24.0M/170M [00:00<00:03, 47.7MB/s]
+ 19%|#8 | 31.6M/170M [00:00<00:02, 56.7MB/s]
+ 22%|##2 | 37.5M/170M [00:00<00:02, 55.6MB/s]
+ 27%|##6 | 45.2M/170M [00:00<00:02, 62.4MB/s]
+ 30%|### | 51.5M/170M [00:01<00:02, 46.6MB/s]
+ 33%|###3 | 56.6M/170M [00:01<00:02, 42.1MB/s]
+ 38%|###7 | 64.0M/170M [00:01<00:02, 44.0MB/s]
+ 45%|####5 | 77.0M/170M [00:01<00:01, 64.1MB/s]
+ 50%|####9 | 84.1M/170M [00:01<00:01, 61.5MB/s]
+ 53%|#####3 | 90.6M/170M [00:01<00:01, 50.9MB/s]
+ 57%|#####6 | 96.2M/170M [00:02<00:02, 34.6MB/s]
+ 60%|###### | 102M/170M [00:02<00:01, 37.6MB/s]
+ 63%|######3 | 107M/170M [00:02<00:01, 40.0MB/s]
+ 66%|######6 | 112M/170M [00:02<00:01, 42.5MB/s]
+ 71%|####### | 120M/170M [00:02<00:01, 48.8MB/s]
+ 74%|#######4 | 126M/170M [00:02<00:00, 47.5MB/s]
+ 80%|######## | 136M/170M [00:03<00:00, 51.7MB/s]
+ 86%|########5 | 146M/170M [00:03<00:00, 59.3MB/s]
+ 89%|########9 | 152M/170M [00:03<00:00, 59.4MB/s]
+ 94%|#########4| 160M/170M [00:03<00:00, 58.0MB/s]
+ 99%|#########8| 168M/170M [00:03<00:00, 57.2MB/s]
+100%|##########| 170M/170M [00:03<00:00, 50.1MB/s]
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/torch/nn/functional.py:3897: 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)
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/torchvision/models/detection/anchor_utils.py:124: 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=& [...]
@@ -581,7 +581,7 @@ 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> ( 3 minutes 36.224 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes 48.383 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 05f2262a3d..8f22edac4b 100644
--- a/docs/how_to/deploy_models/deploy_prequantized.html
+++ b/docs/how_to/deploy_models/deploy_prequantized.html
@@ -500,9 +500,8 @@ training. Other models require a full post training calibration.</p>
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]
- 57%|#####6 | 7.68M/13.6M [00:00<00:00, 51.2MB/s]
- 93%|#########2| 12.6M/13.6M [00:00<00:00, 38.0MB/s]
-100%|##########| 13.6M/13.6M [00:00<00:00, 42.3MB/s]
+ 59%|#####8 | 7.99M/13.6M [00:00<00:00, 57.8MB/s]
+100%|##########| 13.6M/13.6M [00:00<00:00, 66.1MB/s]
</pre></div>
</div>
</div>
@@ -593,7 +592,7 @@ 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)
- 90.1348 90.0830 94.1666 89.8210 0.4250
+ 90.3151 90.2691 90.8183 90.0883 0.1735
</pre></div>
</div>
<div class="admonition note">
@@ -632,7 +631,7 @@ 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 16.961 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 20.333 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 50cd519dfb..900520eec8 100644
--- a/docs/how_to/deploy_models/deploy_prequantized_tflite.html
+++ b/docs/how_to/deploy_models/deploy_prequantized_tflite.html
@@ -585,7 +585,7 @@ 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)
- 119.6617 119.8028 120.9767 117.9153 0.6340
+ 120.5423 120.4685 125.6759 119.5865 0.6641
</pre></div>
</div>
<div class="admonition note">
@@ -613,7 +613,7 @@ 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> ( 2 minutes 29.156 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 2 minutes 31.977 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 6150d8ed98..54a3dcc5b9 100644
--- a/docs/how_to/deploy_models/deploy_quantized.html
+++ b/docs/how_to/deploy_models/deploy_quantized.html
@@ -526,7 +526,7 @@ 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 43.517 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 42.223 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 21817bcc40..93a3a3755a 100644
--- a/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
+++ b/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
@@ -468,24 +468,24 @@ to your device.</p>
Downloading /workspace/.mxnet/models/ssd_512_resnet50_v1_voc-9c8b225a.zip from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/ssd_512_resnet50_v1_voc-9c8b225a.zip...
0%| | 0/132723 [00:00<?, ?KB/s]
- 5%|4 | 6315/132723 [00:00<00:02, 63139.97KB/s]
- 11%|#1 | 14790/132723 [00:00<00:01, 75844.69KB/s]
- 17%|#6 | 22375/132723 [00:00<00:01, 68936.78KB/s]
- 22%|##2 | 29470/132723 [00:00<00:01, 60860.59KB/s]
- 27%|##6 | 35699/132723 [00:00<00:01, 50204.26KB/s]
- 33%|###3 | 44062/132723 [00:00<00:01, 59246.80KB/s]
- 38%|###7 | 50407/132723 [00:00<00:01, 58438.31KB/s]
- 44%|####4 | 58798/132723 [00:00<00:01, 65518.96KB/s]
- 50%|##### | 66934/132723 [00:01<00:00, 70021.38KB/s]
- 57%|#####6 | 75433/132723 [00:01<00:00, 74341.81KB/s]
- 63%|######2 | 83067/132723 [00:01<00:00, 66702.48KB/s]
- 69%|######8 | 91466/132723 [00:01<00:00, 71382.30KB/s]
- 75%|#######5 | 99853/132723 [00:01<00:00, 74863.44KB/s]
- 82%|########1 | 108184/132723 [00:01<00:00, 77268.05KB/s]
- 87%|########7 | 116066/132723 [00:01<00:00, 76981.59KB/s]
- 94%|#########3| 124530/132723 [00:01<00:00, 79203.96KB/s]
-100%|#########9| 132534/132723 [00:01<00:00, 72683.40KB/s]
-100%|##########| 132723/132723 [00:01<00:00, 68954.83KB/s]
+ 3%|2 | 3682/132723 [00:00<00:03, 36817.58KB/s]
+ 8%|8 | 10682/132723 [00:00<00:02, 56330.46KB/s]
+ 14%|#3 | 18572/132723 [00:00<00:01, 66632.38KB/s]
+ 20%|#9 | 26510/132723 [00:00<00:01, 71660.76KB/s]
+ 25%|##5 | 33677/132723 [00:00<00:01, 67776.40KB/s]
+ 31%|###1 | 41468/132723 [00:00<00:01, 71094.61KB/s]
+ 37%|###7 | 49280/132723 [00:00<00:01, 71541.97KB/s]
+ 43%|####3 | 57112/132723 [00:00<00:01, 73622.94KB/s]
+ 49%|####9 | 65102/132723 [00:00<00:00, 75536.45KB/s]
+ 55%|#####5 | 73005/132723 [00:01<00:00, 76594.96KB/s]
+ 61%|######1 | 80985/132723 [00:01<00:00, 77562.88KB/s]
+ 67%|######7 | 89008/132723 [00:01<00:00, 78363.15KB/s]
+ 73%|#######3 | 97044/132723 [00:01<00:00, 78962.42KB/s]
+ 79%|#######9 | 104947/132723 [00:01<00:00, 70184.36KB/s]
+ 85%|########5 | 112902/132723 [00:01<00:00, 72770.75KB/s]
+ 91%|######### | 120572/132723 [00:01<00:00, 73881.25KB/s]
+ 97%|#########6| 128556/132723 [00:01<00:00, 75593.83KB/s]
+100%|##########| 132723/132723 [00:01<00:00, 73010.88KB/s]
</pre></div>
</div>
<p>Create TVM runtime and do inference
@@ -524,7 +524,7 @@ 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> ( 3 minutes 45.898 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> ( 3 minutes 56.461 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 76d9fd9cd0..73c1ef0b83 100644
--- a/docs/how_to/deploy_models/sg_execution_times.html
+++ b/docs/how_to/deploy_models/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>16:16.748</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
+<p><strong>16:51.562</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 86%" />
@@ -354,43 +354,43 @@
</colgroup>
<tbody>
<tr class="row-odd"><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>03:45.898</p></td>
+<td><p>03:56.461</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><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>03:36.224</p></td>
+<td><p>03:48.383</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>02:29.156</p></td>
+<td><p>02:31.977</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>01:43.517</p></td>
+<td><p>01:42.223</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>01:16.961</p></td>
+<td><p>01:20.333</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="deploy_model_on_adreno.html#sphx-glr-how-to-deploy-models-deploy-model-on-adreno-py"><span class="std std-ref">Deploy the Pretrained Model on Adreno™</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_adreno.py</span></code>)</p></td>
-<td><p>00:54.886</p></td>
+<td><p>00:56.111</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="deploy_model_on_adreno_tvmc.html#sphx-glr-how-to-deploy-models-deploy-model-on-adreno-tvmc-py"><span class="std std-ref">Deploy the Pretrained Model on Adreno™ with tvmc Interface</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_adreno_tvmc.py</span></code>)</p></td>
-<td><p>00:51.589</p></td>
+<td><p>00:53.265</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>00:42.608</p></td>
+<td><p>00:44.676</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>00:28.162</p></td>
+<td><p>00:29.423</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>00:27.741</p></td>
+<td><p>00:28.704</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 5ecd581d17..602806a045 100644
--- a/docs/how_to/extend_tvm/bring_your_own_datatypes.html
+++ b/docs/how_to/extend_tvm/bring_your_own_datatypes.html
@@ -624,7 +624,7 @@ 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.zip80284e7a-7990-4b90-baf4-58da2c8595f7 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.zipe820fe04-b70a-4359-a47c-6cf9420def47 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
</pre></div>
</div>
<p>It’s easy to execute MobileNet with native TVM:</p>
diff --git a/docs/how_to/extend_tvm/sg_execution_times.html b/docs/how_to/extend_tvm/sg_execution_times.html
index eb7b34ea9c..0024d8d997 100644
--- a/docs/how_to/extend_tvm/sg_execution_times.html
+++ b/docs/how_to/extend_tvm/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>00:54.320</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
+<p><strong>00:57.036</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -354,19 +354,19 @@
</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>00:50.530</p></td>
+<td><p>00:53.001</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>00:02.709</p></td>
+<td><p>00:02.890</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>00:01.074</p></td>
+<td><p>00:01.137</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 c6e1dd837a..0e3663d77a 100644
--- a/docs/how_to/extend_tvm/use_pass_instrument.html
+++ b/docs/how_to/extend_tvm/use_pass_instrument.html
@@ -531,10 +531,10 @@ 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: 22219us [22219us] (48.74%; 48.74%)
-FoldScaleAxis: 23364us [7us] (51.26%; 51.26%)
- FoldConstant: 23357us [1666us] (51.24%; 99.97%)
- InferType: 21691us [21691us] (47.59%; 92.87%)
+InferType: 22452us [22452us] (47.72%; 47.72%)
+FoldScaleAxis: 24594us [9us] (52.28%; 52.28%)
+ FoldConstant: 24585us [1726us] (52.26%; 99.96%)
+ InferType: 22859us [22859us] (48.59%; 92.98%)
</pre></div>
</div>
</div>
@@ -556,10 +556,10 @@ 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: 21601us [21601us] (48.24%; 48.24%)
-FoldScaleAxis: 23175us [5us] (51.76%; 51.76%)
- FoldConstant: 23170us [1708us] (51.75%; 99.98%)
- InferType: 21463us [21463us] (47.93%; 92.63%)
+InferType: 22534us [22534us] (48.18%; 48.18%)
+FoldScaleAxis: 24235us [8us] (51.82%; 51.82%)
+ FoldConstant: 24227us [1791us] (51.80%; 99.97%)
+ InferType: 22436us [22436us] (47.97%; 92.61%)
</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 fd5ba3f3a6..2a2c341d42 100644
--- a/docs/how_to/optimize_operators/opt_conv_cuda.html
+++ b/docs/how_to/optimize_operators/opt_conv_cuda.html
@@ -580,7 +580,7 @@ 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: 54.182239 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Convolution: 54.190113 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 c422e12fd7..27a7ff3d3d 100644
--- a/docs/how_to/optimize_operators/opt_conv_tensorcore.html
+++ b/docs/how_to/optimize_operators/opt_conv_tensorcore.html
@@ -862,7 +862,7 @@ 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: 13.373636 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>conv2d with tensor core: 13.361943 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 bed06a3aea..a287cd8b30 100644
--- a/docs/how_to/optimize_operators/opt_gemm.html
+++ b/docs/how_to/optimize_operators/opt_gemm.html
@@ -477,8 +477,8 @@ 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: 0.018695
-Baseline: 3.251712
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Numpy running time: 0.019399
+Baseline: 3.484613
</pre></div>
</div>
<p>In TVM, we can always inspect lower level IR to debug or optimize our schedule.
@@ -537,7 +537,7 @@ 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: 0.296724
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt1: 0.331902
</pre></div>
</div>
<p>Here is the generated IR after blocking.</p>
@@ -594,7 +594,7 @@ 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: 0.342606
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt2: 0.350420
</pre></div>
</div>
<p>Here is the generated IR after vectorization.</p>
@@ -649,7 +649,7 @@ 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: 0.114652
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt3: 0.122769
</pre></div>
</div>
<p>Here is the generated IR after loop permutation.</p>
@@ -726,7 +726,7 @@ 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: 0.110320
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt4: 0.109488
</pre></div>
</div>
<p>Here is the generated IR after array packing.</p>
@@ -804,7 +804,7 @@ 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: 0.111773
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt5: 0.112409
</pre></div>
</div>
<p>Here is the generated IR after blocking.</p>
@@ -884,7 +884,7 @@ class Module:
<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: 0.147087
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt6: 0.148717
</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 9396572ceb..846da86b76 100644
--- a/docs/how_to/optimize_operators/sg_execution_times.html
+++ b/docs/how_to/optimize_operators/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>00:34.823</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
+<p><strong>00:36.149</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -354,15 +354,15 @@
</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>00:31.905</p></td>
+<td><p>00:33.396</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>00:01.706</p></td>
+<td><p>00:01.610</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>00:01.212</p></td>
+<td><p>00:01.143</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 fe89577272..81b539f648 100644
--- a/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
+++ b/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>09:58.198</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
+<p><strong>10:20.053</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 85%" />
@@ -354,27 +354,27 @@
</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>06:09.023</p></td>
+<td><p>06:24.225</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>01:41.602</p></td>
+<td><p>01:44.830</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>01:07.297</p></td>
+<td><p>01:09.353</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>00:32.629</p></td>
+<td><p>00:32.826</p></td>
<td><p>0.0 MB</p></td>
</tr>
<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>00:14.156</p></td>
+<td><p>00:14.692</p></td>
<td><p>0.0 MB</p></td>
</tr>
<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>00:13.491</p></td>
+<td><p>00:14.127</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 cf4ba503d2..5476bf1d61 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
@@ -512,183 +512,840 @@ class Module:
T.func_attr({"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True})
blockIdx_x = T.launch_thread("blockIdx.x", 32)
conv2d_nchw = T.allocate([14], "float32", "local")
- pad_temp_shared = T.allocate([2592], "float32", "shared")
- kernel_shared = T.allocate([4608], "float32", "shared")
+ pad_temp_shared = T.allocate([2016], "float32", "shared")
+ kernel_shared = T.allocate([1536], "float32", "shared")
threadIdx_x = T.launch_thread("threadIdx.x", 56)
- conv2d_nchw_1 = T.Buffer((14,), data=conv2d_nchw, scope="local", align=32)
+ conv2d_nchw_1 = T.Buffer((49,), data=conv2d_nchw, scope="local", align=16)
conv2d_nchw_1[0] = T.float32(0)
- conv2d_nchw_1[1] = T.float32(0)
- conv2d_nchw_1[2] = T.float32(0)
- conv2d_nchw_1[3] = T.float32(0)
- conv2d_nchw_1[4] = T.float32(0)
- conv2d_nchw_1[5] = T.float32(0)
- conv2d_nchw_1[6] = T.float32(0)
conv2d_nchw_1[7] = T.float32(0)
+ conv2d_nchw_1[1] = T.float32(0)
conv2d_nchw_1[8] = T.float32(0)
+ conv2d_nchw_1[2] = T.float32(0)
conv2d_nchw_1[9] = T.float32(0)
+ conv2d_nchw_1[3] = T.float32(0)
conv2d_nchw_1[10] = T.float32(0)
+ conv2d_nchw_1[4] = T.float32(0)
conv2d_nchw_1[11] = T.float32(0)
+ conv2d_nchw_1[5] = T.float32(0)
conv2d_nchw_1[12] = T.float32(0)
+ conv2d_nchw_1[6] = T.float32(0)
conv2d_nchw_1[13] = T.float32(0)
- for rc_outer_outer in range(16):
- cse_var_1: T.int32 = rc_outer_outer * 1568
+ for rc_outer_outer, rx_outer_outer in T.grid(16, 3):
+ cse_var_2: T.int32 = rc_outer_outer * 1568
+ cse_var_1: T.int32 = rc_outer_outer * 288
threadIdx_x_1 = T.env_thread("threadIdx.x")
- pad_temp_shared_1 = T.Buffer((2592,), data=pad_temp_shared, scope="shared")
+ pad_temp_shared_1 = T.Buffer((2016,), data=pad_temp_shared, scope="shared")
data_1 = T.Buffer((25088,), data=data.data)
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(9 <= threadIdx_x_1 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + threadIdx_x_1 // 9 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 56] = T.if_then_else(9 <= (threadIdx_x_1 + 56) % 81 and (threadIdx_x_1 + 56) % 81 < 72 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 56) // 81 * 49 + (threadIdx_x_1 + 56) % 81 // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 112] = T.if_then_else(9 <= (threadIdx_x_1 + 31) % 81 and (threadIdx_x_1 + 31) % 81 < 72 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 112) // 81 * 49 + (threadIdx_x_1 + 31) % 81 // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 168] = T.if_then_else(3 <= threadIdx_x_1 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 168) // 81 * 49 + (threadIdx_x_1 + 6) // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 224] = T.if_then_else(9 <= (threadIdx_x_1 + 62) % 81 and (threadIdx_x_1 + 62) % 81 < 72 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 224) // 81 * 49 + (threadIdx_x_1 + 62) % 81 // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 280] = T.if_then_else(9 <= (threadIdx_x_1 + 37) % 81 and (threadIdx_x_1 + 37) % 81 < 72 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 280) // 81 * 49 + (threadIdx_x_1 + 37) % 81 // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 336] = T.if_then_else(1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 336) // 81 * 49 + (threadIdx_x_1 + 12) // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 392] = T.if_then_else(9 <= (threadIdx_x_1 + 68) % 81 and (threadIdx_x_1 + 68) % 81 < 72 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 392) // 81 * 49 + (threadIdx_x_1 + 68) % 81 // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 448] = T.if_then_else(9 <= (threadIdx_x_1 + 43) % 81 and (threadIdx_x_1 + 43) % 81 < 72 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 448) // 81 * 49 + (threadIdx_x_1 + 43) % 81 // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 504] = T.if_then_else(threadIdx_x_1 < 54 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 504) // 81 * 49 + threadIdx_x_1 // 9 * 7 + threadIdx_x_1 % 9 + 6], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(7 <= threadIdx_x_1 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + threadIdx_x_1 + rx_outer_outer - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 560] = T.if_then_else(9 <= (threadIdx_x_1 + 74) % 81 and (threadIdx_x_1 + 74) % 81 < 72 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 560) // 81 * 49 + (threadIdx_x_1 + 74) % 81 // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 56] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 8) % 9 and (threadIdx_x_1 // 7 + 8) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 56) // 63 * 49 + (threadIdx_x_1 // 7 + 8) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 616] = T.if_then_else(9 <= (threadIdx_x_1 + 49) % 81 and (threadIdx_x_1 + 49) % 81 < 72 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 616) // 81 * 49 + (threadIdx_x_1 + 49) % 81 // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 112] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 7) % 9 and (threadIdx_x_1 // 7 + 7) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 112) // 63 * 49 + (threadIdx_x_1 // 7 + 7) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 672] = T.if_then_else(threadIdx_x_1 < 48 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 672) // 81 * 49 + (threadIdx_x_1 + 24) // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 168] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 6) % 9 and (threadIdx_x_1 // 7 + 6) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 168) // 63 * 49 + (threadIdx_x_1 // 7 + 6) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 728] = T.if_then_else(9 <= (threadIdx_x_1 + 80) % 81 and (threadIdx_x_1 + 80) % 81 < 72 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 728) // 81 * 49 + (threadIdx_x_1 + 80) % 81 // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 224] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 5) % 9 and (threadIdx_x_1 // 7 + 5) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 224) // 63 * 49 + (threadIdx_x_1 // 7 + 5) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 784] = T.if_then_else(9 <= (threadIdx_x_1 + 55) % 81 and (threadIdx_x_1 + 55) % 81 < 72 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 784) // 81 * 49 + (threadIdx_x_1 + 55) % 81 // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 280] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 4) % 9 and (threadIdx_x_1 // 7 + 4) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 280) // 63 * 49 + (threadIdx_x_1 // 7 + 4) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 840] = T.if_then_else(9 <= (threadIdx_x_1 + 30) % 81 and (threadIdx_x_1 + 30) % 81 < 72 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 840) // 81 * 49 + (threadIdx_x_1 + 30) % 81 // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 336] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 3) % 9 and (threadIdx_x_1 // 7 + 3) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 336) // 63 * 49 + (threadIdx_x_1 // 7 + 3) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 896] = T.if_then_else(4 <= threadIdx_x_1 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 896) // 81 * 49 + (threadIdx_x_1 + 5) // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 392] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 2) % 9 and (threadIdx_x_1 // 7 + 2) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 392) // 63 * 49 + (threadIdx_x_1 // 7 + 2) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 952] = T.if_then_else(9 <= (threadIdx_x_1 + 61) % 81 and (threadIdx_x_1 + 61) % 81 < 72 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 952) // 81 * 49 + (threadIdx_x_1 + 61) % 81 // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 448] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 448) // 63 * 49 + threadIdx_x_1 + rx_outer_outer - 1], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1008] = T.if_then_else(1 <= (threadIdx_x_1 // 9 + 4) % 9 and (threadIdx_x_1 + 36) % 81 < 72 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1008) // 81 * 49 + (threadIdx_x_1 // 9 + 4) % 9 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 504] = T.if_then_else(7 <= threadIdx_x_1 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + threadIdx_x_1 + rx_outer_outer + 384], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1064] = T.if_then_else(1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1064) // 81 * 49 + (threadIdx_x_1 + 11) // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 560] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 8) % 9 and (threadIdx_x_1 // 7 + 8) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 560) // 63 * 49 + (threadIdx_x_1 // 7 + 8) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1120] = T.if_then_else(9 <= (threadIdx_x_1 + 67) % 81 and (threadIdx_x_1 + 67) % 81 < 72 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1120) // 81 * 49 + (threadIdx_x_1 + 67) % 81 // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 616] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 7) % 9 and (threadIdx_x_1 // 7 + 7) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 616) // 63 * 49 + (threadIdx_x_1 // 7 + 7) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1176] = T.if_then_else(9 <= (threadIdx_x_1 + 42) % 81 and (threadIdx_x_1 + 42) % 81 < 72 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1176) // 81 * 49 + (threadIdx_x_1 + 42) % 81 // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 672] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 6) % 9 and (threadIdx_x_1 // 7 + 6) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 672) // 63 * 49 + (threadIdx_x_1 // 7 + 6) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1232] = T.if_then_else(threadIdx_x_1 < 55 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1232) // 81 * 49 + (threadIdx_x_1 + 17) // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 728] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 5) % 9 and (threadIdx_x_1 // 7 + 5) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 728) // 63 * 49 + (threadIdx_x_1 // 7 + 5) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1288] = T.if_then_else(9 <= (threadIdx_x_1 + 73) % 81 and (threadIdx_x_1 + 73) % 81 < 72 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1288) // 81 * 49 + (threadIdx_x_1 + 73) % 81 // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 784] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 4) % 9 and (threadIdx_x_1 // 7 + 4) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 784) // 63 * 49 + (threadIdx_x_1 // 7 + 4) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1344] = T.if_then_else(9 <= (threadIdx_x_1 + 48) % 81 and (threadIdx_x_1 + 48) % 81 < 72 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1344) // 81 * 49 + (threadIdx_x_1 + 48) % 81 // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 840] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 3) % 9 and (threadIdx_x_1 // 7 + 3) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 840) // 63 * 49 + (threadIdx_x_1 // 7 + 3) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1400] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1400) // 81 * 49 + (threadIdx_x_1 + 23) // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 896] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 2) % 9 and (threadIdx_x_1 // 7 + 2) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 896) // 63 * 49 + (threadIdx_x_1 // 7 + 2) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1456] = T.if_then_else(9 <= (threadIdx_x_1 + 79) % 81 and (threadIdx_x_1 + 79) % 81 < 72 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1456) // 81 * 49 + (threadIdx_x_1 + 79) % 81 // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 952] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 952) // 63 * 49 + threadIdx_x_1 + rx_outer_outer - 1], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1512] = T.if_then_else(1 <= (threadIdx_x_1 // 9 + 6) % 9 and (threadIdx_x_1 + 54) % 81 < 72 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1512) // 81 * 49 + (threadIdx_x_1 // 9 + 6) % 9 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1008] = T.if_then_else(7 <= threadIdx_x_1 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + threadIdx_x_1 + rx_outer_outer + 776], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1568] = T.if_then_else(9 <= (threadIdx_x_1 + 29) % 81 and (threadIdx_x_1 + 29) % 81 < 72 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1568) // 81 * 49 + (threadIdx_x_1 + 29) % 81 // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1064] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 8) % 9 and (threadIdx_x_1 // 7 + 8) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1064) // 63 * 49 + (threadIdx_x_1 // 7 + 8) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1624] = T.if_then_else(5 <= threadIdx_x_1 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1624) // 81 * 49 + (threadIdx_x_1 + 4) // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1120] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 7) % 9 and (threadIdx_x_1 // 7 + 7) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1120) // 63 * 49 + (threadIdx_x_1 // 7 + 7) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1680] = T.if_then_else(9 <= (threadIdx_x_1 + 60) % 81 and (threadIdx_x_1 + 60) % 81 < 72 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1680) // 81 * 49 + (threadIdx_x_1 + 60) % 81 // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1176] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 6) % 9 and (threadIdx_x_1 // 7 + 6) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1176) // 63 * 49 + (threadIdx_x_1 // 7 + 6) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1736] = T.if_then_else(9 <= (threadIdx_x_1 + 35) % 81 and (threadIdx_x_1 + 35) % 81 < 72 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1736) // 81 * 49 + (threadIdx_x_1 + 35) % 81 // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1232] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 5) % 9 and (threadIdx_x_1 // 7 + 5) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1232) // 63 * 49 + (threadIdx_x_1 // 7 + 5) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1792] = T.if_then_else(1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1792) // 81 * 49 + (threadIdx_x_1 + 10) // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1288] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 4) % 9 and (threadIdx_x_1 // 7 + 4) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1288) // 63 * 49 + (threadIdx_x_1 // 7 + 4) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1848] = T.if_then_else(9 <= (threadIdx_x_1 + 66) % 81 and (threadIdx_x_1 + 66) % 81 < 72 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1848) // 81 * 49 + (threadIdx_x_1 + 66) % 81 // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1344] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 3) % 9 and (threadIdx_x_1 // 7 + 3) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1344) // 63 * 49 + (threadIdx_x_1 // 7 + 3) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1904] = T.if_then_else(9 <= (threadIdx_x_1 + 41) % 81 and (threadIdx_x_1 + 41) % 81 < 72 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1904) // 81 * 49 + (threadIdx_x_1 + 41) % 81 // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1400] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 2) % 9 and (threadIdx_x_1 // 7 + 2) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1400) // 63 * 49 + (threadIdx_x_1 // 7 + 2) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 1960] = T.if_then_else(1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 1960) // 81 * 49 + (threadIdx_x_1 + 16) // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1456] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1456) // 63 * 49 + threadIdx_x_1 + rx_outer_outer - 1], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2016] = T.if_then_else(1 <= (threadIdx_x_1 // 9 + 8) % 9 and (threadIdx_x_1 + 72) % 81 < 72 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2016) // 81 * 49 + (threadIdx_x_1 // 9 + 8) % 9 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1512] = T.if_then_else(7 <= threadIdx_x_1 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + threadIdx_x_1 + rx_outer_outer + 1168], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2072] = T.if_then_else(9 <= (threadIdx_x_1 + 47) % 81 and (threadIdx_x_1 + 47) % 81 < 72 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2072) // 81 * 49 + (threadIdx_x_1 + 47) % 81 // 9 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1568] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 8) % 9 and (threadIdx_x_1 // 7 + 8) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1568) // 63 * 49 + (threadIdx_x_1 // 7 + 8) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2128] = T.if_then_else(threadIdx_x_1 < 50 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2128) // 81 * 49 + (threadIdx_x_1 + 22) // 9 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1624] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 7) % 9 and (threadIdx_x_1 // 7 + 7) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1624) // 63 * 49 + (threadIdx_x_1 // 7 + 7) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2184] = T.if_then_else(9 <= (threadIdx_x_1 + 78) % 81 and (threadIdx_x_1 + 78) % 81 < 72 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2184) // 81 * 49 + (threadIdx_x_1 + 78) % 81 // 9 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1680] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 6) % 9 and (threadIdx_x_1 // 7 + 6) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1680) // 63 * 49 + (threadIdx_x_1 // 7 + 6) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2240] = T.if_then_else(9 <= (threadIdx_x_1 + 53) % 81 and (threadIdx_x_1 + 53) % 81 < 72 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2240) // 81 * 49 + (threadIdx_x_1 + 53) % 81 // 9 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1736] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 5) % 9 and (threadIdx_x_1 // 7 + 5) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1736) // 63 * 49 + (threadIdx_x_1 // 7 + 5) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2296] = T.if_then_else(9 <= (threadIdx_x_1 + 28) % 81 and (threadIdx_x_1 + 28) % 81 < 72 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2296) // 81 * 49 + (threadIdx_x_1 + 28) % 81 // 9 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1792] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 4) % 9 and (threadIdx_x_1 // 7 + 4) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1792) // 63 * 49 + (threadIdx_x_1 // 7 + 4) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2352] = T.if_then_else(6 <= threadIdx_x_1 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2352) // 81 * 49 + (threadIdx_x_1 + 3) // 9 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1848] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 3) % 9 and (threadIdx_x_1 // 7 + 3) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1848) // 63 * 49 + (threadIdx_x_1 // 7 + 3) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2408] = T.if_then_else(9 <= (threadIdx_x_1 + 59) % 81 and (threadIdx_x_1 + 59) % 81 < 72 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2408) // 81 * 49 + (threadIdx_x_1 + 59) % 81 // 9 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
+ pad_temp_shared_1[threadIdx_x_1 + 1904] = T.if_then_else(1 <= (threadIdx_x_1 // 7 + 2) % 9 and (threadIdx_x_1 // 7 + 2) % 9 < 8 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1904) // 63 * 49 + (threadIdx_x_1 // 7 + 2) % 9 * 7 + rx_outer_outer + threadIdx_x_1 % 7 - 8], T.float32(0))
with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2464] = T.if_then_else(9 <= (threadIdx_x_1 + 34) % 81 and (threadIdx_x_1 + 34) % 81 < 72 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2464) // 81 * 49 + (threadIdx_x_1 + 34) % 81 // 9 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- pad_temp_shared_1[threadIdx_x_1 + 2520] = T.if_then_else(1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2520) // 81 * 49 + threadIdx_x_1 // 9 * 7 + threadIdx_x_1 % 9 - 1], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 56):
- if T.likely(threadIdx_x_1 < 16):
- pad_temp_shared_1[threadIdx_x_1 + 2576] = T.if_then_else(threadIdx_x_1 < 7 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_1 + (threadIdx_x_1 + 2576) // 81 * 49 + (threadIdx_x_1 + 65) // 9 * 7 + threadIdx_x_1 - 6], T.float32(0))
- kernel_shared_1 = T.Buffer((4608,), data=kernel_shared, scope="shared")
- for ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer in range(21):
- threadIdx_x_2 = T.launch_thread("threadIdx.x", 56)
- kernel_1 = T.Buffer((2359296,), data=kernel.data)
- if T.likely(ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8 < 144):
- kernel_shared_1[ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4] = kernel_1[blockIdx_x * 73728 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8) // 9 * 4608 + rc_outer_outer * 288 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4) % 288]
- if T.likely(ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8 < 144):
- kernel_shared_1[ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 1] = kernel_1[blockIdx_x * 73728 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8) // 9 * 4608 + rc_outer_outer * 288 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 1) % 288]
- if T.likely(ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8 < 144):
- kernel_shared_1[ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 2] = kernel_1[blockIdx_x * 73728 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8) // 9 * 4608 + rc_outer_outer * 288 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 2) % 288]
- if T.likely(ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8 < 144):
- kernel_shared_1[ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 3] = kernel_1[blockIdx_x * 73728 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7 + threadIdx_x_2 // 8) // 9 * 4608 + rc_outer_outer * 288 + (ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224 + threadIdx_x_2 * 4 + 3) % 288]
- for rc_outer_inner, ry_outer_inner, rc_inner in T.grid(8, 3, 4):
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 9] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 18] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 27] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 36] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 45] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 54] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3]
- conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 9] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 18] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 27] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 36] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 45] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 54] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 288]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 1] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 10] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 19] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 28] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 37] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 46] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 55] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 1] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 10] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 19] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 28] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 37] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 46] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 55] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 289]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 2] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 11] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 20] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 29] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 38] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 47] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 56] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 2] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 11] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 20] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 29] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 38] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 47] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 324 + rc_inner * 81 + ry_outer_inner * 9 + threadIdx_x % 7 + 56] * kernel_shared_1[threadIdx_x // 7 * 576 + rc_outer_inner * 36 + rc_inner * 9 + ry_outer_inner * 3 + 290]
- for i1_inner, i2_inner in T.grid(2, 7):
+ pad_temp_shared_1[threadIdx_x_1 + 1960] = T.if_then_else(threadIdx_x_1 < 49 and 1 <= rx_outer_outer + threadIdx_x_1 % 7 and rx_outer_outer + threadIdx_x_1 % 7 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 1960) // 63 * 49 + threadIdx_x_1 + rx_outer_outer - 1], T.float32(0))
+ threadIdx_x_2 = T.env_thread("threadIdx.x")
+ kernel_shared_1 = T.Buffer((1536,), data=kernel_shared, scope="shared")
+ kernel_1 = T.Buffer((2359296,), data=kernel.data)
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2] = kernel_1[blockIdx_x * 73728 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 56] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 56) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 112] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 112) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 48]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 168] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 168) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 72) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 224] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 224) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 96]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 280] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 280) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 88) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 336] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 336) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 392] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 392) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 24]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 448] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 448) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 504] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 504) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 72]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 560] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 560) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 616] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 616) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 120]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 672] = kernel_1[blockIdx_x * 73728 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 32256]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 728] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 728) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 784] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 784) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 48]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 840] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 840) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 72) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 896] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 896) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 96]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 952] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 952) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 88) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1008] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1008) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1064] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1064) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 24]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1120] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1120) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1176] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1176) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 72]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1232] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1232) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1288] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1288) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 120]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1344] = kernel_1[blockIdx_x * 73728 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 64512]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1400] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1400) // 96 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 96 * 3 + rx_outer_outer]
+ with T.launch_thread(threadIdx_x_2, 56):
+ kernel_shared_1[threadIdx_x_2 + 1456] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1456) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 48]
+ with T.launch_thread(threadIdx_x_2, 56):
+ if T.likely(threadIdx_x_2 < 24):
+ kernel_shared_1[threadIdx_x_2 + 1512] = kernel_1[blockIdx_x * 73728 + (threadIdx_x_2 + 1512) // 96 * 4608 + cse_var_1 + threadIdx_x_2 * 3 + rx_outer_outer + 216]
+ for rc_outer_inner in range(2):
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 1] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 1] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 2] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 2] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 3] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 3] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 4] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 4] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 5] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 5] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 6] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 6] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 768]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 63] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 63] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 64] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 64] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 65] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 65] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 66] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 66] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 67] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 67] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 68] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 68] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 69] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 3]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 69] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 771]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 126] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 126] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 127] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 127] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 128] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 128] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 129] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 129] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 130] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 130] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 131] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 131] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 132] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 6]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 132] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 774]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 189] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 189] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 190] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 190] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 191] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 191] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 192] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 192] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 193] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 193] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 194] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 194] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 195] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 9]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 195] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 777]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 252] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 252] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 253] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 253] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 254] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 254] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 255] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 255] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 256] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 256] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 257] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 257] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 258] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 12]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 258] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 780]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 315] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 315] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 316] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 316] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 317] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 317] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 318] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 318] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 319] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 319] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 320] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 320] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 321] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 15]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 321] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 783]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 378] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 378] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 379] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 379] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 380] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 380] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 381] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 381] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 382] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 382] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 383] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 383] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 384] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 18]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 384] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 786]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 441] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 441] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 442] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 442] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 443] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 443] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 444] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 444] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 445] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 445] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 446] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 446] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 447] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 21]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 447] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 789]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 504] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 504] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 505] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 505] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 506] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 506] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 507] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 507] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 508] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 508] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 509] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 509] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 510] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 24]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 510] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 792]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 567] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 567] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 568] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 568] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 569] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 569] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 570] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 570] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 571] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 571] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 572] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 572] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 573] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 27]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 573] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 795]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 630] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 630] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 631] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 631] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 632] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 632] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 633] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 633] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 634] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 634] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 635] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 635] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 636] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 30]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 636] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 798]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 693] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 693] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 694] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 694] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 695] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 695] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 696] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 696] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 697] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 697] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 698] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 698] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 699] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 33]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 699] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 801]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 756] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 756] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 757] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 757] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 758] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 758] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 759] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 759] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 760] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 760] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 761] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 761] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 762] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 36]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 762] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 804]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 819] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 819] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 820] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 820] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 821] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 821] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 822] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 822] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 823] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 823] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 824] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 824] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 825] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 39]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 825] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 807]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 882] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 882] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 883] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 883] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 884] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 884] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 885] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 885] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 886] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 886] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 887] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 887] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 888] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 42]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 888] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 810]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 945] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 945] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 946] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 946] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 947] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 947] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 948] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 948] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 949] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 949] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 950] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 950] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 951] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 45]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 951] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 813]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 7] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 7] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 8] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 8] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 9] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 9] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 10] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 10] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 11] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 11] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 12] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 12] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 13] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 1]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 13] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 769]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 70] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 70] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 71] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 71] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 72] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 72] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 73] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 73] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 74] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 74] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 75] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 75] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 76] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 4]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 76] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 772]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 133] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 133] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 134] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 134] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 135] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 135] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 136] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 136] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 137] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 137] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 138] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 138] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 139] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 7]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 139] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 775]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 196] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 196] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 197] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 197] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 198] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 198] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 199] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 199] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 200] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 200] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 201] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 201] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 202] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 10]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 202] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 778]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 259] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 259] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 260] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 260] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 261] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 261] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 262] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 262] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 263] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 263] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 264] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 264] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 265] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 13]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 265] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 781]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 322] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 322] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 323] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 323] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 324] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 324] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 325] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 325] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 326] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 326] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 327] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 327] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 328] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 16]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 328] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 784]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 385] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 385] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 386] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 386] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 387] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 387] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 388] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 388] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 389] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 389] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 390] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 390] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 391] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 19]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 391] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 787]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 448] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 448] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 449] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 449] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 450] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 450] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 451] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 451] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 452] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 452] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 453] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 453] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 454] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 22]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 454] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 790]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 511] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 511] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 512] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 512] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 513] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 513] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 514] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 514] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 515] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 515] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 516] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 516] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 517] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 25]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 517] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 793]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 574] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 574] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 575] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 575] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 576] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 576] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 577] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 577] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 578] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 578] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 579] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 579] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 580] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 28]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 580] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 796]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 637] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 637] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 638] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 638] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 639] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 639] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 640] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 640] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 641] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 641] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 642] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 642] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 643] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 31]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 643] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 799]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 700] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 700] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 701] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 701] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 702] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 702] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 703] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 703] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 704] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 704] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 705] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 705] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 706] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 34]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 706] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 802]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 763] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 763] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 764] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 764] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 765] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 765] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 766] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 766] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 767] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 767] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 768] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 768] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 769] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 37]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 769] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 805]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 826] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 826] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 827] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 827] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 828] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 828] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 829] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 829] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 830] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 830] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 831] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 831] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 832] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 40]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 832] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 808]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 889] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 889] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 890] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 890] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 891] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 891] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 892] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 892] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 893] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 893] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 894] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 894] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 895] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 43]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 895] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 811]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 952] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 952] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 953] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 953] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 954] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 954] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 955] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 955] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 956] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 956] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 957] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 957] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 958] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 46]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 958] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 814]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 14] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 14] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 15] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 15] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 16] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 16] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 17] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 17] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 18] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 18] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 19] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 19] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 20] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 2]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 20] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 770]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 77] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 77] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 78] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 78] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 79] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 79] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 80] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 80] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 81] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 81] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 82] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 82] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 83] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 5]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 83] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 773]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 140] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 140] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 141] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 141] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 142] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 142] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 143] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 143] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 144] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 144] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 145] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 145] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 146] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 8]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 146] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 776]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 203] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 203] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 204] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 204] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 205] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 205] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 206] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 206] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 207] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 207] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 208] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 208] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 209] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 11]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 209] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 779]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 266] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 266] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 267] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 267] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 268] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 268] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 269] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 269] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 270] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 270] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 271] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 271] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 272] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 14]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 272] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 782]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 329] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 329] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 330] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 330] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 331] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 331] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 332] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 332] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 333] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 333] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 334] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 334] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 335] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 17]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 335] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 785]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 392] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 392] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 393] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 393] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 394] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 394] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 395] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 395] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 396] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 396] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 397] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 397] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 398] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 20]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 398] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 788]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 455] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 455] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 456] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 456] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 457] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 457] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 458] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 458] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 459] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 459] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 460] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 460] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 461] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 23]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 461] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 791]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 518] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 518] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 519] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 519] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 520] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 520] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 521] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 521] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 522] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 522] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 523] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 523] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 524] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 26]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 524] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 794]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 581] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 581] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 582] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 582] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 583] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 583] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 584] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 584] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 585] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 585] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 586] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 586] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 587] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 29]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 587] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 797]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 644] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 644] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 645] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 645] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 646] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 646] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 647] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 647] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 648] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 648] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 649] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 649] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 650] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 32]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 650] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 800]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 707] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 707] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 708] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 708] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 709] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 709] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 710] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 710] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 711] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 711] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 712] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 712] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 713] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 35]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 713] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 803]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 770] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 770] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 771] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 771] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 772] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 772] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 773] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 773] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 774] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 774] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 775] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 775] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 776] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 38]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 776] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 806]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 833] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 833] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 834] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 834] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 835] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 835] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 836] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 836] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 837] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 837] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 838] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 838] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 839] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 41]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 839] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 809]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 896] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 896] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 897] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 897] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 898] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 898] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 899] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 899] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 900] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 900] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 901] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 901] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 902] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 44]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 902] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 812]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 959] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 959] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 960] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 960] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 961] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 961] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 962] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 962] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 963] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 963] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 964] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 964] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 965] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 47]
+ conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[rc_outer_inner * 1008 + threadIdx_x % 7 * 7 + 965] * kernel_shared_1[threadIdx_x // 7 * 96 + rc_outer_inner * 48 + 815]
+ for i3_inner in range(7):
compute_1 = T.Buffer((25088,), data=compute.data)
bias_1 = T.Buffer((512,), data=bias.data)
- compute_1[blockIdx_x * 784 + threadIdx_x // 7 * 98 + i1_inner * 49 + i2_inner * 7 + threadIdx_x % 7] = T.max(conv2d_nchw_1[i1_inner * 7 + i2_inner] + bias_1[blockIdx_x * 16 + threadIdx_x // 7 * 2 + i1_inner], T.float32(0))
+ compute_1[blockIdx_x * 784 + threadIdx_x * 7 + i3_inner] = T.max(conv2d_nchw_1[i3_inner] + bias_1[blockIdx_x * 16 + threadIdx_x // 7], T.float32(0))
+ compute_1[blockIdx_x * 784 + threadIdx_x * 7 + i3_inner + 392] = T.max(conv2d_nchw_1[i3_inner + 7] + bias_1[blockIdx_x * 16 + threadIdx_x // 7 + 8], T.float32(0))
</pre></div>
</div>
</div>
@@ -722,7 +1379,7 @@ class Module:
<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: 0.275 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 0.325 ms
</pre></div>
</div>
</div>
@@ -751,36 +1408,36 @@ conv2d_nchw_nn_o_i, conv2d_nchw_nn_i = s[conv2d_nchw].split(conv2d_nchw_nn, fact
conv2d_nchw_nn_o_o_i, conv2d_nchw_nn_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_i, factor=1)
conv2d_nchw_nn_o_o_o_i, conv2d_nchw_nn_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_i, factor=1)
conv2d_nchw_nn_o_o_o_o, conv2d_nchw_nn_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_o_i, factor=1)
-conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=2)
+conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=1)
conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=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=8)
-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=7)
+conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=2)
+conv2d_nchw_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=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=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_o_o_i, conv2d_nchw_xx_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_i, factor=7)
+conv2d_nchw_xx_o_o_o_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=4)
-conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=8)
+conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=16)
+conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=2)
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_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=1)
conv2d_nchw_rx_o_o, conv2d_nchw_rx_o_i = s[conv2d_nchw].split(conv2d_nchw_rx_o_i, factor=1)
s[conv2d_nchw].reorder(conv2d_nchw_nn_o_o_o_o, conv2d_nchw_ff_o_o_o_o, conv2d_nchw_yy_o_o_o_o, conv2d_nchw_xx_o_o_o_o, conv2d_nchw_nn_o_o_o_i, conv2d_nchw_ff_o_o_o_i, conv2d_nchw_yy_o_o_o_i, conv2d_nchw_xx_o_o_o_i, conv2d_nchw_nn_o_o_i, conv2d_nchw_ff_o_o_i, conv2d_nchw_yy_o_o_i, conv2d_nchw_xx_o_o_i, conv2d_nchw_rc_o_o, conv2d_nchw_ry_o_o, conv2d_nchw_rx_o_o, conv2d_nchw_rc_o_i, conv2d_nchw_ry_o_i, conv2d_nchw_rx_o_i, conv2d_nchw_nn_o_i, conv2d_nchw_ff_o_i, conv2d_nchw_yy_o_i, conv2d_nc [...]
compute_i0_o_i, compute_i0_i = s[compute].split(compute_i0, factor=1)
compute_i0_o_o_i, compute_i0_o_i = s[compute].split(compute_i0_o_i, factor=1)
compute_i0_o_o_o, compute_i0_o_o_i = s[compute].split(compute_i0_o_o_i, factor=1)
-compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=2)
+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=8)
-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=7)
-compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=1)
+compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=2)
+compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
+compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=7)
compute_i2_o_o_o, compute_i2_o_o_i = s[compute].split(compute_i2_o_o_i, factor=1)
-compute_i3_o_i, compute_i3_i = s[compute].split(compute_i3, factor=1)
-compute_i3_o_o_i, compute_i3_o_i = s[compute].split(compute_i3_o_i, factor=7)
+compute_i3_o_i, compute_i3_i = s[compute].split(compute_i3, factor=7)
+compute_i3_o_o_i, compute_i3_o_i = s[compute].split(compute_i3_o_i, factor=1)
compute_i3_o_o_o, compute_i3_o_o_i = s[compute].split(compute_i3_o_o_i, factor=1)
s[compute].reorder(compute_i0_o_o_o, compute_i1_o_o_o, compute_i2_o_o_o, compute_i3_o_o_o, compute_i0_o_o_i, compute_i1_o_o_i, compute_i2_o_o_i, compute_i3_o_o_i, compute_i0_o_i, compute_i1_o_i, compute_i2_o_i, compute_i3_o_i, compute_i0_i, compute_i1_i, compute_i2_i, compute_i3_i)
s[conv2d_nchw].compute_at(s[compute], compute_i3_o_i)
@@ -798,7 +1455,7 @@ 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=4)
+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)
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=56)
s[kernel_shared].bind(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
@@ -807,7 +1464,7 @@ pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, pad_temp_shared_ax0_ax1_fus
s[pad_temp_shared].vectorize(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=56)
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", 64)
+s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 1024)
s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
@@ -827,141 +1484,771 @@ CUDA source code:
#endif
extern "C" __global__ void __launch_bounds__(56) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
float conv2d_nchw[14];
- __shared__ float pad_temp_shared[2592];
- __shared__ float kernel_shared[4608];
+ __shared__ float pad_temp_shared[2016];
+ __shared__ float kernel_shared[1536];
conv2d_nchw[0] = 0.000000e+00f;
- conv2d_nchw[1] = 0.000000e+00f;
- conv2d_nchw[2] = 0.000000e+00f;
- conv2d_nchw[3] = 0.000000e+00f;
- conv2d_nchw[4] = 0.000000e+00f;
- conv2d_nchw[5] = 0.000000e+00f;
- conv2d_nchw[6] = 0.000000e+00f;
conv2d_nchw[7] = 0.000000e+00f;
+ conv2d_nchw[1] = 0.000000e+00f;
conv2d_nchw[8] = 0.000000e+00f;
+ conv2d_nchw[2] = 0.000000e+00f;
conv2d_nchw[9] = 0.000000e+00f;
+ conv2d_nchw[3] = 0.000000e+00f;
conv2d_nchw[10] = 0.000000e+00f;
+ conv2d_nchw[4] = 0.000000e+00f;
conv2d_nchw[11] = 0.000000e+00f;
+ conv2d_nchw[5] = 0.000000e+00f;
conv2d_nchw[12] = 0.000000e+00f;
+ conv2d_nchw[6] = 0.000000e+00f;
conv2d_nchw[13] = 0.000000e+00f;
for (int rc_outer_outer = 0; rc_outer_outer < 16; ++rc_outer_outer) {
- __syncthreads();
- pad_temp_shared[((int)threadIdx.x)] = ((((9 <= ((int)threadIdx.x)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((rc_outer_outer * 1568) + ((((int)threadIdx.x) / 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 56)] = (((((9 <= ((((int)threadIdx.x) + 56) % 81)) && (((((int)threadIdx.x) + 56) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 56) / 81) * 49)) + ((((((int)threadIdx.x) + 56) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 112)] = (((((9 <= ((((int)threadIdx.x) + 31) % 81)) && (((((int)threadIdx.x) + 31) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 112) / 81) * 49)) + ((((((int)threadIdx.x) + 31) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 168)] = ((((3 <= ((int)threadIdx.x)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 168) / 81) * 49)) + (((((int)threadIdx.x) + 6) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 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 * 1568) + (((((int)threadIdx.x) + 224) / 81) * 49)) + ((((((int)threadIdx.x) + 62) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 280)] = (((((9 <= ((((int)threadIdx.x) + 37) % 81)) && (((((int)threadIdx.x) + 37) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 280) / 81) * 49)) + ((((((int)threadIdx.x) + 37) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 336)] = (((1 <= ((((int)threadIdx.x) + 3) % 9)) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 336) / 81) * 49)) + (((((int)threadIdx.x) + 12) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 392)] = (((((9 <= ((((int)threadIdx.x) + 68) % 81)) && (((((int)threadIdx.x) + 68) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 392) / 81) * 49)) + ((((((int)threadIdx.x) + 68) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- 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 * 1568) + (((((int)threadIdx.x) + 448) / 81) * 49)) + ((((((int)threadIdx.x) + 43) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 504)] = ((((((int)threadIdx.x) < 54) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 504) / 81) * 49)) + ((((int)threadIdx.x) / 9) * 7)) + (((int)threadIdx.x) % 9)) + 6)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 560)] = (((((9 <= ((((int)threadIdx.x) + 74) % 81)) && (((((int)threadIdx.x) + 74) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 560) / 81) * 49)) + ((((((int)threadIdx.x) + 74) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 616)] = (((((9 <= ((((int)threadIdx.x) + 49) % 81)) && (((((int)threadIdx.x) + 49) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 616) / 81) * 49)) + ((((((int)threadIdx.x) + 49) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 672)] = ((((((int)threadIdx.x) < 48) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 672) / 81) * 49)) + (((((int)threadIdx.x) + 24) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 728)] = (((((9 <= ((((int)threadIdx.x) + 80) % 81)) && (((((int)threadIdx.x) + 80) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 728) / 81) * 49)) + ((((((int)threadIdx.x) + 80) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 784)] = (((((9 <= ((((int)threadIdx.x) + 55) % 81)) && (((((int)threadIdx.x) + 55) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 784) / 81) * 49)) + ((((((int)threadIdx.x) + 55) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 840)] = (((((9 <= ((((int)threadIdx.x) + 30) % 81)) && (((((int)threadIdx.x) + 30) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 840) / 81) * 49)) + ((((((int)threadIdx.x) + 30) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 896)] = ((((4 <= ((int)threadIdx.x)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 896) / 81) * 49)) + (((((int)threadIdx.x) + 5) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 952)] = (((((9 <= ((((int)threadIdx.x) + 61) % 81)) && (((((int)threadIdx.x) + 61) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 952) / 81) * 49)) + ((((((int)threadIdx.x) + 61) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1008)] = (((((1 <= (((((int)threadIdx.x) / 9) + 4) % 9)) && (((((int)threadIdx.x) + 36) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1008) / 81) * 49)) + ((((((int)threadIdx.x) / 9) + 4) % 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1064)] = (((1 <= ((((int)threadIdx.x) + 2) % 9)) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1064) / 81) * 49)) + (((((int)threadIdx.x) + 11) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1120)] = (((((9 <= ((((int)threadIdx.x) + 67) % 81)) && (((((int)threadIdx.x) + 67) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1120) / 81) * 49)) + ((((((int)threadIdx.x) + 67) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1176)] = (((((9 <= ((((int)threadIdx.x) + 42) % 81)) && (((((int)threadIdx.x) + 42) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1176) / 81) * 49)) + ((((((int)threadIdx.x) + 42) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1232)] = ((((((int)threadIdx.x) < 55) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1232) / 81) * 49)) + (((((int)threadIdx.x) + 17) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1288)] = (((((9 <= ((((int)threadIdx.x) + 73) % 81)) && (((((int)threadIdx.x) + 73) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1288) / 81) * 49)) + ((((((int)threadIdx.x) + 73) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1344)] = (((((9 <= ((((int)threadIdx.x) + 48) % 81)) && (((((int)threadIdx.x) + 48) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1344) / 81) * 49)) + ((((((int)threadIdx.x) + 48) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1400)] = ((((((int)threadIdx.x) < 49) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1400) / 81) * 49)) + (((((int)threadIdx.x) + 23) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1456)] = (((((9 <= ((((int)threadIdx.x) + 79) % 81)) && (((((int)threadIdx.x) + 79) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1456) / 81) * 49)) + ((((((int)threadIdx.x) + 79) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1512)] = (((((1 <= (((((int)threadIdx.x) / 9) + 6) % 9)) && (((((int)threadIdx.x) + 54) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1512) / 81) * 49)) + ((((((int)threadIdx.x) / 9) + 6) % 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1568)] = (((((9 <= ((((int)threadIdx.x) + 29) % 81)) && (((((int)threadIdx.x) + 29) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1568) / 81) * 49)) + ((((((int)threadIdx.x) + 29) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1624)] = ((((5 <= ((int)threadIdx.x)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1624) / 81) * 49)) + (((((int)threadIdx.x) + 4) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1680)] = (((((9 <= ((((int)threadIdx.x) + 60) % 81)) && (((((int)threadIdx.x) + 60) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1680) / 81) * 49)) + ((((((int)threadIdx.x) + 60) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1736)] = (((((9 <= ((((int)threadIdx.x) + 35) % 81)) && (((((int)threadIdx.x) + 35) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1736) / 81) * 49)) + ((((((int)threadIdx.x) + 35) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1792)] = (((1 <= ((((int)threadIdx.x) + 1) % 9)) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1792) / 81) * 49)) + (((((int)threadIdx.x) + 10) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1848)] = (((((9 <= ((((int)threadIdx.x) + 66) % 81)) && (((((int)threadIdx.x) + 66) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1848) / 81) * 49)) + ((((((int)threadIdx.x) + 66) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1904)] = (((((9 <= ((((int)threadIdx.x) + 41) % 81)) && (((((int)threadIdx.x) + 41) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1904) / 81) * 49)) + ((((((int)threadIdx.x) + 41) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 1960)] = (((1 <= ((((int)threadIdx.x) + 7) % 9)) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1960) / 81) * 49)) + (((((int)threadIdx.x) + 16) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2016)] = (((((1 <= (((((int)threadIdx.x) / 9) + 8) % 9)) && (((((int)threadIdx.x) + 72) % 81) < 72)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2016) / 81) * 49)) + ((((((int)threadIdx.x) / 9) + 8) % 9) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2072)] = (((((9 <= ((((int)threadIdx.x) + 47) % 81)) && (((((int)threadIdx.x) + 47) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2072) / 81) * 49)) + ((((((int)threadIdx.x) + 47) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2128)] = ((((((int)threadIdx.x) < 50) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2128) / 81) * 49)) + (((((int)threadIdx.x) + 22) / 9) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2184)] = (((((9 <= ((((int)threadIdx.x) + 78) % 81)) && (((((int)threadIdx.x) + 78) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2184) / 81) * 49)) + ((((((int)threadIdx.x) + 78) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2240)] = (((((9 <= ((((int)threadIdx.x) + 53) % 81)) && (((((int)threadIdx.x) + 53) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2240) / 81) * 49)) + ((((((int)threadIdx.x) + 53) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2296)] = (((((9 <= ((((int)threadIdx.x) + 28) % 81)) && (((((int)threadIdx.x) + 28) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2296) / 81) * 49)) + ((((((int)threadIdx.x) + 28) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2352)] = ((((6 <= ((int)threadIdx.x)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2352) / 81) * 49)) + (((((int)threadIdx.x) + 3) / 9) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2408)] = (((((9 <= ((((int)threadIdx.x) + 59) % 81)) && (((((int)threadIdx.x) + 59) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2408) / 81) * 49)) + ((((((int)threadIdx.x) + 59) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2464)] = (((((9 <= ((((int)threadIdx.x) + 34) % 81)) && (((((int)threadIdx.x) + 34) % 81) < 72)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2464) / 81) * 49)) + ((((((int)threadIdx.x) + 34) % 81) / 9) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 2520)] = (((1 <= (((int)threadIdx.x) % 9)) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2520) / 81) * 49)) + ((((int)threadIdx.x) / 9) * 7)) + (((int)threadIdx.x) % 9)) - 1)] : 0.000000e+00f);
- if (((int)threadIdx.x) < 16) {
- pad_temp_shared[(((int)threadIdx.x) + 2576)] = ((((((int)threadIdx.x) < 7) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 2576) / 81) * 49)) + (((((int)threadIdx.x) + 65) / 9) * 7)) + ((int)threadIdx.x)) - 6)] : 0.000000e+00f);
- }
- for (int ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer = 0; ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer < 21; ++ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer) {
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) < 144) {
- kernel_shared[((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4))] = kernel[((((((int)blockIdx.x) * 73728) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) / 9) * 4608)) + (rc_outer_outer * 288)) + (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) % 288))];
+ for (int rx_outer_outer = 0; rx_outer_outer < 3; ++rx_outer_outer) {
+ __syncthreads();
+ pad_temp_shared[((int)threadIdx.x)] = ((((7 <= ((int)threadIdx.x)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((rc_outer_outer * 1568) + ((int)threadIdx.x)) + rx_outer_outer) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 56)] = (((((1 <= (((((int)threadIdx.x) / 7) + 8) % 9)) && ((((((int)threadIdx.x) / 7) + 8) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 56) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 8) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 112)] = (((((1 <= (((((int)threadIdx.x) / 7) + 7) % 9)) && ((((((int)threadIdx.x) / 7) + 7) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 112) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 7) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 168)] = (((((1 <= (((((int)threadIdx.x) / 7) + 6) % 9)) && ((((((int)threadIdx.x) / 7) + 6) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 168) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 6) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((1 <= (((((int)threadIdx.x) / 7) + 5) % 9)) && ((((((int)threadIdx.x) / 7) + 5) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 224) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 5) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 280)] = (((((1 <= (((((int)threadIdx.x) / 7) + 4) % 9)) && ((((((int)threadIdx.x) / 7) + 4) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 280) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 4) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 336)] = (((((1 <= (((((int)threadIdx.x) / 7) + 3) % 9)) && ((((((int)threadIdx.x) / 7) + 3) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 336) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 3) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 392)] = (((((1 <= (((((int)threadIdx.x) / 7) + 2) % 9)) && ((((((int)threadIdx.x) / 7) + 2) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 392) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 2) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 448)] = ((((((int)threadIdx.x) < 49) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 448) / 63) * 49)) + ((int)threadIdx.x)) + rx_outer_outer) - 1)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 504)] = ((((7 <= ((int)threadIdx.x)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((rc_outer_outer * 1568) + ((int)threadIdx.x)) + rx_outer_outer) + 384)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 560)] = (((((1 <= (((((int)threadIdx.x) / 7) + 8) % 9)) && ((((((int)threadIdx.x) / 7) + 8) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 560) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 8) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 616)] = (((((1 <= (((((int)threadIdx.x) / 7) + 7) % 9)) && ((((((int)threadIdx.x) / 7) + 7) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 616) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 7) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 672)] = (((((1 <= (((((int)threadIdx.x) / 7) + 6) % 9)) && ((((((int)threadIdx.x) / 7) + 6) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 672) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 6) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 728)] = (((((1 <= (((((int)threadIdx.x) / 7) + 5) % 9)) && ((((((int)threadIdx.x) / 7) + 5) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 728) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 5) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 784)] = (((((1 <= (((((int)threadIdx.x) / 7) + 4) % 9)) && ((((((int)threadIdx.x) / 7) + 4) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 784) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 4) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 840)] = (((((1 <= (((((int)threadIdx.x) / 7) + 3) % 9)) && ((((((int)threadIdx.x) / 7) + 3) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 840) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 3) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 896)] = (((((1 <= (((((int)threadIdx.x) / 7) + 2) % 9)) && ((((((int)threadIdx.x) / 7) + 2) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 896) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 2) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 952)] = ((((((int)threadIdx.x) < 49) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 952) / 63) * 49)) + ((int)threadIdx.x)) + rx_outer_outer) - 1)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1008)] = ((((7 <= ((int)threadIdx.x)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((rc_outer_outer * 1568) + ((int)threadIdx.x)) + rx_outer_outer) + 776)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1064)] = (((((1 <= (((((int)threadIdx.x) / 7) + 8) % 9)) && ((((((int)threadIdx.x) / 7) + 8) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1064) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 8) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1120)] = (((((1 <= (((((int)threadIdx.x) / 7) + 7) % 9)) && ((((((int)threadIdx.x) / 7) + 7) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1120) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 7) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1176)] = (((((1 <= (((((int)threadIdx.x) / 7) + 6) % 9)) && ((((((int)threadIdx.x) / 7) + 6) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1176) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 6) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1232)] = (((((1 <= (((((int)threadIdx.x) / 7) + 5) % 9)) && ((((((int)threadIdx.x) / 7) + 5) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1232) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 5) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1288)] = (((((1 <= (((((int)threadIdx.x) / 7) + 4) % 9)) && ((((((int)threadIdx.x) / 7) + 4) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1288) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 4) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1344)] = (((((1 <= (((((int)threadIdx.x) / 7) + 3) % 9)) && ((((((int)threadIdx.x) / 7) + 3) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1344) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 3) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1400)] = (((((1 <= (((((int)threadIdx.x) / 7) + 2) % 9)) && ((((((int)threadIdx.x) / 7) + 2) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1400) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 2) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1456)] = ((((((int)threadIdx.x) < 49) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1456) / 63) * 49)) + ((int)threadIdx.x)) + rx_outer_outer) - 1)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1512)] = ((((7 <= ((int)threadIdx.x)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((rc_outer_outer * 1568) + ((int)threadIdx.x)) + rx_outer_outer) + 1168)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1568)] = (((((1 <= (((((int)threadIdx.x) / 7) + 8) % 9)) && ((((((int)threadIdx.x) / 7) + 8) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1568) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 8) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1624)] = (((((1 <= (((((int)threadIdx.x) / 7) + 7) % 9)) && ((((((int)threadIdx.x) / 7) + 7) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1624) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 7) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1680)] = (((((1 <= (((((int)threadIdx.x) / 7) + 6) % 9)) && ((((((int)threadIdx.x) / 7) + 6) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1680) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 6) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1736)] = (((((1 <= (((((int)threadIdx.x) / 7) + 5) % 9)) && ((((((int)threadIdx.x) / 7) + 5) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1736) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 5) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1792)] = (((((1 <= (((((int)threadIdx.x) / 7) + 4) % 9)) && ((((((int)threadIdx.x) / 7) + 4) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1792) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 4) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1848)] = (((((1 <= (((((int)threadIdx.x) / 7) + 3) % 9)) && ((((((int)threadIdx.x) / 7) + 3) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1848) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 3) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1904)] = (((((1 <= (((((int)threadIdx.x) / 7) + 2) % 9)) && ((((((int)threadIdx.x) / 7) + 2) % 9) < 8)) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[((((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1904) / 63) * 49)) + ((((((int)threadIdx.x) / 7) + 2) % 9) * 7)) + rx_outer_outer) + (((int)threadIdx.x) % 7)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[(((int)threadIdx.x) + 1960)] = ((((((int)threadIdx.x) < 49) && (1 <= (rx_outer_outer + (((int)threadIdx.x) % 7)))) && ((rx_outer_outer + (((int)threadIdx.x) % 7)) < 8)) ? data[(((((rc_outer_outer * 1568) + (((((int)threadIdx.x) + 1960) / 63) * 49)) + ((int)threadIdx.x)) + rx_outer_outer) - 1)] : 0.000000e+00f);
+ kernel_shared[((int)threadIdx.x)] = kernel[((((((int)blockIdx.x) * 73728) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 56)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 56) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 56) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 112)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 112) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 48)];
+ kernel_shared[(((int)threadIdx.x) + 168)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 168) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 72) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 224)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 224) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 96)];
+ kernel_shared[(((int)threadIdx.x) + 280)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 280) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 88) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 336)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 336) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 48) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 392)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 392) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 24)];
+ kernel_shared[(((int)threadIdx.x) + 448)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 448) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 64) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 504)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 504) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 72)];
+ kernel_shared[(((int)threadIdx.x) + 560)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 560) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 80) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 616)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 616) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 120)];
+ kernel_shared[(((int)threadIdx.x) + 672)] = kernel[(((((((int)blockIdx.x) * 73728) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 32256)];
+ kernel_shared[(((int)threadIdx.x) + 728)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 728) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 56) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 784)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 784) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 48)];
+ kernel_shared[(((int)threadIdx.x) + 840)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 840) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 72) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 896)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 896) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 96)];
+ kernel_shared[(((int)threadIdx.x) + 952)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 952) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 88) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1008)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1008) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 48) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1064)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1064) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 24)];
+ kernel_shared[(((int)threadIdx.x) + 1120)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1120) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 64) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1176)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1176) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 72)];
+ kernel_shared[(((int)threadIdx.x) + 1232)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1232) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 80) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1288)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1288) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 120)];
+ kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[(((((((int)blockIdx.x) * 73728) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 64512)];
+ kernel_shared[(((int)threadIdx.x) + 1400)] = kernel[(((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1400) / 96) * 4608)) + (rc_outer_outer * 288)) + (((((int)threadIdx.x) + 56) % 96) * 3)) + rx_outer_outer)];
+ kernel_shared[(((int)threadIdx.x) + 1456)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1456) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 48)];
+ if (((int)threadIdx.x) < 24) {
+ kernel_shared[(((int)threadIdx.x) + 1512)] = kernel[((((((((int)blockIdx.x) * 73728) + (((((int)threadIdx.x) + 1512) / 96) * 4608)) + (rc_outer_outer * 288)) + (((int)threadIdx.x) * 3)) + rx_outer_outer) + 216)];
}
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) < 144) {
- kernel_shared[(((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 1)] = kernel[((((((int)blockIdx.x) * 73728) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) / 9) * 4608)) + (rc_outer_outer * 288)) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 1) % 288))];
- }
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) < 144) {
- kernel_shared[(((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 2)] = kernel[((((((int)blockIdx.x) * 73728) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) / 9) * 4608)) + (rc_outer_outer * 288)) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 2) % 288))];
- }
- if (((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) < 144) {
- kernel_shared[(((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 3)] = kernel[((((((int)blockIdx.x) * 73728) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 7) + (((int)threadIdx.x) >> 3)) / 9) * 4608)) + (rc_outer_outer * 288)) + ((((ax0_ax1_fused_ax2_fused_ax3_fused_outer_outer * 224) + (((int)threadIdx.x) * 4)) + 3) % 288))];
- }
- }
- __syncthreads();
- for (int rc_outer_inner = 0; rc_outer_inner < 8; ++rc_outer_inner) {
- for (int ry_outer_inner = 0; ry_outer_inner < 3; ++ry_outer_inner) {
- for (int rc_inner = 0; rc_inner < 4; ++rc_inner) {
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7))] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 9)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 18)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 27)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 36)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 45)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 54)] * kernel_shared[(((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3))]));
- conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7))] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 9)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 18)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 27)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 36)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 45)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 54)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 288)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 1)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 10)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 19)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 28)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 37)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 46)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 55)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 1)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 10)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 19)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 28)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 37)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 46)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 55)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 289)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 2)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 11)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 20)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 29)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 38)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 47)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 56)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 2)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 11)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 20)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 29)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 38)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 47)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((rc_outer_inner * 324) + (rc_inner * 81)) + (ry_outer_inner * 9)) + (((int)threadIdx.x) % 7)) + 56)] * kernel_shared[((((((((int)threadIdx.x) / 7) * 576) + (rc_outer_inner * 36)) + (rc_inner * 9)) + (ry_outer_inner * 3)) + 290)]));
- }
+ __syncthreads();
+ for (int rc_outer_inner = 0; rc_outer_inner < 2; ++rc_outer_inner) {
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7))] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7))] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 1)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 2)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 3)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 4)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 5)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48))]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 6)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 768)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 63)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 63)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 64)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 64)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 65)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 65)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 66)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 66)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 67)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 67)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 68)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 68)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 69)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 3)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 69)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 771)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 126)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 126)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 127)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 127)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 128)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 128)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 129)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 129)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 130)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 130)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 131)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 131)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 132)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 6)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 132)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 774)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 189)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 189)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 190)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 190)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 191)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 191)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 192)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 192)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 193)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 193)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 194)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 194)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 195)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 9)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 195)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 777)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 252)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 252)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 253)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 253)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 254)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 254)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 255)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 255)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 256)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 256)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 257)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 257)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 258)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 12)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 258)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 780)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 315)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 315)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 316)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 316)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 317)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 317)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 318)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 318)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 319)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 319)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 320)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 320)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 321)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 15)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 321)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 783)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 378)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 378)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 379)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 379)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 380)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 380)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 381)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 381)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 382)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 382)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 383)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 383)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 384)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 18)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 384)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 786)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 441)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 441)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 442)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 442)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 443)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 443)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 444)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 444)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 445)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 445)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 446)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 446)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 447)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 21)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 447)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 789)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 504)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 504)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 505)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 505)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 506)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 506)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 507)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 507)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 508)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 508)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 509)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 509)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 510)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 24)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 510)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 792)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 567)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 567)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 568)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 568)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 569)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 569)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 570)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 570)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 571)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 571)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 572)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 572)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 573)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 27)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 573)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 795)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 630)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 630)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 631)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 631)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 632)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 632)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 633)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 633)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 634)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 634)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 635)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 635)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 636)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 30)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 636)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 798)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 693)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 693)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 694)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 694)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 695)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 695)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 696)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 696)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 697)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 697)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 698)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 698)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 699)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 33)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 699)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 801)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 756)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 756)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 757)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 757)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 758)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 758)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 759)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 759)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 760)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 760)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 761)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 761)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 762)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 36)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 762)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 804)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 819)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 819)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 820)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 820)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 821)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 821)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 822)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 822)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 823)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 823)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 824)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 824)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 825)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 39)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 825)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 807)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 882)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 882)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 883)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 883)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 884)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 884)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 885)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 885)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 886)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 886)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 887)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 887)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 888)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 42)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 888)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 810)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 945)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 945)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 946)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 946)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 947)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 947)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 948)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 948)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 949)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 949)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 950)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 950)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 951)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 45)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 951)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 813)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 7)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 7)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 8)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 8)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 9)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 9)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 10)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 10)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 11)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 11)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 12)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 12)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 13)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 1)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 13)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 769)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 70)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 70)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 71)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 71)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 72)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 72)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 73)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 73)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 74)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 74)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 75)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 75)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 76)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 4)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 76)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 772)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 133)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 133)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 134)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 134)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 135)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 135)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 136)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 136)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 137)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 137)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 138)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 138)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 139)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 7)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 139)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 775)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 196)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 196)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 197)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 197)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 198)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 198)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 199)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 199)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 200)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 200)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 201)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 201)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 202)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 10)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 202)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 778)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 259)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 259)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 260)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 260)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 261)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 261)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 262)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 262)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 263)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 263)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 264)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 264)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 265)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 13)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 265)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 781)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 322)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 322)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 323)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 323)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 324)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 324)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 325)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 325)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 326)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 326)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 327)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 327)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 328)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 16)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 328)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 784)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 385)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 385)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 386)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 386)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 387)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 387)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 388)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 388)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 389)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 389)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 390)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 390)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 391)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 19)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 391)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 787)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 448)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 448)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 449)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 449)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 450)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 450)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 451)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 451)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 452)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 452)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 453)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 453)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 454)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 22)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 454)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 790)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 511)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 511)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 512)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 512)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 513)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 513)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 514)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 514)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 515)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 515)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 516)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 516)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 517)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 25)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 517)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 793)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 574)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 574)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 575)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 575)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 576)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 576)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 577)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 577)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 578)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 578)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 579)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 579)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 580)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 28)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 580)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 796)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 637)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 637)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 638)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 638)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 639)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 639)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 640)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 640)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 641)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 641)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 642)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 642)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 643)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 31)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 643)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 799)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 700)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 700)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 701)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 701)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 702)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 702)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 703)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 703)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 704)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 704)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 705)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 705)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 706)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 34)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 706)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 802)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 763)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 763)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 764)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 764)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 765)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 765)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 766)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 766)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 767)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 767)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 768)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 768)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 769)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 37)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 769)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 805)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 826)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 826)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 827)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 827)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 828)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 828)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 829)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 829)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 830)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 830)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 831)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 831)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 832)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 40)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 832)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 808)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 889)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 889)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 890)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 890)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 891)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 891)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 892)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 892)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 893)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 893)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 894)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 894)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 895)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 43)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 895)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 811)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 952)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 952)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 953)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 953)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 954)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 954)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 955)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 955)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 956)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 956)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 957)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 957)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 958)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 46)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 958)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 814)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 14)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 14)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 15)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 15)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 16)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 16)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 17)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 17)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 18)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 18)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 19)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 19)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 20)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 2)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 20)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 770)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 77)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 77)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 78)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 78)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 79)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 79)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 80)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 80)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 81)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 81)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 82)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 82)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 83)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 5)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 83)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 773)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 140)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 140)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 141)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 141)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 142)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 142)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 143)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 143)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 144)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 144)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 145)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 145)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 146)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 8)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 146)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 776)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 203)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 203)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 204)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 204)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 205)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 205)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 206)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 206)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 207)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 207)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 208)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 208)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 209)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 11)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 209)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 779)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 266)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 266)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 267)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 267)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 268)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 268)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 269)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 269)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 270)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 270)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 271)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 271)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 272)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 14)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 272)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 782)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 329)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 329)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 330)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 330)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 331)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 331)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 332)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 332)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 333)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 333)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 334)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 334)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 335)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 17)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 335)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 785)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 392)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 392)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 393)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 393)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 394)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 394)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 395)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 395)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 396)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 396)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 397)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 397)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 398)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 20)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 398)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 788)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 455)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 455)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 456)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 456)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 457)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 457)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 458)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 458)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 459)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 459)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 460)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 460)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 461)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 23)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 461)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 791)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 518)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 518)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 519)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 519)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 520)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 520)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 521)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 521)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 522)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 522)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 523)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 523)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 524)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 26)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 524)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 794)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 581)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 581)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 582)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 582)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 583)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 583)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 584)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 584)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 585)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 585)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 586)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 586)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 587)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 29)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 587)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 797)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 644)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 644)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 645)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 645)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 646)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 646)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 647)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 647)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 648)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 648)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 649)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 649)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 650)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 32)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 650)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 800)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 707)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 707)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 708)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 708)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 709)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 709)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 710)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 710)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 711)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 711)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 712)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 712)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 713)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 35)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 713)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 803)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 770)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 770)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 771)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 771)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 772)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 772)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 773)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 773)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 774)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 774)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 775)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 775)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 776)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 38)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 776)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 806)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 833)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 833)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 834)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 834)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 835)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 835)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 836)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 836)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 837)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 837)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 838)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 838)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 839)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 41)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 839)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 809)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 896)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 896)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 897)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 897)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 898)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 898)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 899)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 899)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 900)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 900)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 901)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 901)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 902)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 44)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 902)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 812)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 959)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 959)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 960)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 960)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 961)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 961)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 962)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 962)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 963)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 963)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 964)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 964)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 965)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 47)]));
+ conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((rc_outer_inner * 1008) + ((((int)threadIdx.x) % 7) * 7)) + 965)] * kernel_shared[((((((int)threadIdx.x) / 7) * 96) + (rc_outer_inner * 48)) + 815)]));
}
}
}
- for (int i1_inner = 0; i1_inner < 2; ++i1_inner) {
- for (int i2_inner = 0; i2_inner < 7; ++i2_inner) {
- compute[(((((((int)blockIdx.x) * 784) + ((((int)threadIdx.x) / 7) * 98)) + (i1_inner * 49)) + (i2_inner * 7)) + (((int)threadIdx.x) % 7))] = max((conv2d_nchw[((i1_inner * 7) + i2_inner)] + bias[(((((int)blockIdx.x) * 16) + ((((int)threadIdx.x) / 7) * 2)) + i1_inner)]), 0.000000e+00f);
- }
+ for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
+ compute[(((((int)blockIdx.x) * 784) + (((int)threadIdx.x) * 7)) + i3_inner)] = max((conv2d_nchw[i3_inner] + bias[((((int)blockIdx.x) * 16) + (((int)threadIdx.x) / 7))]), 0.000000e+00f);
+ compute[((((((int)blockIdx.x) * 784) + (((int)threadIdx.x) * 7)) + i3_inner) + 392)] = max((conv2d_nchw[(i3_inner + 7)] + bias[(((((int)blockIdx.x) * 16) + (((int)threadIdx.x) / 7)) + 8)]), 0.000000e+00f);
}
}
</pre></div>
@@ -996,7 +2283,7 @@ 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> ( 6 minutes 9.023 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 6 minutes 24.225 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 fbfb1dfd69..d87e452f4f 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
@@ -921,7 +921,7 @@ 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)
- 7.8896 7.8893 7.8964 7.8830 0.0054
+ 7.9159 7.9132 7.9271 7.9074 0.0082
</pre></div>
</div>
</div>
@@ -943,7 +943,7 @@ 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 7.297 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 9.353 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-tune-with-autoscheduler-tune-network-cuda-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/eafe360d52540634c9eea0fa89e804bd/tune_network_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_network_cuda.py</span></code></a></p>
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 439c62ac6d..e40e6921e3 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
@@ -940,7 +940,7 @@ 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)
- 747.7804 748.3530 750.0463 744.9418 2.1229
+ 757.4250 755.7490 760.8145 755.7114 2.3968
</pre></div>
</div>
</div>
@@ -962,7 +962,7 @@ 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 41.602 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 44.830 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 f498abdbdb..6867f787af 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
@@ -640,9 +640,9 @@ class Module:
for i0_outer_i1_outer_fused in T.parallel(32):
compute_1 = T.allocate([2048], "float32", "global")
compute_2 = T.Buffer((2048,), data=compute_1)
- for i_outer_inner, nb_j_inner in T.grid(2, 2):
- for i_inner_init in range(32):
- cse_var_1: T.int32 = i_outer_inner * 1024 + i_inner_init * 32 + nb_j_inner * 16
+ for nb_j_inner in range(2):
+ for i_inner_init in range(64):
+ cse_var_1: T.int32 = i_inner_init * 32 + nb_j_inner * 16
compute_2[cse_var_1] = T.float32(0)
compute_2[cse_var_1 + 1] = T.float32(0)
compute_2[cse_var_1 + 2] = T.float32(0)
@@ -659,47 +659,47 @@ class Module:
compute_2[cse_var_1 + 13] = T.float32(0)
compute_2[cse_var_1 + 14] = T.float32(0)
compute_2[cse_var_1 + 15] = T.float32(0)
- for elem_idx, i_inner in T.grid(T.Let(placeholder_5[cse_var_2 + 1] - placeholder_5[cse_var_2], where={cse_var_2: i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner}), 32):
+ for elem_idx, i_inner in T.grid(T.Let(placeholder_5[cse_var_2 + 1] - placeholder_5[cse_var_2], where={cse_var_2: i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner}), 64):
cse_var_2 = T.int32()
placeholder_5 = T.Buffer((33,), "int32", data=placeholder_3.data)
cse_var_21: T.int32 = elem_idx * 16
- cse_var_20: T.int32 = i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner
- cse_var_19: T.int32 = i_outer_inner * 1024 + i_inner * 32 + nb_j_inner * 16
- cse_var_18: T.int32 = i0_outer_i1_outer_fused // 16 * 16384 + i_outer_inner * 8192 + i_inner * 256
- cse_var_17: T.int32 = cse_var_19 + 9
- cse_var_16: T.int32 = cse_var_19 + 8
- cse_var_15: T.int32 = cse_var_19 + 7
- cse_var_14: T.int32 = cse_var_19 + 6
- cse_var_13: T.int32 = cse_var_19 + 5
- cse_var_12: T.int32 = cse_var_19 + 4
- cse_var_11: T.int32 = cse_var_19 + 3
- cse_var_10: T.int32 = cse_var_19 + 2
- cse_var_9: T.int32 = cse_var_19 + 15
- cse_var_8: T.int32 = cse_var_19 + 14
- cse_var_7: T.int32 = cse_var_19 + 13
- cse_var_6: T.int32 = cse_var_19 + 12
- cse_var_5: T.int32 = cse_var_19 + 11
- cse_var_4: T.int32 = cse_var_19 + 10
- cse_var_3: T.int32 = cse_var_19 + 1
+ cse_var_20: T.int32 = i_inner * 32 + nb_j_inner * 16
+ cse_var_19: T.int32 = i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner
+ cse_var_18: T.int32 = i0_outer_i1_outer_fused // 16 * 16384 + i_inner * 256
+ cse_var_17: T.int32 = cse_var_20 + 9
+ cse_var_16: T.int32 = cse_var_20 + 8
+ cse_var_15: T.int32 = cse_var_20 + 7
+ cse_var_14: T.int32 = cse_var_20 + 6
+ cse_var_13: T.int32 = cse_var_20 + 5
+ cse_var_12: T.int32 = cse_var_20 + 4
+ cse_var_11: T.int32 = cse_var_20 + 3
+ cse_var_10: T.int32 = cse_var_20 + 2
+ cse_var_9: T.int32 = cse_var_20 + 15
+ cse_var_8: T.int32 = cse_var_20 + 14
+ cse_var_7: T.int32 = cse_var_20 + 13
+ cse_var_6: T.int32 = cse_var_20 + 12
+ cse_var_5: T.int32 = cse_var_20 + 11
+ cse_var_4: T.int32 = cse_var_20 + 10
+ cse_var_3: T.int32 = cse_var_20 + 1
placeholder_6 = T.Buffer((78656,), data=placeholder_1.data)
placeholder_7 = T.Buffer((32768,), data=placeholder.data)
placeholder_8 = T.Buffer((4916,), "int32", data=placeholder_2.data)
- compute_2[cse_var_19] = compute_2[cse_var_19] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_3] = compute_2[cse_var_3] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 1] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_10] = compute_2[cse_var_10] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 2] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_11] = compute_2[cse_var_11] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 3] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_12] = compute_2[cse_var_12] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 4] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_13] = compute_2[cse_var_13] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 5] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_14] = compute_2[cse_var_14] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 6] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_15] = compute_2[cse_var_15] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 7] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_16] = compute_2[cse_var_16] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 8] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_17] = compute_2[cse_var_17] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 9] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_4] = compute_2[cse_var_4] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 10] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_5] = compute_2[cse_var_5] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 11] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_6] = compute_2[cse_var_6] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 12] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_7] = compute_2[cse_var_7] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 13] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_8] = compute_2[cse_var_8] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 14] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
- compute_2[cse_var_9] = compute_2[cse_var_9] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21 + 15] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
+ compute_2[cse_var_20] = compute_2[cse_var_20] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_3] = compute_2[cse_var_3] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 1] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_10] = compute_2[cse_var_10] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 2] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_11] = compute_2[cse_var_11] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 3] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_12] = compute_2[cse_var_12] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 4] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_13] = compute_2[cse_var_13] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 5] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_14] = compute_2[cse_var_14] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 6] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_15] = compute_2[cse_var_15] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 7] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_16] = compute_2[cse_var_16] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 8] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_17] = compute_2[cse_var_17] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 9] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_4] = compute_2[cse_var_4] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 10] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_5] = compute_2[cse_var_5] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 11] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_6] = compute_2[cse_var_6] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 12] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_7] = compute_2[cse_var_7] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 13] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_8] = compute_2[cse_var_8] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 14] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
+ compute_2[cse_var_9] = compute_2[cse_var_9] + placeholder_6[placeholder_5[cse_var_19] * 16 + cse_var_21 + 15] * T.max(placeholder_7[cse_var_18 + placeholder_8[placeholder_5[cse_var_19] + elem_idx]], T.float32(0))
for i0_inner in range(64):
cse_var_22: T.int32 = i0_outer_i1_outer_fused // 16 * 32768 + i0_inner * 512 + i0_outer_i1_outer_fused % 16 * 32
compute_3 = T.Buffer((65536,), data=compute.data)
@@ -738,7 +738,7 @@ class Module:
<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: 1.728 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 47422496c5..882bee7129 100644
--- a/docs/how_to/tune_with_autotvm/sg_execution_times.html
+++ b/docs/how_to/tune_with_autotvm/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>00:33.815</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
+<p><strong>00:43.909</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -354,7 +354,7 @@
</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>00:33.781</p></td>
+<td><p>00:43.874</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>
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 bd74e0172c..5a4670112b 100644
--- a/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
+++ b/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
@@ -695,655 +695,8 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5021072
-No: 2 GFLOPS: 90.55/90.55 result: MeasureResult(costs=(0.002556609098360656,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.1473910808563232, timestamp=1679006697.4090955) [('tile_f', [-1, 1, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8254840
-No: 3 GFLOPS: 0.00/90.55 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 742, in __call__
- yield remote, remote.load_module(os.path.split(build_result.filename)[1])
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 706, in run_through_rpc
- costs = time_f(*args).results
- File "/workspace/python/tvm/runtime/module.py", line 357, in evaluator
- blob = feval(*args)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 262, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./packed_func.pxi", line 251, in tvm._ffi._cy3.core.FuncCall3
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
-tvm._ffi.base.TVMError: Traceback (most recent call last):
- 4: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 3: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 2: tvm::runtime::RPCWrappedFunc::operator()(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../src/runtime/rpc/rpc_module.cc:129
- 1: tvm::runtime::RPCClientSession::CallFunc(void*, TVMValue const*, int const*, int, std::function<void (tvm::runtime::TVMArgs)> const&)
- at ../src/runtime/rpc/rpc_endpoint.cc:1012
- 0: tvm::runtime::RPCEndpoint::CallFunc(void*, TVMValue const*, int const*, int, std::function<void (tvm::runtime::TVMArgs)>)
- at ../src/runtime/rpc/rpc_endpoint.cc:804
- File "../src/runtime/rpc/rpc_endpoint.cc", line 804
-TVMError:
----------------------------------------------------------------
-An error occurred during the execution of TVM.
-For more information, please see: https://tvm.apache.org/docs/errors.html
----------------------------------------------------------------
- Check failed: (code == RPCCode::kReturn) is false: code=kShutdown
-
-During handling of the above exception, another exception occurred:
-
-Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 706, in run_through_rpc
- costs = time_f(*args).results
- File "/usr/lib/python3.7/contextlib.py", line 130, in __exit__
- self.gen.throw(type, value, traceback)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 746, in __call__
- remote.remove(build_result.filename)
- File "/workspace/python/tvm/rpc/client.py", line 144, in remove
- self._remote_funcs["remove"] = self.get_function("tvm.rpc.server.remove")
- File "/workspace/python/tvm/rpc/client.py", line 72, in get_function
- return self._sess.get_function(name)
- File "/workspace/python/tvm/runtime/module.py", line 171, in get_function
- self.handle, c_str(name), ctypes.c_int(query_imports), ctypes.byref(ret_handle)
- File "/workspace/python/tvm/_ffi/base.py", line 348, in check_call
- raise get_last_ffi_error()
-tvm._ffi.base.TVMError: Traceback (most recent call last):
- 52: 0xffffffffffffffff
- 51: _start
- 50: __libc_start_main
- 49: _Py_UnixMain
- 48: 0x0000000000650da0
- 47: 0x0000000000650afa
- 46: _PyFunction_FastCallDict
- 45: _PyEval_EvalCodeWithName
- 44: _PyEval_EvalFrameDefault
- 43: _PyFunction_FastCallKeywords
- 42: _PyEval_EvalCodeWithName
- 41: _PyEval_EvalFrameDefault
- 40: _PyMethodDef_RawFastCallKeywords
- 39: 0x0000000000546369
- 38: _PyEval_EvalCodeWithName
- 37: _PyEval_EvalFrameDefault
- 36: _PyFunction_FastCallKeywords
- 35: _PyEval_EvalCodeWithName
- 34: _PyEval_EvalFrameDefault
- 33: _PyFunction_FastCallDict
- 32: _PyEval_EvalCodeWithName
- 31: _PyEval_EvalFrameDefault
- 30: _PyObject_FastCallDict
- 29: 0x00000000004c06e1
- 28: _PyFunction_FastCallDict
- 27: _PyEval_EvalFrameDefault
- 26: _PyMethodDescr_FastCallKeywords
- 25: 0x00000000005dcb58
- 24: 0x00000000005dc83f
- 23: 0x00000000004ba127
- 22: _PyEval_EvalFrameDefault
- 21: _PyFunction_FastCallKeywords
- 20: _PyEval_EvalFrameDefault
- 19: _PyFunction_FastCallKeywords
- 18: _PyEval_EvalFrameDefault
- 17: _PyFunction_FastCallKeywords
- 16: _PyEval_EvalCodeWithName
- 15: _PyEval_EvalFrameDefault
- 14: 0x0000000000537c30
- 13: _PyObject_FastCallKeywords
- 12: 0x00007f972211dfa2
- 11: _ctypes_callproc
- 10: ffi_call
- 9: ffi_call_unix64
- 8: TVMModGetFunction
- at ../src/runtime/c_runtime_api.cc:408
- 7: tvm::runtime::ModuleNode::GetFunction(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, bool)
- at ../src/runtime/module.cc:66
- 6: tvm::runtime::RPCModuleNode::GetFunction(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, tvm::runtime::ObjectPtr<tvm::runtime::Object> const&)
- at ../src/runtime/rpc/rpc_module.cc:185
- 5: tvm::runtime::RPCClientSession::GetFunction(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)
- at ../src/runtime/rpc/rpc_endpoint.cc:1007
- 4: tvm::runtime::TVMRetValue tvm::runtime::RPCEndpoint::SysCallRemote<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&>(tvm::runtime::RPCCode, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)
- at ../src/runtime/rpc/rpc_endpoint.h:223
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&>(int&&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/rpc/rpc_endpoint.cc:684
- File "../src/runtime/rpc/rpc_endpoint.cc", line 684
-TVMError:
----------------------------------------------------------------
-An error occurred during the execution of TVM.
-For more information, please see: https://tvm.apache.org/docs/errors.html
----------------------------------------------------------------
- Check failed: (code == RPCCode::kReturn) is false: code=1
-
-Traceback (most recent call last):
- 52: 0xffffffffffffffff
- 51: _start
- 50: __libc_start_main
- 49: _Py_UnixMain
- 48: 0x0000000000650da0
- 47: 0x0000000000650afa
- 46: _PyFunction_FastCallDict
- 45: _PyEval_EvalCodeWithName
- 44: _PyEval_EvalFrameDefault
- 43: _PyFunction_FastCallKeywords
- 42: _PyEval_EvalCodeWithName
- 41: _PyEval_EvalFrameDefault
- 40: _PyMethodDef_RawFastCallKeywords
- 39: 0x0000000000546369
- 38: _PyEval_EvalCodeWithName
- 37: _PyEval_EvalFrameDefault
- 36: _PyFunction_FastCallKeywords
- 35: _PyEval_EvalCodeWithName
- 34: _PyEval_EvalFrameDefault
- 33: _PyFunction_FastCallDict
- 32: _PyEval_EvalCodeWithName
- 31: _PyEval_EvalFrameDefault
- 30: _PyObject_FastCallDict
- 29: 0x00000000004c06e1
- 28: _PyFunction_FastCallDict
- 27: _PyEval_EvalFrameDefault
- 26: _PyMethodDescr_FastCallKeywords
- 25: 0x00000000005dcb58
- 24: 0x00000000005dc83f
- 23: 0x00000000004ba127
- 22: _PyEval_EvalFrameDefault
- 21: _PyFunction_FastCallKeywords
- 20: _PyEval_EvalFrameDefault
- 19: _PyFunction_FastCall [('tile_f', [-1, 64, 1, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1590266
-No: 4 GFLOPS: 0.00/90.55 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
- func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
- func = build(s, args, target=target, runtime=runtime)
- File "/workspace/python/tvm/driver/build_module.py", line 227, in build
- input_mod = lower(inputs, args, name=name, binds=binds)
- File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
- return ffi.lower_schedule(inp, args, name, binds, simple_mode)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
-tvm._ffi.base.TVMError: Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
- raise InstantiationError("Skipped because of invalid gpu kernel")
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
-
-Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 32]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4604973
-No: 5 GFLOPS: 0.00/90.55 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
- func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
- func = build(s, args, target=target, runtime=runtime)
- File "/workspace/python/tvm/driver/build_module.py", line 227, in build
- input_mod = lower(inputs, args, name=name, binds=binds)
- File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
- return ffi.lower_schedule(inp, args, name, binds, simple_mode)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
-tvm._ffi.base.TVMError: Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
- raise InstantiationError("Skipped because of invalid gpu kernel")
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
-
-Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 128, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4284533
-No: 6 GFLOPS: 0.00/90.55 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
- func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
- func = build(s, args, target=target, runtime=runtime)
- File "/workspace/python/tvm/driver/build_module.py", line 227, in build
- input_mod = lower(inputs, args, name=name, binds=binds)
- File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
- return ffi.lower_schedule(inp, args, name, binds, simple_mode)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
-tvm._ffi.base.TVMError: Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
- raise InstantiationError("Skipped because of invalid gpu kernel")
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
-
-Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 8, 8, 8]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4163217
-No: 7 GFLOPS: 222.61/222.61 result: MeasureResult(costs=(0.0010399281666666667,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.479459047317505, timestamp=1679006708.9954762) [('tile_f', [-1, 4, 4, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4304154
-No: 8 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
- func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
- func = build(s, args, target=target, runtime=runtime)
- File "/workspace/python/tvm/driver/build_module.py", line 227, in build
- input_mod = lower(inputs, args, name=name, binds=binds)
- File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
- return ffi.lower_schedule(inp, args, name, binds, simple_mode)
- File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
- File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
- File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
-tvm._ffi.base.TVMError: Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
- raise InstantiationError("Skipped because of invalid gpu kernel")
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
-
-Traceback (most recent call last):
- 24: TVMFuncCall
- at ../src/runtime/c_runtime_api.cc:477
- 23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 22: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 21: operator()
- at ../include/tvm/runtime/packed_func.h:1734
- 20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
- at ../include/tvm/runtime/packed_func.h:1674
- 19: run<>
- at ../include/tvm/runtime/packed_func.h:1634
- 18: run<tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1634
- 14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
- at ../include/tvm/runtime/packed_func.h:1649
- 13: operator()
- at ../src/driver/driver_api.cc:402
- 12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
- at ../src/driver/driver_api.cc:388
- 11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
- at ../src/driver/driver_api.cc:283
- 10: tvm::transform::Pass::operator()(tvm::IRModule) const
- at ../src/ir/transform.cc:258
- 9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:451
- 7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/ir/transform.cc:274
- 6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
- at ../src/tir/ir/transform.cc:100
- 5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
- at ../include/tvm/runtime/packed_func.h:1753
- 4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
- at ../include/tvm/runtime/packed_func.h:1697
- 3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
- at ../include/tvm/runtime/packed_func.h:1621
- 2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
- at ../include/tvm/runtime/packed_func.h:1217
- 1: Call
- at ../include/tvm/runtime/packed_func.h:1213
- 0: operator()
- at ../src/runtime/c_runtime_api.cc:534
- File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 32, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 256, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7420641
-No: 9 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 16, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 512, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7003199
+No: 2 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1465,8 +818,8 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 1, 128]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 8, 64]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9851811
-No: 10 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 128, 1]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 256, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8195490
+No: 3 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1588,8 +941,8 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 1, 512]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 32]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,1887379
-No: 11 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 4, 32]), ('tile_y', [-1, 7, 1, 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', 0), ('unroll_explicit', 1)],None,5966815
+No: 4 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1711,8 +1064,9 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 4, 64]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 16]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8060787
-No: 12 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 1, 128]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 64, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,3768371
+No: 5 GFLOPS: 64.38/64.38 result: MeasureResult(costs=(0.003595801525,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.958209753036499, timestamp=1679043027.2978892) [('tile_f', [-1, 8, 8, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4103962
+No: 6 GFLOPS: 0.00/64.38 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1834,8 +1188,8 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 8, 4]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 8, 16]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6133285
-No: 13 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 16, 4, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 16, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9538473
+No: 7 GFLOPS: 0.00/64.38 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -1957,8 +1311,10 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 256, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4634399
-No: 14 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 2, 32]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2696950
+No: 8 GFLOPS: 413.24/413.24 result: MeasureResult(costs=(0.0005602043184357541,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.21183180809021, timestamp=1679043028.17347) [('tile_f', [-1, 8, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9141003
+No: 9 GFLOPS: 130.28/413.24 result: MeasureResult(costs=(0.0017769945714285714,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7461109161376953, timestamp=1679043034.4908655) [('tile_f', [-1, 2, 32, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5491991
+No: 10 GFLOPS: 0.00/413.24 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2080,8 +1436,9 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 32, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8529752
-No: 15 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 128, 2, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 512]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,385731
+No: 11 GFLOPS: 31.12/413.24 result: MeasureResult(costs=(0.0074379817142857145,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.8672966957092285, timestamp=1679043035.3957975) [('tile_f', [-1, 8, 4, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7647352
+No: 12 GFLOPS: 0.00/413.24 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2203,8 +1560,8 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 32, 4]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 512]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,190431
-No: 16 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 1, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8262046
+No: 13 GFLOPS: 0.00/413.24 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2326,8 +1683,8 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 8, 8, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 32, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1374150
-No: 17 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 2, 256]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 256, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,3744838
+No: 14 GFLOPS: 0.00/413.24 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2449,8 +1806,10 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 16, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,3274074
-No: 18 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 64, 2, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 64]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4040756
+No: 15 GFLOPS: 817.74/817.74 result: MeasureResult(costs=(0.00028309882245430813,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.182854175567627, timestamp=1679043039.8110738) [('tile_f', [-1, 1, 2, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 8]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7843944
+No: 16 GFLOPS: 374.30/817.74 result: MeasureResult(costs=(0.0006184928658536585,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2676329612731934, timestamp=1679043040.6763678) [('tile_f', [-1, 1, 32, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,3721170
+No: 17 GFLOPS: 0.00/817.74 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2572,8 +1931,9 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 2, 4, 64]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 256]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,3091648
-No: 19 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 4, 128]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7338315
+No: 18 GFLOPS: 4.06/817.74 result: MeasureResult(costs=(0.0570745505,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.9982800483703613, timestamp=1679043042.8743176) [('tile_f', [-1, 4, 4, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5425817
+No: 19 GFLOPS: 0.00/817.74 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2695,8 +2055,8 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 256, 1]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,3293892
-No: 20 GFLOPS: 0.00/222.61 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 1, 8]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 64]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6750177
+No: 20 GFLOPS: 0.00/817.74 result: Traceback (most recent call last):
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
@@ -2818,7 +2178,7 @@ Traceback (most recent call last):
File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, 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, 1, 2, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 8, 64]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6171504
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 8, 1, 32]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 8, 32]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4994188
</pre></div>
</div>
<p>Finally we can inspect the best config from log file, check correctness,
@@ -2857,9 +2217,9 @@ and measure running time.</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Finish loading 20 records
Best config:
-[('tile_f', [-1, 4, 4, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4304154
+[('tile_f', [-1, 1, 2, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 8]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7843944
Finish loading 20 records
-Time cost of this operator: 0.001350
+Time cost of this operator: 0.000578
</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 b8793bc6fc..a0ddd31fa6 100644
--- a/docs/how_to/work_with_microtvm/micro_autotune.html
+++ b/docs/how_to/work_with_microtvm/micro_autotune.html
@@ -649,10 +649,10 @@ the tuned operator.</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>########## 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 315.9 98.731 (1, 2, 10, 10, 3) 2 1 [315.9]
-tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.104 0.97 (1, 6, 10, 10) 1 1 [3.104]
-tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.955 0.299 (1, 1, 10, 10, 3) 1 1 [0.955]
-Total_time - 319.959 - - - - -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 316.2 98.746 (1, 2, 10, 10, 3) 2 1 [316.2]
+tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.042 0.95 (1, 6, 10, 10) 1 1 [3.042]
+tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.974 0.304 (1, 1, 10, 10, 3) 1 1 [0.974]
+Total_time - 320.215 - - - - -
</pre></div>
</div>
</div>
@@ -704,13 +704,13 @@ Total_time -
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>########## 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 103.8 97.509 (1, 6, 10, 10, 1) 2 1 [103.8]
-tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.802 1.692 (1, 6, 10, 10) 1 1 [1.802]
-tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.85 0.798 (1, 3, 10, 10, 1) 1 1 [0.85]
-Total_time - 106.451 - - - - -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 103.4 97.512 (1, 6, 10, 10, 1) 2 1 [103.4]
+tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.781 1.679 (1, 6, 10, 10) 1 1 [1.781]
+tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.857 0.808 (1, 3, 10, 10, 1) 1 1 [0.857]
+Total_time - 106.038 - - - - -
</pre></div>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 20.489 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 24.983 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-work-with-microtvm-micro-autotune-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/9ccca8fd489a1486ac71b55a55c320c5/micro_autotune.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_autotune.py</span></code></a></p>
diff --git a/docs/how_to/work_with_microtvm/micro_pytorch.html b/docs/how_to/work_with_microtvm/micro_pytorch.html
index 9ff6b26bc4..273e57c9f0 100644
--- a/docs/how_to/work_with_microtvm/micro_pytorch.html
+++ b/docs/how_to/work_with_microtvm/micro_pytorch.html
@@ -460,8 +460,7 @@ download a cat image and preprocess it to use as the model input.</p>
Downloading: "https://download.pytorch.org/models/quantized/mobilenet_v2_qnnpack_37f702c5.pth" to /workspace/.cache/torch/hub/checkpoints/mobilenet_v2_qnnpack_37f702c5.pth
0%| | 0.00/3.42M [00:00<?, ?B/s]
- 61%|###### | 2.09M/3.42M [00:00<00:00, 8.07MB/s]
-100%|##########| 3.42M/3.42M [00:00<00:00, 13.0MB/s]
+100%|##########| 3.42M/3.42M [00:00<00:00, 50.5MB/s]
/workspace/python/tvm/relay/frontend/pytorch_utils.py:47: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
return LooseVersion(torch_ver) > ver
/venv/apache-tvm-py3.7/lib/python3.7/site-packages/setuptools/_distutils/version.py:346: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
@@ -587,7 +586,7 @@ via the host <cite>main.cc`</cite> or if a Zephyr emulated board is selected as
Torch top-1 id: 282, class name: tiger cat
</pre></div>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 18.097 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 22.427 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-work-with-microtvm-micro-pytorch-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/12b9ecc04c41abaa12022061771821d1/micro_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">micro_pytorch.py</span></code></a></p>
diff --git a/docs/how_to/work_with_microtvm/micro_train.html b/docs/how_to/work_with_microtvm/micro_train.html
index 375578b5df..528b39ca6b 100644
--- a/docs/how_to/work_with_microtvm/micro_train.html
+++ b/docs/how_to/work_with_microtvm/micro_train.html
@@ -529,7 +529,7 @@ 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/tmpisplwp6d/images/random'
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>'/tmp/tmpwvh5kmcm/images/random'
</pre></div>
</div>
</div>
@@ -589,8 +589,8 @@ 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], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [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/tmpisplwp6d/images/target contains 8144 images
-/tmp/tmpisplwp6d/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], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0]" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>/tmp/tmpwvh5kmcm/images/target contains 8144 images
+/tmp/tmpwvh5kmcm/images/random contains 5000 images
</pre></div>
</div>
</div>
@@ -702,13 +702,13 @@ 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 - 48s - loss: 0.2149 - accuracy: 0.9244 - val_loss: 0.1410 - val_accuracy: 0.9532 - 48s/epoch - 146ms/step
+328/328 - 48s - loss: 0.2170 - accuracy: 0.9255 - val_loss: 0.1736 - val_accuracy: 0.9452 - 48s/epoch - 146ms/step
Epoch 2/3
-328/328 - 44s - loss: 0.1080 - accuracy: 0.9608 - val_loss: 0.1752 - val_accuracy: 0.9392 - 44s/epoch - 133ms/step
+328/328 - 44s - loss: 0.0959 - accuracy: 0.9645 - val_loss: 0.1144 - val_accuracy: 0.9603 - 44s/epoch - 133ms/step
Epoch 3/3
-328/328 - 43s - loss: 0.0610 - accuracy: 0.9776 - val_loss: 0.1257 - val_accuracy: 0.9535 - 43s/epoch - 132ms/step
+328/328 - 47s - loss: 0.0731 - accuracy: 0.9731 - val_loss: 0.1512 - val_accuracy: 0.9551 - 47s/epoch - 143ms/step
-<keras.callbacks.History object at 0x7fa6d4bbb950>
+<keras.callbacks.History object at 0x7fed646bcf90>
</pre></div>
</div>
</div>
@@ -972,7 +972,7 @@ as intended.</p>
<p>From here, we could modify the model to read live images from the camera - we have another
Arduino tutorial for how to do that <a class="reference external" href="https://github.com/guberti/tvm-arduino-demos/tree/master/examples/person_detection">on GitHub</a>. Alternatively, we could also
<a class="reference external" href="https://tvm.apache.org/docs/how_to/work_with_microtvm/micro_autotune.html">use TVM’s autotuning capabilities</a> to dramatically improve the model’s performance.</p>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 4 minutes 42.118 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 4 minutes 51.956 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 5f8d8c0343..a4cc44367e 100644
--- a/docs/how_to/work_with_microtvm/sg_execution_times.html
+++ b/docs/how_to/work_with_microtvm/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>07:45.881</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
+<p><strong>08:06.016</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 82%" />
@@ -354,27 +354,27 @@
</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">5. Training Vision Models for microTVM on Arduino</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_train.py</span></code>)</p></td>
-<td><p>04:42.118</p></td>
+<td><p>04:51.956</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">6. Model Tuning with microTVM</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_autotune.py</span></code>)</p></td>
-<td><p>01:20.489</p></td>
+<td><p>01:24.983</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="micro_pytorch.html#sphx-glr-how-to-work-with-microtvm-micro-pytorch-py"><span class="std std-ref">4. microTVM PyTorch Tutorial</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_pytorch.py</span></code>)</p></td>
-<td><p>01:18.097</p></td>
+<td><p>01:22.427</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><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">3. microTVM Ahead-of-Time (AOT) Compilation</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_aot.py</span></code>)</p></td>
-<td><p>00:10.158</p></td>
+<td><p>00:10.690</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="micro_custom_ide.html#sphx-glr-how-to-work-with-microtvm-micro-custom-ide-py"><span class="std std-ref">9. Bring microTVM to your own development environment</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_custom_ide.py</span></code>)</p></td>
-<td><p>00:07.843</p></td>
+<td><p>00:08.286</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">2. microTVM TFLite Tutorial</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_tflite.py</span></code>)</p></td>
-<td><p>00:07.177</p></td>
+<td><p>00:07.673</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">7. Running TVM on bare metal Arm(R) Cortex(R)-M55 CPU and Ethos(TM)-U55 NPU with CMSIS-NN</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_ethosu.py</span></code>)</p></td>
diff --git a/docs/how_to/work_with_relay/sg_execution_times.html b/docs/how_to/work_with_relay/sg_execution_times.html
index dfdc09158c..a849b4d16c 100644
--- a/docs/how_to/work_with_relay/sg_execution_times.html
+++ b/docs/how_to/work_with_relay/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>00:46.294</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
+<p><strong>00:48.171</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -354,15 +354,15 @@
</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>00:34.066</p></td>
+<td><p>00:35.830</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>00:10.478</p></td>
+<td><p>00:10.696</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>00:01.744</p></td>
+<td><p>00:01.639</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 44013fc603..816241307b 100644
--- a/docs/how_to/work_with_schedules/intrin_math.html
+++ b/docs/how_to/work_with_schedules/intrin_math.html
@@ -540,7 +540,7 @@ 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 0x7fa587119e60>
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span><function my_cuda_math_rule at 0x7fec19fa0a70>
</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 e4779a4a54..9599aea50a 100644
--- a/docs/how_to/work_with_schedules/sg_execution_times.html
+++ b/docs/how_to/work_with_schedules/sg_execution_times.html
@@ -345,7 +345,7 @@
<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>00:08.367</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
+<p><strong>00:07.806</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -354,35 +354,35 @@
</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>00:05.758</p></td>
+<td><p>00:05.107</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>00:01.214</p></td>
+<td><p>00:01.227</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>00:00.579</p></td>
+<td><p>00:00.615</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>00:00.560</p></td>
+<td><p>00:00.589</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="extern_op.html#sphx-glr-how-to-work-with-schedules-extern-op-py"><span class="std std-ref">External Tensor Functions</span></a> (<code class="docutils literal notranslate"><span class="pre">extern_op.py</span></code>)</p></td>
-<td><p>00:00.116</p></td>
+<td><p>00:00.123</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><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>00:00.060</p></td>
+<td><p>00:00.062</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="schedule_primitives.html#sphx-glr-how-to-work-with-schedules-schedule-primitives-py"><span class="std std-ref">Schedule Primitives in TVM</span></a> (<code class="docutils literal notranslate"><span class="pre">schedule_primitives.py</span></code>)</p></td>
-<td><p>00:00.052</p></td>
+<td><p>00:00.056</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>00:00.026</p></td>
+<td><p>00:00.028</p></td>
<td><p>0.0 MB</p></td>
</tr>
</tbody>
diff --git a/docs/reference/api/python/auto_scheduler.html b/docs/reference/api/python/auto_scheduler.html
index d2bd610df0..601546b086 100644
--- a/docs/reference/api/python/auto_scheduler.html
+++ b/docs/reference/api/python/auto_scheduler.html
@@ -1621,7 +1621,7 @@ history states as starting point to perform Evolutionary Search).</p></li>
<dl class="py class">
<dt class="sig sig-object py" id="tvm.auto_scheduler.SketchPolicy">
-<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">tvm.auto_scheduler.</span></span><span class="sig-name descname"><span class="pre">SketchPolicy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">program_cost_model</span></span><span class="o"><span class="pre">=</span></span><span class="defau [...]
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">tvm.auto_scheduler.</span></span><span class="sig-name descname"><span class="pre">SketchPolicy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">program_cost_model</span></span><span class="o"><span class="pre">=</span></span><span class="defau [...]
<dd><p>The search policy that searches in a hierarchical search space defined by sketches.
The policy randomly samples programs from the space defined by sketches and use evolutionary
search to fine-tune them.</p>
@@ -1905,7 +1905,7 @@ Candidates:
<dl class="py function">
<dt class="sig sig-object py" id="tvm.auto_scheduler.auto_schedule">
-<span class="sig-prename descclassname"><span class="pre">tvm.auto_scheduler.</span></span><span class="sig-name descname"><span class="pre">auto_schedule</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_policy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em clas [...]
+<span class="sig-prename descclassname"><span class="pre">tvm.auto_scheduler.</span></span><span class="sig-name descname"><span class="pre">auto_schedule</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_policy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em clas [...]
<dd><p>THIS API IS DEPRECATED.</p>
<p>Run auto scheduling search for a task.</p>
<dl class="field-list simple">
diff --git a/docs/reference/api/typedoc/classes/bytestreamreader.html b/docs/reference/api/typedoc/classes/bytestreamreader.html
index 6cf98a4645..2201efd68f 100644
--- a/docs/reference/api/typedoc/classes/bytestreamreader.html
+++ b/docs/reference/api/typedoc/classes/bytestreamreader.html
@@ -119,7 +119,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L45">rpc_server.ts:45</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L45">rpc_server.ts:45</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -141,7 +141,7 @@
<div class="tsd-signature tsd-kind-icon">bytes<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Uint8Array</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L45">rpc_server.ts:45</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L45">rpc_server.ts:45</a></li>
</ul>
</aside>
</section>
@@ -151,7 +151,7 @@
<div class="tsd-signature tsd-kind-icon">offset<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span><span class="tsd-signature-symbol"> = 0</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L44">rpc_server.ts:44</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L44">rpc_server.ts:44</a></li>
</ul>
</aside>
</section>
@@ -168,7 +168,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L65">rpc_server.ts:65</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L65">rpc_server.ts:65</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">Uint8Array</span></h4>
@@ -185,7 +185,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L51">rpc_server.ts:51</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L51">rpc_server.ts:51</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
@@ -202,7 +202,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L59">rpc_server.ts:59</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L59">rpc_server.ts:59</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
diff --git a/docs/reference/api/typedoc/classes/cachedcallstack.html b/docs/reference/api/typedoc/classes/cachedcallstack.html
index dcd862856c..104290b698 100644
--- a/docs/reference/api/typedoc/classes/cachedcallstack.html
+++ b/docs/reference/api/typedoc/classes/cachedcallstack.html
@@ -144,7 +144,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L223">memory.ts:223</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L223">memory.ts:223</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -172,7 +172,7 @@
<div class="tsd-signature tsd-kind-icon">temp<wbr>Args<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Array</span><span class="tsd-signature-symbol"><</span><a href="../interfaces/disposable.html" class="tsd-signature-type">Disposable</a><span class="tsd-signature-symbol">></span><span class="tsd-signature-symbol"> = []</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L208">memory.ts:208</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L208">memory.ts:208</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -194,7 +194,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L312">memory.ts:312</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L312">memory.ts:312</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -226,7 +226,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L284">memory.ts:284</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L284">memory.ts:284</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -262,7 +262,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L388">memory.ts:388</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L388">memory.ts:388</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -300,7 +300,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L376">memory.ts:376</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L376">memory.ts:376</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -340,7 +340,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L267">memory.ts:267</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L267">memory.ts:267</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -373,7 +373,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L243">memory.ts:243</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L243">memory.ts:243</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -390,7 +390,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L321">memory.ts:321</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L321">memory.ts:321</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -422,7 +422,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L252">memory.ts:252</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L252">memory.ts:252</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -444,7 +444,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L359">memory.ts:359</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L359">memory.ts:359</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -470,7 +470,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L342">memory.ts:342</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L342">memory.ts:342</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -496,7 +496,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L350">memory.ts:350</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L350">memory.ts:350</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -522,7 +522,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L326">memory.ts:326</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L326">memory.ts:326</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -548,7 +548,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L363">memory.ts:363</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L363">memory.ts:363</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -574,7 +574,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L346">memory.ts:346</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L346">memory.ts:346</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -600,7 +600,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L334">memory.ts:334</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L334">memory.ts:334</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
diff --git a/docs/reference/api/typedoc/classes/dldatatype.html b/docs/reference/api/typedoc/classes/dldatatype.html
index 8b4dd8a4d7..c760e18de8 100644
--- a/docs/reference/api/typedoc/classes/dldatatype.html
+++ b/docs/reference/api/typedoc/classes/dldatatype.html
@@ -119,7 +119,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L359">runtime.ts:359</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L359">runtime.ts:359</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -147,7 +147,7 @@
<div class="tsd-signature tsd-kind-icon">bits<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L357">runtime.ts:357</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L357">runtime.ts:357</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -162,7 +162,7 @@
<div class="tsd-signature tsd-kind-icon">code<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L355">runtime.ts:355</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L355">runtime.ts:355</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -177,7 +177,7 @@
<div class="tsd-signature tsd-kind-icon">lanes<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L359">runtime.ts:359</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L359">runtime.ts:359</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -199,7 +199,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L376">runtime.ts:376</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L376">runtime.ts:376</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
@@ -216,7 +216,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L367">runtime.ts:367</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L367">runtime.ts:367</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">string</span></h4>
diff --git a/docs/reference/api/typedoc/classes/dldevice.html b/docs/reference/api/typedoc/classes/dldevice.html
index fa4b8f4966..b1c802babc 100644
--- a/docs/reference/api/typedoc/classes/dldevice.html
+++ b/docs/reference/api/typedoc/classes/dldevice.html
@@ -118,7 +118,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L299">runtime.ts:299</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L299">runtime.ts:299</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -146,7 +146,7 @@
<div class="tsd-signature tsd-kind-icon">device<wbr>Id<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L297">runtime.ts:297</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L297">runtime.ts:297</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -161,7 +161,7 @@
<div class="tsd-signature tsd-kind-icon">device<wbr>Type<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L295">runtime.ts:295</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L295">runtime.ts:295</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -183,7 +183,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L320">runtime.ts:320</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L320">runtime.ts:320</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -205,7 +205,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L327">runtime.ts:327</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L327">runtime.ts:327</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">string</span></h4>
diff --git a/docs/reference/api/typedoc/classes/environment.html b/docs/reference/api/typedoc/classes/environment.html
index 77c8bd62c3..6bc5f48047 100644
--- a/docs/reference/api/typedoc/classes/environment.html
+++ b/docs/reference/api/typedoc/classes/environment.html
@@ -125,7 +125,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/environment.ts#L86">environment.ts:86</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/environment.ts#L86">environment.ts:86</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -169,7 +169,7 @@
<aside class="tsd-sources">
<p>Implementation of <a href="../interfaces/libraryprovider.html">LibraryProvider</a>.<a href="../interfaces/libraryprovider.html#imports">imports</a></p>
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/environment.ts#L70">environment.ts:70</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/environment.ts#L70">environment.ts:70</a></li>
</ul>
</aside>
</section>
@@ -179,7 +179,7 @@
<div class="tsd-signature tsd-kind-icon">logger<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-symbol">(</span>msg<span class="tsd-signature-symbol">: </span><span class="tsd-signature-type">string</span><span class="tsd-signature-symbol">)</span><span class="tsd-signature-symbol"> => </span><span class="tsd-signature-type">void</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/environment.ts#L69">environment.ts:69</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/environment.ts#L69">environment.ts:69</a></li>
</ul>
</aside>
<div class="tsd-type-declaration">
@@ -210,7 +210,7 @@
<div class="tsd-signature tsd-kind-icon">packedCFunc<wbr>Table<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Array</span><span class="tsd-signature-symbol"><</span><span class="tsd-signature-type">ctypes.FTVMWasmPackedCFunc</span><span class="tsd-signature-symbol"> | </span><span class="tsd-signature-type">undefined</span><span class="tsd-signature-symbol">></span><span class="tsd-signature-symbol"> = [undefined,]</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/environment.ts#L78">environment.ts:78</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/environment.ts#L78">environment.ts:78</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -228,7 +228,7 @@
<div class="tsd-signature tsd-kind-icon">packedCFunc<wbr>Table<wbr>Free<wbr>Id<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Array</span><span class="tsd-signature-symbol"><</span><span class="tsd-signature-type">number</span><span class="tsd-signature-symbol">></span><span class="tsd-signature-symbol"> = []</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/environment.ts#L84">environment.ts:84</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/environment.ts#L84">environment.ts:84</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -250,7 +250,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/environment.ts#L105">environment.ts:105</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/environment.ts#L105">environment.ts:105</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
diff --git a/docs/reference/api/typedoc/classes/ffilibrary.html b/docs/reference/api/typedoc/classes/ffilibrary.html
index 59b62b381c..69a53a1112 100644
--- a/docs/reference/api/typedoc/classes/ffilibrary.html
+++ b/docs/reference/api/typedoc/classes/ffilibrary.html
@@ -131,7 +131,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L50">runtime.ts:50</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L50">runtime.ts:50</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -156,7 +156,7 @@
<div class="tsd-signature tsd-kind-icon">exports<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Record</span><span class="tsd-signature-symbol"><</span><span class="tsd-signature-type">string</span><span class="tsd-signature-symbol">, </span><span class="tsd-signature-type">Function</span><span class="tsd-signature-symbol">></span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L47">runtime.ts:47</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L47">runtime.ts:47</a></li>
</ul>
</aside>
</section>
@@ -166,7 +166,7 @@
<div class="tsd-signature tsd-kind-icon">memory<span class="tsd-signature-symbol">:</span> <a href="memory.html" class="tsd-signature-type">Memory</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L46">runtime.ts:46</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L46">runtime.ts:46</a></li>
</ul>
</aside>
</section>
@@ -176,7 +176,7 @@
<div class="tsd-signature tsd-kind-icon">wasm32<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">boolean</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L45">runtime.ts:45</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L45">runtime.ts:45</a></li>
</ul>
</aside>
</section>
@@ -186,7 +186,7 @@
<div class="tsd-signature tsd-kind-icon">webGPUContext<span class="tsd-signature-symbol">:</span> <a href="webgpucontext.html" class="tsd-signature-type">WebGPUContext</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L48">runtime.ts:48</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L48">runtime.ts:48</a></li>
</ul>
</aside>
</section>
@@ -203,7 +203,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L77">runtime.ts:77</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L77">runtime.ts:77</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -226,7 +226,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L67">runtime.ts:67</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L67">runtime.ts:67</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -243,7 +243,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L85">runtime.ts:85</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L85">runtime.ts:85</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <a href="cachedcallstack.html" class="tsd-signature-type">CachedCallStack</a></h4>
@@ -260,7 +260,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L96">runtime.ts:96</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L96">runtime.ts:96</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -283,7 +283,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L73">runtime.ts:73</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L73">runtime.ts:73</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
diff --git a/docs/reference/api/typedoc/classes/instance.html b/docs/reference/api/typedoc/classes/instance.html
index d776a8a62b..4ab6955ec3 100644
--- a/docs/reference/api/typedoc/classes/instance.html
+++ b/docs/reference/api/typedoc/classes/instance.html
@@ -161,7 +161,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L844">runtime.ts:844</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L844">runtime.ts:844</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -224,7 +224,7 @@
<div class="tsd-signature tsd-kind-icon">exports<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Record</span><span class="tsd-signature-symbol"><</span><span class="tsd-signature-type">string</span><span class="tsd-signature-symbol">, </span><span class="tsd-signature-type">Function</span><span class="tsd-signature-symbol">></span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L834">runtime.ts:834</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L834">runtime.ts:834</a></li>
</ul>
</aside>
</section>
@@ -234,7 +234,7 @@
<div class="tsd-signature tsd-kind-icon">memory<span class="tsd-signature-symbol">:</span> <a href="memory.html" class="tsd-signature-type">Memory</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L833">runtime.ts:833</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L833">runtime.ts:833</a></li>
</ul>
</aside>
</section>
@@ -251,7 +251,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L973">runtime.ts:973</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L973">runtime.ts:973</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -296,7 +296,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L932">runtime.ts:932</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L932">runtime.ts:932</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -318,7 +318,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L901">runtime.ts:901</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L901">runtime.ts:901</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -381,7 +381,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1215">runtime.ts:1215</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1215">runtime.ts:1215</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -412,7 +412,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1000">runtime.ts:1000</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1000">runtime.ts:1000</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -453,7 +453,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1207">runtime.ts:1207</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1207">runtime.ts:1207</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -491,7 +491,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L922">runtime.ts:922</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L922">runtime.ts:922</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -508,7 +508,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1235">runtime.ts:1235</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1235">runtime.ts:1235</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -552,7 +552,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L943">runtime.ts:943</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L943">runtime.ts:943</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -577,7 +577,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1088">runtime.ts:1088</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1088">runtime.ts:1088</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -609,7 +609,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1363">runtime.ts:1363</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1363">runtime.ts:1363</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -640,7 +640,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1123">runtime.ts:1123</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1123">runtime.ts:1123</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -672,7 +672,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1016">runtime.ts:1016</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1016">runtime.ts:1016</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -695,7 +695,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1281">runtime.ts:1281</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1281">runtime.ts:1281</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -729,7 +729,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L986">runtime.ts:986</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L986">runtime.ts:986</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -769,7 +769,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1341">runtime.ts:1341</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1341">runtime.ts:1341</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -817,7 +817,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1055">runtime.ts:1055</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1055">runtime.ts:1055</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -857,7 +857,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1320">runtime.ts:1320</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1320">runtime.ts:1320</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -900,7 +900,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1197">runtime.ts:1197</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1197">runtime.ts:1197</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -938,7 +938,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1491">runtime.ts:1491</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1491">runtime.ts:1491</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -990,7 +990,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1009">runtime.ts:1009</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1009">runtime.ts:1009</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -1014,7 +1014,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1151">runtime.ts:1151</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1151">runtime.ts:1151</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -1046,7 +1046,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1134">runtime.ts:1134</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1134">runtime.ts:1134</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -1078,7 +1078,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1292">runtime.ts:1292</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1292">runtime.ts:1292</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -1110,7 +1110,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L1223">runtime.ts:1223</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L1223">runtime.ts:1223</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -1141,7 +1141,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L957">runtime.ts:957</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L957">runtime.ts:957</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
diff --git a/docs/reference/api/typedoc/classes/memory.html b/docs/reference/api/typedoc/classes/memory.html
index f4de3b3393..7d20d21976 100644
--- a/docs/reference/api/typedoc/classes/memory.html
+++ b/docs/reference/api/typedoc/classes/memory.html
@@ -130,7 +130,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L40">memory.ts:40</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L40">memory.ts:40</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -152,7 +152,7 @@
<div class="tsd-signature tsd-kind-icon">memory<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Memory</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L32">memory.ts:32</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L32">memory.ts:32</a></li>
</ul>
</aside>
</section>
@@ -162,7 +162,7 @@
<div class="tsd-signature tsd-kind-icon">wasm32<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">boolean</span><span class="tsd-signature-symbol"> = true</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L33">memory.ts:33</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L33">memory.ts:33</a></li>
</ul>
</aside>
</section>
@@ -179,7 +179,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L154">memory.ts:154</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L154">memory.ts:154</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -210,7 +210,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L90">memory.ts:90</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L90">memory.ts:90</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -233,7 +233,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L97">memory.ts:97</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L97">memory.ts:97</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -256,7 +256,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L74">memory.ts:74</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L74">memory.ts:74</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -279,7 +279,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L81">memory.ts:81</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L81">memory.ts:81</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -302,7 +302,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L104">memory.ts:104</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L104">memory.ts:104</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -325,7 +325,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L132">memory.ts:132</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L132">memory.ts:132</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -362,7 +362,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L145">memory.ts:145</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L145">memory.ts:145</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -393,7 +393,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L60">memory.ts:60</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L60">memory.ts:60</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -416,7 +416,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L67">memory.ts:67</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L67">memory.ts:67</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -439,7 +439,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L53">memory.ts:53</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L53">memory.ts:53</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -462,7 +462,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L114">memory.ts:114</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L114">memory.ts:114</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -485,7 +485,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L124">memory.ts:124</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L124">memory.ts:124</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
@@ -502,7 +502,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/memory.ts#L175">memory.ts:175</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/memory.ts#L175">memory.ts:175</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
diff --git a/docs/reference/api/typedoc/classes/module.html b/docs/reference/api/typedoc/classes/module.html
index e492f02691..2c460fb802 100644
--- a/docs/reference/api/typedoc/classes/module.html
+++ b/docs/reference/api/typedoc/classes/module.html
@@ -119,7 +119,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L614">runtime.ts:614</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L614">runtime.ts:614</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -169,7 +169,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L626">runtime.ts:626</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L626">runtime.ts:626</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -186,7 +186,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L653">runtime.ts:653</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L653">runtime.ts:653</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -218,7 +218,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L641">runtime.ts:641</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L641">runtime.ts:641</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -250,7 +250,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L687">runtime.ts:687</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L687">runtime.ts:687</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
diff --git a/docs/reference/api/typedoc/classes/ndarray.html b/docs/reference/api/typedoc/classes/ndarray.html
index 8365765e71..aca58a74fe 100644
--- a/docs/reference/api/typedoc/classes/ndarray.html
+++ b/docs/reference/api/typedoc/classes/ndarray.html
@@ -130,7 +130,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L401">runtime.ts:401</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L401">runtime.ts:401</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -158,7 +158,7 @@
<div class="tsd-signature tsd-kind-icon">device<span class="tsd-signature-symbol">:</span> <a href="dldevice.html" class="tsd-signature-type">DLDevice</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L394">runtime.ts:394</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L394">runtime.ts:394</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -173,7 +173,7 @@
<div class="tsd-signature tsd-kind-icon">dtype<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">string</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L390">runtime.ts:390</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L390">runtime.ts:390</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -188,7 +188,7 @@
<div class="tsd-signature tsd-kind-icon">ndim<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L388">runtime.ts:388</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L388">runtime.ts:388</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -203,7 +203,7 @@
<div class="tsd-signature tsd-kind-icon">shape<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Array</span><span class="tsd-signature-symbol"><</span><span class="tsd-signature-type">number</span><span class="tsd-signature-symbol">></span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L392">runtime.ts:392</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L392">runtime.ts:392</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -225,7 +225,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L480">runtime.ts:480</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L480">runtime.ts:480</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -258,7 +258,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L524">runtime.ts:524</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L524">runtime.ts:524</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -290,7 +290,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L465">runtime.ts:465</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L465">runtime.ts:465</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -307,7 +307,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L458">runtime.ts:458</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L458">runtime.ts:458</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -339,7 +339,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L584">runtime.ts:584</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L584">runtime.ts:584</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -363,7 +363,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L553">runtime.ts:553</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L553">runtime.ts:553</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
diff --git a/docs/reference/api/typedoc/classes/packedfunccell.html b/docs/reference/api/typedoc/classes/packedfunccell.html
index 1e9cd155a5..d3ce2b336d 100644
--- a/docs/reference/api/typedoc/classes/packedfunccell.html
+++ b/docs/reference/api/typedoc/classes/packedfunccell.html
@@ -117,7 +117,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L248">runtime.ts:248</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L248">runtime.ts:248</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -146,7 +146,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L255">runtime.ts:255</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L255">runtime.ts:255</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -163,7 +163,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L264">runtime.ts:264</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L264">runtime.ts:264</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
diff --git a/docs/reference/api/typedoc/classes/rpcserver.html b/docs/reference/api/typedoc/classes/rpcserver.html
index 0996ff1814..4704d75ce0 100644
--- a/docs/reference/api/typedoc/classes/rpcserver.html
+++ b/docs/reference/api/typedoc/classes/rpcserver.html
@@ -115,7 +115,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L95">rpc_server.ts:95</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L95">rpc_server.ts:95</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -176,7 +176,7 @@
<div class="tsd-signature tsd-kind-icon">get<wbr>Imports<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-symbol">(</span><span class="tsd-signature-symbol">)</span><span class="tsd-signature-symbol"> => </span><span class="tsd-signature-type">Record</span><span class="tsd-signature-symbol"><</span><span class="tsd-signature-type">string</span><span class="tsd-signature-symbol">, </span><span class="tsd-signature-type">unknown</span><span class="tsd-signat [...]
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L84">rpc_server.ts:84</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L84">rpc_server.ts:84</a></li>
</ul>
</aside>
<div class="tsd-type-declaration">
@@ -201,7 +201,7 @@
<div class="tsd-signature tsd-kind-icon">key<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">string</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L80">rpc_server.ts:80</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L80">rpc_server.ts:80</a></li>
</ul>
</aside>
</section>
@@ -211,7 +211,7 @@
<div class="tsd-signature tsd-kind-icon">logger<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-symbol">(</span>msg<span class="tsd-signature-symbol">: </span><span class="tsd-signature-type">string</span><span class="tsd-signature-symbol">)</span><span class="tsd-signature-symbol"> => </span><span class="tsd-signature-type">void</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L83">rpc_server.ts:83</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L83">rpc_server.ts:83</a></li>
</ul>
</aside>
<div class="tsd-type-declaration">
@@ -242,7 +242,7 @@
<div class="tsd-signature tsd-kind-icon">socket<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">WebSocket</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L81">rpc_server.ts:81</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L81">rpc_server.ts:81</a></li>
</ul>
</aside>
</section>
@@ -252,7 +252,7 @@
<div class="tsd-signature tsd-kind-icon">state<span class="tsd-signature-symbol">:</span> <a href="../enums/rpcserverstate.html" class="tsd-signature-type">RPCServerState</a><span class="tsd-signature-symbol"> = RPCServerState.InitHeader</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L82">rpc_server.ts:82</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L82">rpc_server.ts:82</a></li>
</ul>
</aside>
</section>
@@ -262,7 +262,7 @@
<div class="tsd-signature tsd-kind-icon">url<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">string</span></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/rpc_server.ts#L79">rpc_server.ts:79</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/rpc_server.ts#L79">rpc_server.ts:79</a></li>
</ul>
</aside>
</section>
diff --git a/docs/reference/api/typedoc/classes/runtimecontext.html b/docs/reference/api/typedoc/classes/runtimecontext.html
index 4fbeeec5f9..337a4c1c38 100644
--- a/docs/reference/api/typedoc/classes/runtimecontext.html
+++ b/docs/reference/api/typedoc/classes/runtimecontext.html
@@ -132,7 +132,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L148">runtime.ts:148</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L148">runtime.ts:148</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -172,7 +172,7 @@
<div class="tsd-signature tsd-kind-icon">array<wbr>Get<wbr>Item<span class="tsd-signature-symbol">:</span> <a href="../index.html#packedfunc" class="tsd-signature-type">PackedFunc</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L143">runtime.ts:143</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L143">runtime.ts:143</a></li>
</ul>
</aside>
</section>
@@ -182,7 +182,7 @@
<div class="tsd-signature tsd-kind-icon">array<wbr>Get<wbr>Size<span class="tsd-signature-symbol">:</span> <a href="../index.html#packedfunc" class="tsd-signature-type">PackedFunc</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L144">runtime.ts:144</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L144">runtime.ts:144</a></li>
</ul>
</aside>
</section>
@@ -192,7 +192,7 @@
<div class="tsd-signature tsd-kind-icon">array<wbr>Make<span class="tsd-signature-symbol">:</span> <a href="../index.html#packedfunc" class="tsd-signature-type">PackedFunc</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L145">runtime.ts:145</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L145">runtime.ts:145</a></li>
</ul>
</aside>
</section>
@@ -202,7 +202,7 @@
<div class="tsd-signature tsd-kind-icon">get<wbr>Sys<wbr>Lib<span class="tsd-signature-symbol">:</span> <a href="../index.html#packedfunc" class="tsd-signature-type">PackedFunc</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L146">runtime.ts:146</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L146">runtime.ts:146</a></li>
</ul>
</aside>
</section>
@@ -219,7 +219,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L189">runtime.ts:189</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L189">runtime.ts:189</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -263,7 +263,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L163">runtime.ts:163</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L163">runtime.ts:163</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -280,7 +280,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L208">runtime.ts:208</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L208">runtime.ts:208</a></li>
</ul>
</aside>
<h4 class="tsd-type-parameters-title">Type parameters</h4>
@@ -309,7 +309,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L157">runtime.ts:157</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L157">runtime.ts:157</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -326,7 +326,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/32e500b7f/web/src/runtime.ts#L167">runtime.ts:167</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/d1dea1376/web/src/runtime.ts#L167">runtime.ts:167</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
... 1961 lines suppressed ...