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 2023/01/27 10:12:48 UTC
[tvm-site] branch asf-site updated: deploying docs (apache/tvm@56771a87d1560f8963cf745ad093ffc3d83f3f6a)
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 de3f1b145b deploying docs (apache/tvm@56771a87d1560f8963cf745ad093ffc3d83f3f6a)
de3f1b145b is described below
commit de3f1b145ba338a5034abc3e0bb88a8680c634e2
Author: tvm-bot <95...@users.noreply.github.com>
AuthorDate: Fri Jan 27 10:12:39 2023 +0000
deploying docs (apache/tvm@56771a87d1560f8963cf745ad093ffc3d83f3f6a)
---
docs/_images/sphx_glr_micro_train_001.png | Bin 322817 -> 331892 bytes
docs/_images/sphx_glr_micro_train_thumb.png | Bin 23407 -> 23825 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 | 4 +-
.../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 | 20 +-
.../extend_tvm/bring_your_own_datatypes.rst.txt | 2 +-
.../how_to/extend_tvm/sg_execution_times.rst.txt | 8 +-
.../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 | 2570 ++++++++++----------
.../tune_network_cuda.rst.txt | 4 +-
.../tune_network_x86.rst.txt | 4 +-
.../tune_sparse_x86.rst.txt | 76 +-
.../tune_with_autotvm/sg_execution_times.rst.txt | 8 +-
.../tune_with_autotvm/tune_conv2d_cuda.rst.txt | 100 +-
.../work_with_microtvm/micro_autotune.rst.txt | 16 +-
.../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 | 12 +-
.../work_with_relay/sg_execution_times.rst.txt | 8 +-
.../how_to/work_with_schedules/intrin_math.rst.txt | 2 +-
.../work_with_schedules/sg_execution_times.rst.txt | 14 +-
.../how_to/work_with_schedules/tensorize.rst.txt | 2 +-
.../tutorials/autotvm/sg_execution_times.rst.txt | 4 +-
.../frontend/deploy_classification.rst.txt | 2 +-
.../tutorials/frontend/deploy_detection.rst.txt | 2 +-
.../tutorials/frontend/sg_execution_times.rst.txt | 6 +-
.../tutorials/optimize/sg_execution_times.rst.txt | 6 +-
.../topic/vta/tutorials/sg_execution_times.rst.txt | 6 +-
.../tutorial/auto_scheduler_matmul_x86.rst.txt | 4 +-
docs/_sources/tutorial/autotvm_matmul_x86.rst.txt | 20 +-
docs/_sources/tutorial/autotvm_relay_x86.rst.txt | 58 +-
.../tutorial/cross_compilation_and_rpc.rst.txt | 2 +-
docs/_sources/tutorial/intro_topi.rst.txt | 2 +-
docs/_sources/tutorial/sg_execution_times.rst.txt | 22 +-
.../tutorial/tensor_expr_get_started.rst.txt | 44 +-
docs/commit_hash | 2 +-
docs/how_to/compile_models/from_darknet.html | 2 +-
docs/how_to/compile_models/from_keras.html | 2 +-
docs/how_to/compile_models/from_mxnet.html | 2 +-
docs/how_to/compile_models/from_oneflow.html | 11 +-
docs/how_to/compile_models/from_pytorch.html | 11 +-
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 | 4 +-
.../deploy_models/deploy_model_on_android.html | 2 +-
.../deploy_object_detection_pytorch.html | 39 +-
docs/how_to/deploy_models/deploy_prequantized.html | 8 +-
.../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 | 43 +-
docs/how_to/deploy_models/sg_execution_times.html | 20 +-
.../extend_tvm/bring_your_own_datatypes.html | 2 +-
docs/how_to/extend_tvm/sg_execution_times.html | 8 +-
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 | 2570 ++++++++++----------
.../tune_with_autoscheduler/tune_network_cuda.html | 4 +-
.../tune_with_autoscheduler/tune_network_x86.html | 4 +-
.../tune_with_autoscheduler/tune_sparse_x86.html | 76 +-
.../tune_with_autotvm/sg_execution_times.html | 8 +-
.../how_to/tune_with_autotvm/tune_conv2d_cuda.html | 100 +-
docs/how_to/work_with_microtvm/micro_autotune.html | 16 +-
docs/how_to/work_with_microtvm/micro_pytorch.html | 4 +-
docs/how_to/work_with_microtvm/micro_train.html | 16 +-
.../work_with_microtvm/sg_execution_times.html | 12 +-
.../how_to/work_with_relay/sg_execution_times.html | 8 +-
docs/how_to/work_with_schedules/intrin_math.html | 2 +-
.../work_with_schedules/sg_execution_times.html | 14 +-
docs/how_to/work_with_schedules/tensorize.html | 2 +-
docs/reference/api/python/auto_scheduler.html | 4 +-
.../api/typedoc/classes/bytestreamreader.html | 12 +-
.../api/typedoc/classes/cachedcallstack.html | 34 +-
docs/reference/api/typedoc/classes/dldatatype.html | 12 +-
docs/reference/api/typedoc/classes/dldevice.html | 10 +-
.../reference/api/typedoc/classes/environment.html | 12 +-
docs/reference/api/typedoc/classes/ffilibrary.html | 20 +-
.../api/typedoc/classes/graphexecutor.html | 16 +-
docs/reference/api/typedoc/classes/instance.html | 40 +-
docs/reference/api/typedoc/classes/memory.html | 34 +-
docs/reference/api/typedoc/classes/module.html | 10 +-
docs/reference/api/typedoc/classes/ndarray.html | 22 +-
.../api/typedoc/classes/packedfunccell.html | 6 +-
docs/reference/api/typedoc/classes/rpcserver.html | 14 +-
docs/reference/api/typedoc/classes/scalar.html | 6 +-
.../api/typedoc/classes/webgpucontext.html | 12 +-
docs/reference/api/typedoc/enums/argtypecode.html | 30 +-
.../api/typedoc/enums/aynccallbackcode.html | 4 +-
.../api/typedoc/enums/dldatatypecode.html | 8 +-
.../api/typedoc/enums/rpcserverstate.html | 12 +-
docs/reference/api/typedoc/enums/sizeof.html | 18 +-
docs/reference/api/typedoc/index.html | 112 +-
.../api/typedoc/interfaces/disposable.html | 2 +-
.../api/typedoc/interfaces/functioninfo.html | 6 +-
.../api/typedoc/interfaces/libraryprovider.html | 4 +-
docs/searchindex.js | 2 +-
.../vta/tutorials/autotvm/sg_execution_times.html | 4 +-
.../tutorials/frontend/deploy_classification.html | 2 +-
.../vta/tutorials/frontend/deploy_detection.html | 2 +-
.../vta/tutorials/frontend/sg_execution_times.html | 10 +-
.../vta/tutorials/optimize/sg_execution_times.html | 6 +-
docs/topic/vta/tutorials/sg_execution_times.html | 6 +-
docs/tutorial/auto_scheduler_matmul_x86.html | 4 +-
docs/tutorial/autotvm_matmul_x86.html | 20 +-
docs/tutorial/autotvm_relay_x86.html | 276 +--
docs/tutorial/cross_compilation_and_rpc.html | 2 +-
docs/tutorial/intro_topi.html | 2 +-
docs/tutorial/sg_execution_times.html | 22 +-
docs/tutorial/tensor_expr_get_started.html | 44 +-
129 files changed, 3584 insertions(+), 3522 deletions(-)
diff --git a/docs/_images/sphx_glr_micro_train_001.png b/docs/_images/sphx_glr_micro_train_001.png
index 9d7b73ba75..9ba43750df 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 9ae7d8cdc5..d025dae545 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 1825363025..97ac56db9b 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 14.131 seconds)
+ **Total running time of the script:** ( 1 minutes 17.355 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 ef1b2f5575..e121acc09d 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 947ms/step
+
1/1 [==============================] - ETA: 0s
1/1 [==============================] - 1s 954ms/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 db1ae4d592..0038c25ffb 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.zip28c6522a-69df-4ac8-ab68-ff22f5a4cbfc from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip...
+ Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zipa0ce24eb-8bfb-45e2-aef5-6e7943372f9a 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 80fac46df8..34b29787a9 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]
19%|#9 | 7.99M/41.5M [00:00<00:00, 77.5MB/s]
39%|###8 | 16.0M/41.5M [00:00<00:00, 75.0MB/s]
58%|#####7 | 24.0M/41.5M [00:00<00:00, 66.7MB/s]
90%|########9 | 37.2M/41.5M [00:00<00:00, 85.3MB/s]
100%|##########| 41.5M/41.5M [00:00<00:00, 78.9MB/s]
+
0%| | 0.00/41.5M [00:00<?, ?B/s]
19%|#9 | 7.99M/41.5M [00:00<00:00, 35.3MB/s]
39%|###8 | 16.0M/41.5M [00:00<00:00, 40.9MB/s]
58%|#####7 | 24.0M/41.5M [00:00<00:00, 45.4MB/s]
77%|#######7 | 32.0M/41.5M [00:00<00:00, 52.5MB/s]
92%|#########2| 38.3M/41.5M [00:00<00:00, 49.9MB/s]
100%|##########| 41.5M/41.5M [00:00<00:00, 48.6MB/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 8f814998a1..9f4b0a70bb 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]
27%|##7 | 12.2M/44.7M [00:00<00:00, 128MB/s]
55%|#####4 | 24.4M/44.7M [00:00<00:00, 101MB/s]
80%|#######9 | 35.7M/44.7M [00:00<00:00, 107MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 106MB/s]
+
0%| | 0.00/44.7M [00:00<?, ?B/s]
14%|#4 | 6.30M/44.7M [00:00<00:00, 60.9MB/s]
32%|###2 | 14.3M/44.7M [00:00<00:00, 48.1MB/s]
49%|####8 | 21.7M/44.7M [00:00<00:00, 58.3MB/s]
62%|######1 | 27.7M/44.7M [00:00<00:00, 58.6MB/s]
81%|########1 | 36.4M/44.7M [00:00<00:00, 68.9MB/s]
97%|#########6| 43.3M/44.7M [00:00<00:00, 68.7MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 56.8MB/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 2699c8edae..e62113a294 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 21.545 seconds)
+ **Total running time of the script:** ( 1 minutes 22.479 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 b6548d8399..4674d6f8c4 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
=================
-**06:19.061** total execution time for **how_to_compile_models** files:
+**06:25.962** total execution time for **how_to_compile_models** files:
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:21.545 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:22.479 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``) | 01:14.131 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``) | 01:17.355 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``) | 00:52.717 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``) | 00:53.409 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``) | 00:35.828 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``) | 00:35.917 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``) | 00:30.544 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``) | 00:30.417 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``) | 00:30.531 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``) | 00:30.381 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``) | 00:26.388 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``) | 00:27.487 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``) | 00:24.511 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``) | 00:25.115 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``) | 00:20.244 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``) | 00:20.758 | 0.0 MB |
+-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``) | 00:02.622 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``) | 00:02.644 | 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 f820c0874b..a3381a7532 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
@@ -727,7 +727,7 @@ 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)
- 3341.0008 3340.7200 3343.2723 3338.7640 1.4712
+ 3342.1493 3340.8745 3354.0181 3338.0046 4.3589
@@ -736,7 +736,7 @@ well as provides information about the model's performance
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 3.163 seconds)
+ **Total running time of the script:** ( 1 minutes 3.134 seconds)
.. _sphx_glr_download_how_to_deploy_models_deploy_model_on_adreno.py:
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 1674480280..86530d76d5 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)
- 16.6069 16.7762 17.1110 16.0513 0.3718
+ 15.8913 15.6132 16.6356 15.4781 0.4431
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 6470f579c2..2bb5b28df0 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<00:02, 83.3MB/s]
13%|#2 | 21.3M/170M [00:00<00:01, 116MB/s]
19%|#9 | 32.3M/170M [00:00<00:01, 106MB/s]
25%|##5 | 42.6M/170M [00:00<00:01, 105MB/s]
31%|### | 52.6M/170M [00:00<00:01, 103MB/s]
37%|###6 | 62.5M/170M [00:00<00:01, 102MB/s]
43%|####2 | 72.3M/170M [00:00<00:01, 101MB/s]
48%|####8 | 81.9M/170M [00:00<00:00, 96.7MB/s]
55%|#####4 | 92.9M/170M [00:00<00:00, 102MB/s]
60%|###### | 103M/170M [00:01<00:00, 99.3MB/s]
66%|######6 | 112M/170M [00:01<00:00, 100MB/s]
72%|#######2 | 123M/170M [00:01<00:00, 102MB/s]
78%|#######7 | 132M/170M [00:01<00:00, 96.5MB/s]
84%|########4 | 143M/170M [00:01<00:00, 102MB/s]
90%|######### | 153M/170M [00:01<00:00, 88.6MB/s]
99%|#########8| 167M/170M [00:01<00:00, 105MB/s]
100%|##########| 170M/170M [00:01<00:00, 101MB/s]
+
0%| | 0.00/170M [00:00<?, ?B/s]
5%|4 | 7.99M/170M [00:00<00:04, 35.0MB/s]
9%|9 | 16.0M/170M [00:00<00:03, 44.6MB/s]
14%|#4 | 24.0M/170M [00:00<00:03, 49.0MB/s]
18%|#7 | 30.3M/170M [00:00<00:03, 48.2MB/s]
21%|## | 35.0M/170M [00:00<00:03, 35.7MB/s]
25%|##4 | 42.1M/170M [00:01<00:03, 40.1MB/s]
30%|##9 | 50.2M/170M [00:01<00:02, 50.1MB/s]
35%|###5 | 59.7M/170M [00:01<00:01, 61.7MB/s]
39%|###9 | 66.4M/170M [00:01<00:01, 63.9MB/s]
47%|####7 | 80.0M/170M [00:01<00:01, 73.0MB/s]
52%|#####1 | 88.0M/170M [00:01<00:01, 75.6MB/s]
57%|#####6 | 96.0M/170M [00:01<00:01, 76.4MB/s]
64%|######4 | 109M/170M [00:01<00:00, 93.6MB/s]
70%|######9 | 119M/170M [00:01<00:00, 90.7MB/s]
75%|#######5 | 128M/170M [00:02<00:00, 89.4MB/s]
80%|######## | 136M/170M [00:02<00:00, 73.8MB/s]
85%|########4 | 144M/170M [00:02<00:00, 71.0MB/s]
92%|#########1| 156M/170M [00:02<00:00, 83.6MB/s]
98%|#########7| 166M/170M [00:02<00:00, 83.4MB/s]
100%|##########| 170M/170M [00:02<00:00, 66.9MB/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 27.786 seconds)
+ **Total running time of the script:** ( 3 minutes 27.731 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 c9e5573f32..1f4e7163e5 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]
100%|#########9| 13.5M/13.6M [00:00<00:00, 142MB/s]
100%|##########| 13.6M/13.6M [00:00<00:00, 141MB/s]
+
0%| | 0.00/13.6M [00:00<?, ?B/s]
59%|#####8 | 7.99M/13.6M [00:00<00:00, 36.8MB/s]
100%|##########| 13.6M/13.6M [00:00<00:00, 46.7MB/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)
- 90.2892 90.2321 90.9815 90.0854 0.1888
+ 90.6740 90.6287 95.8723 90.2278 0.5534
@@ -458,7 +458,7 @@ TODO
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 14.094 seconds)
+ **Total running time of the script:** ( 1 minutes 13.967 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 fe538b692a..2739349e85 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)
- 120.8368 120.7806 123.1834 120.0746 0.4294
+ 119.6954 119.7157 120.3397 118.9367 0.2781
@@ -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 34.550 seconds)
+ **Total running time of the script:** ( 2 minutes 35.610 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 fdc04a776d..cb04fcd29d 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 20.755 seconds)
+ **Total running time of the script:** ( 1 minutes 42.340 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 28197d4ecd..3536e3fc16 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 | 6147/132723 [00:00<00:02, 61464.05KB/s]
11%|# | 14511/132723 [00:00<00:01, 74505.85KB/s]
17%|#6 | 21962/132723 [00:00<00:01, 59073.28KB/s]
21%|##1 | 28175/132723 [00:00<00:01, 59562.02KB/s]
27%|##6 | 35273/132723 [00:00<00:01, 63275.70KB/s]
33%|###3 | 44041/132723 [00:00<00:01, 71017.19KB/s]
40%|###9 | 52694/132723 [00:00<00:01, 75854.81KB/s]
46%|####6 | 61385/132723 [00:00<00:00, 79259.51KB/s]
53%|#####2 | 70096/132723 [00:00<00:00, 81651.16KB/s]
59%|#####9 | 78933/132723 [00:01<00:00, 83691.51KB/s]
66%|######5 | 87584/132723 [00:01<00:00, 84539.97KB/s]
73%|#######2 | 96432/132723 [00:01<00:00, 85722.14KB/s]
79%|#######9 | 105229/132723 [00:01<00:00, 86395.14KB/s]
86%|########5 | 114055/132723 [00:01<00:00, 86954.14KB/s]
93%|#########2| 122821/132723 [00:01<00:00, 87162.81KB/s]
99%|########
#9| 131635/132723 [00:01<00:00, 87454.03KB/s]
100%|##########| 132723/132723 [00:01<00:00, 79248.71KB/s]
+
0%| | 0/132723 [00:00<?, ?KB/s]
3%|2 | 3631/132723 [00:00<00:03, 34719.86KB/s]
6%|5 | 7422/132723 [00:00<00:03, 36451.66KB/s]
9%|8 | 11422/132723 [00:00<00:03, 38043.10KB/s]
12%|#1 | 15700/132723 [00:00<00:02, 39898.17KB/s]
15%|#4 | 19728/132723 [00:00<00:02, 40034.30KB/s]
18%|#8 | 24286/132723 [00:00<00:02, 41898.24KB/s]
22%|##1 | 28587/132723 [00:00<00:02, 42259.41KB/s]
25%|##5 | 33566/132723 [00:00<00:02, 44400.12KB/s]
29%|##9 | 38926/132723 [00:00<00:01, 47037.46KB/s]
33%|###2 | 43628/132723 [00:01<00:01, 44831.40KB/s]
37%|###7 | 49178/132723 [00:01<00:01, 47972.57KB/s]
41%|####1 | 54770/132723 [00:01<00:01, 50324.63KB/s]
45%|####5 | 59918/132723 [00:01<00:01, 50577.75KB/s]
50%|####9 | 65806/132723 [00:01<00:01, 52883.19KB/s]
54%|#####4 | 71886/132723 [00:01<00:01, 55129.48KB/s]
59%|#####8 |
78148/132723 [00:01<00:00, 57359.85KB/s]
64%|######3 | 84764/132723 [00:01<00:00, 59986.53KB/s]
69%|######8 | 91414/132723 [00:01<00:00, 61932.02KB/s]
74%|#######3 | 97947/132723 [00:01<00:00, 62946.66KB/s]
79%|#######9 | 105128/132723 [00:02<00:00, 65599.10KB/s]
85%|########4 | 112169/132723 [00:02<00:00, 67039.19KB/s]
90%|######### | 119673/132723 [00:02<00:00, 69434.24KB/s]
96%|#########5| 127136/132723 [00:02<00:00, 70989.63KB/s]
100%|##########| 132723/132723 [00:02<00:00, 55008.50KB/s]
@@ -246,7 +246,7 @@ Display result
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 3 minutes 34.842 seconds)
+ **Total running time of the script:** ( 3 minutes 33.963 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 107623b996..ff02177340 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
@@ -5,26 +5,26 @@
Computation times
=================
-**14:51.380** total execution time for **how_to_deploy_models** files:
+**15:12.069** total execution time for **how_to_deploy_models** files:
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``) | 03:34.842 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``) | 03:33.963 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 03:27.786 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 03:27.731 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``) | 02:34.550 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``) | 02:35.610 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``) | 01:20.755 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``) | 01:42.340 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``) | 01:14.094 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``) | 01:13.967 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno.py` (``deploy_model_on_adreno.py``) | 01:03.163 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno.py` (``deploy_model_on_adreno.py``) | 01:03.134 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``) | 00:40.880 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``) | 00:40.390 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``) | 00:27.715 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``) | 00:27.686 | 0.0 MB |
+------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``) | 00:27.588 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``) | 00:27.241 | 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 39a4c7aa80..9a781f40be 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.zip2e21da86-954a-4776-ad44-0a8bcc9cb300 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
+ Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zip338d35f6-a39e-477d-bc04-2e9c1ba5f341 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 57719ef473..95759526bd 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
=================
-**00:53.112** total execution time for **how_to_extend_tvm** files:
+**00:53.023** 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``) | 00:49.323 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_bring_your_own_datatypes.py` (``bring_your_own_datatypes.py``) | 00:49.270 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``) | 00:02.708 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``) | 00:02.684 | 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``) | 00:01.062 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_extend_tvm_low_level_custom_pass.py` (``low_level_custom_pass.py``) | 00:00.007 | 0.0 MB |
+-------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/extend_tvm/use_pass_instrument.rst.txt b/docs/_sources/how_to/extend_tvm/use_pass_instrument.rst.txt
index 81a9aa1a59..df896aa98e 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: 22115us [22115us] (49.77%; 49.77%)
- FoldScaleAxis: 22321us [8us] (50.23%; 50.23%)
- FoldConstant: 22313us [1754us] (50.21%; 99.97%)
- InferType: 20559us [20559us] (46.27%; 92.14%)
+ InferType: 21208us [21208us] (48.16%; 48.16%)
+ FoldScaleAxis: 22830us [7us] (51.84%; 51.84%)
+ FoldConstant: 22823us [1672us] (51.83%; 99.97%)
+ InferType: 21151us [21151us] (48.03%; 92.67%)
@@ -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: 20728us [20728us] (48.25%; 48.25%)
- FoldScaleAxis: 22232us [6us] (51.75%; 51.75%)
- FoldConstant: 22227us [1738us] (51.74%; 99.98%)
- InferType: 20489us [20489us] (47.69%; 92.18%)
+ InferType: 20662us [20662us] (48.06%; 48.06%)
+ FoldScaleAxis: 22327us [5us] (51.94%; 51.94%)
+ FoldConstant: 22322us [1720us] (51.92%; 99.98%)
+ InferType: 20602us [20602us] (47.92%; 92.29%)
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 c232f2348b..5f8c33144f 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: 36.795871 ms
+ Convolution: 46.436958 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 c81df42083..72f04ac1c8 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
@@ -608,7 +608,7 @@ be able to run on our build server
.. code-block:: none
- conv2d with tensor core: 7.718617 ms
+ conv2d with tensor core: 6.909338 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 74ee4986ee..d84952021d 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: 0.018811
- Baseline: 3.396133
+ Numpy running time: 0.018045
+ Baseline: 3.425894
@@ -227,7 +227,7 @@ fill 32 * 32 * sizeof(float) which is 4KB in the cache whose total size is 32KB
.. code-block:: none
- Opt1: 0.302338
+ Opt1: 0.293250
@@ -318,7 +318,7 @@ In this tutorial, we chose to vectorize the inner loop row data since it is cach
.. code-block:: none
- Opt2: 0.332977
+ Opt2: 0.331320
@@ -406,7 +406,7 @@ the access pattern for A matrix is more cache friendly.
.. code-block:: none
- Opt3: 0.116724
+ Opt3: 0.116212
@@ -523,7 +523,7 @@ flattening.
.. code-block:: none
- Opt4: 0.109741
+ Opt4: 0.109608
@@ -635,7 +635,7 @@ write to C when all the block results are ready.
.. code-block:: none
- Opt5: 0.110908
+ Opt5: 0.111050
@@ -748,7 +748,7 @@ Furthermore, we can also utilize multi-core processors to do the thread-level pa
.. code-block:: none
- Opt6: 0.146963
+ Opt6: 0.146901
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 5e4eac07cc..63801843f6 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
=================
-**00:34.845** total execution time for **how_to_optimize_operators** files:
+**00:34.690** total execution time for **how_to_optimize_operators** files:
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``) | 00:32.301 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``) | 00:32.165 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.475 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.449 | 0.0 MB |
+-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``) | 00:01.068 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``) | 00:01.075 | 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 4e3b22c967..14420f9660 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
=================
-**09:23.537** total execution time for **how_to_tune_with_autoscheduler** files:
+**09:16.438** 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``) | 05:42.565 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_conv2d_layer_cuda.py` (``tune_conv2d_layer_cuda.py``) | 05:35.291 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``) | 01:39.693 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``) | 01:39.443 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``) | 01:05.839 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``) | 01:06.161 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``) | 00:28.284 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``) | 00:28.467 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``) | 00:14.117 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``) | 00:14.108 | 0.0 MB |
+----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``) | 00:13.038 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``) | 00:12.969 | 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 e0972c5114..26fab967cd 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
@@ -244,13 +244,13 @@ cooperative fetching, unrolling and operator fusion.
def main(data: T.Buffer((1, 512, 7, 7), "float32"), kernel: T.Buffer((512, 512, 3, 3), "float32"), bias: T.Buffer((1, 512, 1, 1), "float32"), compute: T.Buffer((1, 512, 7, 7), "float32")):
T.func_attr({"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True})
blockIdx_x = T.env_thread("blockIdx.x")
- T.launch_thread(blockIdx_x, 112)
+ T.launch_thread(blockIdx_x, 28)
conv2d_nchw = T.allocate([7], "float32", "local")
- pad_temp_shared = T.allocate([432], "float32", "shared")
- kernel_shared = T.allocate([4608], "float32", "shared")
+ pad_temp_shared = T.allocate([216], "float32", "shared")
+ kernel_shared = T.allocate([9216], "float32", "shared")
threadIdx_x = T.env_thread("threadIdx.x")
- T.launch_thread(threadIdx_x, 32)
- conv2d_nchw_1 = T.Buffer((7,), data=conv2d_nchw, scope="local", align=16)
+ T.launch_thread(threadIdx_x, 128)
+ conv2d_nchw_1 = T.Buffer((1,), data=conv2d_nchw, scope="local", align=4)
conv2d_nchw_1[0] = T.float32(0)
conv2d_nchw_1[1] = T.float32(0)
conv2d_nchw_1[2] = T.float32(0)
@@ -258,786 +258,676 @@ cooperative fetching, unrolling and operator fusion.
conv2d_nchw_1[4] = T.float32(0)
conv2d_nchw_1[5] = T.float32(0)
conv2d_nchw_1[6] = T.float32(0)
- for rc_outer_outer in range(32):
- cse_var_2: T.int32 = rc_outer_outer * 784
- cse_var_1: T.int32 = rc_outer_outer * 144
+ for rc_outer_outer in range(64):
+ cse_var_1: T.int32 = rc_outer_outer * 72
threadIdx_x_1 = T.env_thread("threadIdx.x")
- pad_temp_shared_1 = T.Buffer((432,), data=pad_temp_shared, scope="shared")
+ pad_temp_shared_1 = T.Buffer((216,), data=pad_temp_shared, scope="shared")
data_1 = T.Buffer((25088,), data=data.data)
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(1 <= threadIdx_x_1 % 27 // 9 + blockIdx_x % 7 and threadIdx_x_1 % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_2 + threadIdx_x_1 // 27 * 49 + threadIdx_x_1 % 27 // 9 * 7 + blockIdx_x % 7 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 32] = T.if_then_else(1 <= (threadIdx_x_1 + 5) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 5) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 32) // 27 * 49 + (threadIdx_x_1 + 5) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 64] = T.if_then_else(1 <= (threadIdx_x_1 + 10) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 10) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 64) // 27 * 49 + (threadIdx_x_1 + 10) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 96] = T.if_then_else(1 <= (threadIdx_x_1 + 15) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 15) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 96) // 27 * 49 + (threadIdx_x_1 + 15) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 128] = T.if_then_else(1 <= (threadIdx_x_1 + 20) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 20) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 128) // 27 * 49 + (threadIdx_x_1 + 20) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 160] = T.if_then_else(1 <= (threadIdx_x_1 + 25) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 25) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 160) // 27 * 49 + (threadIdx_x_1 + 25) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 192] = T.if_then_else(1 <= (threadIdx_x_1 + 3) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 3) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 192) // 27 * 49 + (threadIdx_x_1 + 3) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 224] = T.if_then_else(1 <= (threadIdx_x_1 + 8) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 8) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 224) // 27 * 49 + (threadIdx_x_1 + 8) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 256] = T.if_then_else(1 <= (threadIdx_x_1 + 13) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 13) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 256) // 27 * 49 + (threadIdx_x_1 + 13) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 288] = T.if_then_else(1 <= blockIdx_x % 7 + (threadIdx_x_1 // 9 + 2) % 3 and blockIdx_x % 7 + (threadIdx_x_1 // 9 + 2) % 3 < 8 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 288) // 27 * 49 + blockIdx_x % 7 * 7 + (threadIdx_x_1 // 9 + 2) % 3 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 320] = T.if_then_else(1 <= (threadIdx_x_1 + 23) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 23) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 320) // 27 * 49 + (threadIdx_x_1 + 23) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 352] = T.if_then_else(1 <= (threadIdx_x_1 + 1) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 1) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 352) // 27 * 49 + (threadIdx_x_1 + 1) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 384] = T.if_then_else(1 <= (threadIdx_x_1 + 6) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 6) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 384) // 27 * 49 + (threadIdx_x_1 + 6) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- if T.likely(threadIdx_x_1 < 16):
- pad_temp_shared_1[threadIdx_x_1 + 416] = T.if_then_else((threadIdx_x_1 + 11) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 416) // 27 * 49 + (threadIdx_x_1 + 11) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ with T.launch_thread(threadIdx_x_1, 128):
+ pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(1 <= threadIdx_x_1 % 27 // 9 + blockIdx_x % 7 and threadIdx_x_1 % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[rc_outer_outer * 392 + threadIdx_x_1 // 27 * 49 + threadIdx_x_1 % 27 // 9 * 7 + blockIdx_x % 7 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
+ with T.launch_thread(threadIdx_x_1, 128):
+ if T.likely(threadIdx_x_1 < 88):
+ pad_temp_shared_1[threadIdx_x_1 + 128] = T.if_then_else(1 <= (threadIdx_x_1 + 20) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 20) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[rc_outer_outer * 392 + (threadIdx_x_1 + 128) // 27 * 49 + (threadIdx_x_1 + 20) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
threadIdx_x_2 = T.env_thread("threadIdx.x")
- kernel_shared_1 = T.Buffer((4608,), data=kernel_shared, scope="shared")
+ kernel_shared_1 = T.Buffer((9216,), data=kernel_shared, scope="shared")
kernel_1 = T.Buffer((2359296,), data=kernel.data)
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 32] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + (threadIdx_x_2 + 32) // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 64] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + (threadIdx_x_2 + 64) // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 96] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 96]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 128] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 128) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 160] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 160) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 192] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 192) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 224] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 224) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 256] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 256) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 288] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 9216]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 320] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 320) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 352] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 352) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 384] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 384) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 416] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 416) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 448] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 448) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 480] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 480) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 512] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 512) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 544] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 544) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 576] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 18432]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 608] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 608) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 640] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 640) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 672] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 672) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 704] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 704) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 736] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 736) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 768] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 768) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 800] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 800) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 832] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 832) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 864] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 27648]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 896] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 896) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 928] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 928) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 960] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 960) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 992] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 992) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1024] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1024) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1056] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1056) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1088] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1088) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1120] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1120) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1152] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 36864]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1184] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1184) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1216] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1216) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1248] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1248) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1280] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1280) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1312] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1312) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1344] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1344) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1376] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1376) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1408] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1408) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1440] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 46080]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1472] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1472) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1504] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1504) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1536] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1536) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1568] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1568) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1600] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1600) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1632] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1632) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1664] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1664) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1696] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1696) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1728] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 55296]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1760] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1760) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1792] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1792) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1824] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1824) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1856] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1856) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1888] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1888) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1920] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1920) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1952] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1952) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1984] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1984) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2016] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 64512]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2048] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2048) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2080] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2080) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2112] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2112) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2144] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2144) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2176] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2176) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2208] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2208) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2240] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2240) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2272] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2272) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2304] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 73728]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2336] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2336) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2368] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2368) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2400] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2400) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2432] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2432) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2464] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2464) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2496] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2496) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2528] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2528) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2560] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2560) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2592] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 82944]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2624] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2624) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2656] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2656) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2688] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2688) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2720] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2720) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2752] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2752) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2784] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2784) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2816] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2816) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2848] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2848) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2880] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 92160]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2912] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2912) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2944] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2944) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2976] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2976) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3008] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3008) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3040] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3040) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3072] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3072) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3104] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3104) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3136] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3136) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3168] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 101376]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3200] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3200) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3232] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3232) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3264] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3264) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3296] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3296) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3328] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3328) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3360] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3360) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3392] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3392) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3424] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3424) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3456] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 110592]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3488] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3488) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3520] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3520) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3552] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3552) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3584] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3584) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3616] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3616) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3648] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3648) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3680] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3680) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3712] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3712) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3744] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 119808]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3776] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3776) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3808] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3808) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3840] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3840) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3872] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3872) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3904] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3904) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3936] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3936) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3968] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3968) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4000] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4000) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4032] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 129024]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4064] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4064) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4096] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4096) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4128] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4128) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4160] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4160) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4192] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4192) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4224] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4224) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4256] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4256) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4288] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4288) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4320] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 138240]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4352] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4352) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4384] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4384) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4416] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4416) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4448] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4448) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4480] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4480) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4512] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4512) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4544] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4544) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4576] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4576) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- for ry_outer_inner in range(3):
- cse_var_115: T.int32 = ry_outer_inner * 9
- cse_var_114: T.int32 = cse_var_115 + 88
- cse_var_113: T.int32 = cse_var_115 + 87
- cse_var_112: T.int32 = cse_var_115 + 86
- cse_var_111: T.int32 = cse_var_115 + 85
- cse_var_110: T.int32 = cse_var_115 + 84
- cse_var_109: T.int32 = cse_var_115 + 83
- cse_var_108: T.int32 = cse_var_115 + 82
- cse_var_107: T.int32 = cse_var_115 + 7
- cse_var_106: T.int32 = cse_var_115 + 61
- cse_var_105: T.int32 = cse_var_115 + 60
- cse_var_104: T.int32 = cse_var_115 + 6
- cse_var_103: T.int32 = cse_var_115 + 59
- cse_var_102: T.int32 = cse_var_115 + 58
- cse_var_101: T.int32 = cse_var_115 + 57
- cse_var_100: T.int32 = cse_var_115 + 56
- cse_var_99: T.int32 = cse_var_115 + 55
- cse_var_98: T.int32 = cse_var_115 + 5
- cse_var_97: T.int32 = cse_var_115 + 412
- cse_var_96: T.int32 = cse_var_115 + 411
- cse_var_95: T.int32 = cse_var_115 + 410
- cse_var_94: T.int32 = cse_var_115 + 409
- cse_var_93: T.int32 = cse_var_115 + 408
- cse_var_92: T.int32 = cse_var_115 + 407
- cse_var_91: T.int32 = cse_var_115 + 406
- cse_var_90: T.int32 = cse_var_115 + 4
- cse_var_89: T.int32 = cse_var_115 + 385
- cse_var_88: T.int32 = cse_var_115 + 384
- cse_var_87: T.int32 = cse_var_115 + 383
- cse_var_86: T.int32 = cse_var_115 + 382
- cse_var_85: T.int32 = cse_var_115 + 381
- cse_var_84: T.int32 = cse_var_115 + 380
- cse_var_83: T.int32 = cse_var_115 + 379
- cse_var_82: T.int32 = cse_var_115 + 358
- cse_var_81: T.int32 = cse_var_115 + 357
- cse_var_80: T.int32 = cse_var_115 + 356
- cse_var_79: T.int32 = cse_var_115 + 355
- cse_var_78: T.int32 = cse_var_115 + 354
- cse_var_77: T.int32 = cse_var_115 + 353
- cse_var_76: T.int32 = cse_var_115 + 352
- cse_var_75: T.int32 = cse_var_115 + 34
- cse_var_74: T.int32 = cse_var_115 + 331
- cse_var_73: T.int32 = cse_var_115 + 330
- cse_var_72: T.int32 = cse_var_115 + 33
- cse_var_71: T.int32 = cse_var_115 + 329
- cse_var_70: T.int32 = cse_var_115 + 328
- cse_var_69: T.int32 = cse_var_115 + 327
- cse_var_68: T.int32 = cse_var_115 + 326
- cse_var_67: T.int32 = cse_var_115 + 325
- cse_var_66: T.int32 = cse_var_115 + 32
- cse_var_65: T.int32 = cse_var_115 + 31
- cse_var_64: T.int32 = cse_var_115 + 304
- cse_var_63: T.int32 = cse_var_115 + 303
- cse_var_62: T.int32 = cse_var_115 + 302
- cse_var_61: T.int32 = cse_var_115 + 301
- cse_var_60: T.int32 = cse_var_115 + 300
- cse_var_59: T.int32 = cse_var_115 + 30
- cse_var_58: T.int32 = cse_var_115 + 3
- cse_var_57: T.int32 = cse_var_115 + 299
- cse_var_56: T.int32 = cse_var_115 + 298
- cse_var_55: T.int32 = cse_var_115 + 29
- cse_var_54: T.int32 = cse_var_115 + 28
- cse_var_53: T.int32 = cse_var_115 + 277
- cse_var_52: T.int32 = cse_var_115 + 276
- cse_var_51: T.int32 = cse_var_115 + 275
- cse_var_50: T.int32 = cse_var_115 + 274
- cse_var_49: T.int32 = cse_var_115 + 273
- cse_var_48: T.int32 = cse_var_115 + 272
- cse_var_47: T.int32 = cse_var_115 + 271
- cse_var_46: T.int32 = cse_var_115 + 250
- cse_var_45: T.int32 = cse_var_115 + 249
- cse_var_44: T.int32 = cse_var_115 + 248
- cse_var_43: T.int32 = cse_var_115 + 247
- cse_var_42: T.int32 = cse_var_115 + 246
- cse_var_41: T.int32 = cse_var_115 + 245
- cse_var_40: T.int32 = cse_var_115 + 244
- cse_var_39: T.int32 = cse_var_115 + 223
- cse_var_38: T.int32 = cse_var_115 + 222
- cse_var_37: T.int32 = cse_var_115 + 221
- cse_var_36: T.int32 = cse_var_115 + 220
- cse_var_35: T.int32 = cse_var_115 + 219
- cse_var_34: T.int32 = cse_var_115 + 218
- cse_var_33: T.int32 = cse_var_115 + 217
- cse_var_32: T.int32 = cse_var_115 + 2
- cse_var_31: T.int32 = cse_var_115 + 196
- cse_var_30: T.int32 = cse_var_115 + 195
- cse_var_29: T.int32 = cse_var_115 + 194
- cse_var_28: T.int32 = cse_var_115 + 193
- cse_var_27: T.int32 = cse_var_115 + 192
- cse_var_26: T.int32 = cse_var_115 + 191
- cse_var_25: T.int32 = cse_var_115 + 190
- cse_var_24: T.int32 = cse_var_115 + 169
- cse_var_23: T.int32 = cse_var_115 + 168
- cse_var_22: T.int32 = cse_var_115 + 167
- cse_var_21: T.int32 = cse_var_115 + 166
- cse_var_20: T.int32 = cse_var_115 + 165
- cse_var_19: T.int32 = cse_var_115 + 164
- cse_var_18: T.int32 = cse_var_115 + 163
- cse_var_17: T.int32 = cse_var_115 + 142
- cse_var_16: T.int32 = cse_var_115 + 141
- cse_var_15: T.int32 = cse_var_115 + 140
- cse_var_14: T.int32 = cse_var_115 + 139
- cse_var_13: T.int32 = cse_var_115 + 138
- cse_var_12: T.int32 = cse_var_115 + 137
- cse_var_11: T.int32 = cse_var_115 + 136
- cse_var_10: T.int32 = cse_var_115 + 115
- cse_var_9: T.int32 = cse_var_115 + 114
- cse_var_8: T.int32 = cse_var_115 + 113
- cse_var_7: T.int32 = cse_var_115 + 112
- cse_var_6: T.int32 = cse_var_115 + 111
- cse_var_5: T.int32 = cse_var_115 + 110
- cse_var_4: T.int32 = cse_var_115 + 109
- cse_var_3: T.int32 = cse_var_115 + 1
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_3] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_32] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 27] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_54] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_55] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 54] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_99] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_100] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 81] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_108] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_109] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 108] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_4] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_5] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 135] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_11] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_12] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 162] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_18] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_19] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 189] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_25] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_26] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 216] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_33] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_34] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 243] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_40] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_41] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 270] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_47] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_48] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 297] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_56] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_57] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 324] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_67] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_68] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 351] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_76] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_77] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 378] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_83] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_84] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 405] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_91] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_92] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_3] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_32] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_58] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_54] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_55] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_59] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_99] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_100] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_101] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_108] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_109] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_110] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_4] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_5] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_6] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_11] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_12] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_13] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_18] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_19] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_20] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_25] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_26] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_27] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_33] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_34] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_35] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_40] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_41] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_42] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_47] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_48] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_49] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_56] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_57] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_60] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_67] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_68] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_69] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_76] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_77] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_78] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_83] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_84] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_85] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_91] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_92] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_93] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_32] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_58] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_90] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_55] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_59] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_65] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_100] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_101] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_102] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_109] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_110] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_111] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_5] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_6] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_7] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_12] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_13] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_14] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_19] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_20] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_21] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_26] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_27] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_28] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_34] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_35] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_36] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_41] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_42] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_43] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_48] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_49] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_50] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_57] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_60] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_61] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_68] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_69] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_70] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_77] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_78] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_79] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_84] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_85] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_86] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_92] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_93] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_94] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_58] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_90] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_98] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_59] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_65] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_66] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_101] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_102] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_103] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_110] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_111] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_112] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_6] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_7] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_8] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_13] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_14] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_15] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_20] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_21] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_22] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_27] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_28] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_29] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_35] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_36] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_37] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_42] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_43] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_44] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_49] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_50] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_51] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_60] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_61] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_62] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_69] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_70] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_71] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_78] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_79] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_80] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_85] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_86] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_87] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_93] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_94] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_95] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_90] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_98] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_104] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_65] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_66] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_72] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_102] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_103] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_105] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_111] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_112] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_113] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_7] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_8] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_9] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_14] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_15] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_16] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_21] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_22] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_23] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_28] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_29] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_30] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_36] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_37] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_38] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_43] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_44] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_45] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_50] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_51] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_52] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_61] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_62] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_63] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_70] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_71] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_73] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_79] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_80] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_81] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_86] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_87] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_88] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_94] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_95] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_96] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_98] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_104] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_107] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_66] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_72] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_75] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_103] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_105] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_106] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_112] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_113] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_114] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_8] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_9] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_10] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_15] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_16] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_17] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_22] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_23] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_24] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_29] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_30] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_31] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_37] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_38] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_39] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_44] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_45] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_46] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_51] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_52] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_53] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_62] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_63] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_64] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_71] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_73] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_74] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_80] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_81] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_82] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_87] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_88] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_89] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_95] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_96] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_97] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_104] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_107] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 8] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_72] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_75] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 35] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_105] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_106] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 62] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_113] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_114] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 89] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_9] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_10] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 116] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_16] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_17] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 143] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_23] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_24] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 170] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_30] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_31] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 197] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_38] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_39] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 224] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_45] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_46] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 251] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_52] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_53] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 278] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_63] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_64] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 305] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_73] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_74] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 332] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_81] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_82] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 359] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_88] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_89] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 386] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_96] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_97] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 413] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- 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 // 7 * 1568 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + i3_inner] = T.max(conv2d_nchw_1[i3_inner] + bias_1[blockIdx_x // 7 * 32 + threadIdx_x], T.float32(0))
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 128] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 128) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 256] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 256) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 384] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 384) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 512] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 512) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 640] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 640) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 768] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 768) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 896] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 896) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1024] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1024) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1152] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 73728]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1280] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1280) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1408] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1408) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1536] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1536) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1664] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1664) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1792] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1792) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1920] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1920) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2048] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2048) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2176] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2176) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2304] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 147456]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2432] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2432) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2560] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2560) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2688] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2688) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2816] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2816) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2944] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2944) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3072] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3072) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3200] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3200) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3328] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3328) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3456] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 221184]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3584] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3584) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3712] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3712) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3840] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3840) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3968] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3968) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4096] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4096) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4224] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4224) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4352] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4352) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4480] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4480) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4608] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 294912]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4736] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4736) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4864] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4864) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4992] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4992) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5120] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5120) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5248] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5248) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5376] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5376) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5504] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5504) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5632] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5632) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5760] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 368640]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5888] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5888) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6016] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6016) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6144] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6144) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6272] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6272) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6400] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6400) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6528] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6528) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6656] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6656) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6784] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6784) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6912] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 442368]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7040] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7040) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7168] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7168) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7296] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7296) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7424] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7424) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7552] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7552) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7680] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7680) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7808] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7808) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7936] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7936) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8064] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 516096]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8192] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8192) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8320] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8320) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8448] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8448) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8576] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8576) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8704] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8704) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8832] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8832) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8960] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8960) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 9088] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 9088) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[0] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[8] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[27] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[35] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[54] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[62] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[81] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[82] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[83] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[84] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[85] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[86] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[87] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[82] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[83] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[84] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[85] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[86] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[87] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[88] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[83] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[84] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[85] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[86] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[87] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[88] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[89] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[108] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[109] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[110] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[111] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[112] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[113] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[114] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[109] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[110] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[111] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[112] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[113] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[114] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[115] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[110] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[111] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[112] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[113] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[114] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[115] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[116] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[135] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[136] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[137] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[138] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[139] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[140] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[141] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[136] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[137] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[138] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[139] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[140] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[141] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[142] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[137] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[138] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[139] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[140] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[141] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[142] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[143] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[162] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[163] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[164] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[165] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[166] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[167] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[168] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[163] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[164] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[165] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[166] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[167] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[168] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[169] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[164] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[165] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[166] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[167] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[168] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[169] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[170] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[189] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[190] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[191] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[192] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[193] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[194] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[195] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[190] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[191] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[192] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[193] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[194] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[195] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[196] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[191] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[192] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[193] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[194] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[195] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[196] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[197] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[9] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[17] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[36] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[44] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[63] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[71] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[90] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[91] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[92] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[93] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[94] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[95] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[96] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[91] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[92] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[93] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[94] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[95] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[96] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[97] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[92] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[93] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[94] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[95] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[96] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[97] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[98] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[117] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[118] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[119] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[120] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[121] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[122] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[123] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[118] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[119] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[120] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[121] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[122] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[123] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[124] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[119] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[120] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[121] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[122] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[123] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[124] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[125] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[144] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[145] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[146] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[147] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[148] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[149] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[150] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[145] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[146] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[147] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[148] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[149] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[150] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[151] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[146] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[147] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[148] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[149] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[150] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[151] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[152] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[171] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[172] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[173] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[174] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[175] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[176] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[177] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[172] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[173] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[174] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[175] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[176] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[177] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[178] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[173] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[174] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[175] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[176] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[177] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[178] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[179] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[198] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[199] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[200] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[201] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[202] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[203] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[204] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[199] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[200] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[201] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[202] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[203] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[204] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[205] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[200] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[201] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[202] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[203] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[204] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[205] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[206] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[18] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[26] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[45] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[53] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[72] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[73] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[74] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[75] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[76] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[77] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[78] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[73] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[74] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[75] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[76] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[77] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[78] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[79] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[74] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[75] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[76] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[77] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[78] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[79] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[80] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[99] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[100] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[101] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[102] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[103] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[104] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[105] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[100] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[101] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[102] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[103] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[104] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[105] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[106] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[101] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[102] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[103] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[104] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[105] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[106] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[107] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[126] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[127] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[128] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[129] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[130] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[131] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[132] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[127] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[128] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[129] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[130] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[131] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[132] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[133] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[128] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[129] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[130] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[131] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[132] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[133] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[134] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[153] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[154] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[155] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[156] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[157] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[158] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[159] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[154] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[155] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[156] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[157] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[158] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[159] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[160] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[155] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[156] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[157] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[158] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[159] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[160] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[161] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[180] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[181] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[182] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[183] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[184] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[185] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[186] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[181] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[182] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[183] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[184] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[185] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[186] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[187] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[182] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[183] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[184] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[185] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[186] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[187] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[188] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[207] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[208] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[209] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[210] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[211] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[212] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[213] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[208] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[209] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[210] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[211] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[212] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[213] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[214] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[209] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[210] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[211] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[212] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[213] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[214] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[215] * kernel_shared_1[threadIdx_x * 72 + 71]
+ compute_1 = T.Buffer((25088,), data=compute.data)
+ bias_1 = T.Buffer((512,), data=bias.data)
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7] = T.max(conv2d_nchw_1[0] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 1] = T.max(conv2d_nchw_1[1] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 2] = T.max(conv2d_nchw_1[2] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 3] = T.max(conv2d_nchw_1[3] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 4] = T.max(conv2d_nchw_1[4] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 5] = T.max(conv2d_nchw_1[5] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 6] = T.max(conv2d_nchw_1[6] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
@@ -1087,7 +977,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 0.251 ms
+ Execution time of this operator: 0.216 ms
@@ -1137,17 +1027,17 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
conv2d_nchw_nn_o_o_o_o, conv2d_nchw_nn_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_o_i, factor=1)
conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=1)
conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=1)
- conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=32)
+ conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=128)
conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=1)
conv2d_nchw_yy_o_i, conv2d_nchw_yy_i = s[conv2d_nchw].split(conv2d_nchw_yy, factor=1)
conv2d_nchw_yy_o_o_i, conv2d_nchw_yy_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_i, factor=1)
conv2d_nchw_yy_o_o_o_i, conv2d_nchw_yy_o_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_o_i, factor=1)
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_o_i, conv2d_nchw_xx_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_i, 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=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=16)
+ 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=7)
+ conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=8)
conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=1)
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)
@@ -1158,14 +1048,14 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
compute_i0_o_o_i, compute_i0_o_i = s[compute].split(compute_i0_o_i, factor=1)
compute_i0_o_o_o, compute_i0_o_o_i = s[compute].split(compute_i0_o_o_i, factor=1)
compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=1)
- compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=32)
+ compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=128)
compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=1)
compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=1)
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=7)
+ 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=1)
- compute_i3_o_o_o, compute_i3_o_o_i = s[compute].split(compute_i3_o_o_i, factor=1)
+ compute_i3_o_o_o, compute_i3_o_o_i = s[compute].split(compute_i3_o_o_i, factor=7)
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)
kernel_shared = s.cache_read(kernel, "shared", [conv2d_nchw])
@@ -1184,14 +1074,14 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[kernel_shared].fuse(kernel_shared_ax0, kernel_shared_ax1, kernel_shared_ax2, kernel_shared_ax3)
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
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=32)
+ 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=128)
s[kernel_shared].bind(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[pad_temp_shared].fuse(pad_temp_shared_ax0, pad_temp_shared_ax1, pad_temp_shared_ax2, pad_temp_shared_ax3)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
s[pad_temp_shared].vectorize(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
- pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=32)
+ 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=128)
s[pad_temp_shared].bind(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
- s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 512)
+ s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 1024)
s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
@@ -1209,10 +1099,10 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
#define int64_t long long
#define uint64_t unsigned long long
#endif
- extern "C" __global__ void __launch_bounds__(32) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
+ extern "C" __global__ void __launch_bounds__(128) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
float conv2d_nchw[7];
- __shared__ float pad_temp_shared[432];
- __shared__ float kernel_shared[4608];
+ __shared__ float pad_temp_shared[216];
+ __shared__ float kernel_shared[9216];
conv2d_nchw[0] = 0.000000e+00f;
conv2d_nchw[1] = 0.000000e+00f;
conv2d_nchw[2] = 0.000000e+00f;
@@ -1220,511 +1110,597 @@ They can be used for debugging and learning the behavior of the auto-scheduler.
conv2d_nchw[4] = 0.000000e+00f;
conv2d_nchw[5] = 0.000000e+00f;
conv2d_nchw[6] = 0.000000e+00f;
- for (int rc_outer_outer = 0; rc_outer_outer < 32; ++rc_outer_outer) {
+ for (int rc_outer_outer = 0; rc_outer_outer < 64; ++rc_outer_outer) {
__syncthreads();
- pad_temp_shared[((int)threadIdx.x)] = (((((1 <= (((((int)threadIdx.x) % 27) / 9) + (((int)blockIdx.x) % 7))) && ((((((int)threadIdx.x) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + ((((int)threadIdx.x) / 27) * 49)) + (((((int)threadIdx.x) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 32)] = (((((1 <= ((((((int)threadIdx.x) + 5) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 5) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 32) / 27) * 49)) + ((((((int)threadIdx.x) + 5) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] : 0.0 [...]
- pad_temp_shared[(((int)threadIdx.x) + 64)] = (((((1 <= ((((((int)threadIdx.x) + 10) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 10) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 64) / 27) * 49)) + ((((((int)threadIdx.x) + 10) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : [...]
- pad_temp_shared[(((int)threadIdx.x) + 96)] = (((((1 <= ((((((int)threadIdx.x) + 15) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 15) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 96) / 27) * 49)) + ((((((int)threadIdx.x) + 15) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : [...]
- pad_temp_shared[(((int)threadIdx.x) + 128)] = (((((1 <= ((((((int)threadIdx.x) + 20) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 20) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 128) / 27) * 49)) + ((((((int)threadIdx.x) + 20) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] [...]
- pad_temp_shared[(((int)threadIdx.x) + 160)] = (((((1 <= ((((((int)threadIdx.x) + 25) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 25) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 160) / 27) * 49)) + ((((((int)threadIdx.x) + 25) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] [...]
- pad_temp_shared[(((int)threadIdx.x) + 192)] = (((((1 <= ((((((int)threadIdx.x) + 3) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 3) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 192) / 27) * 49)) + ((((((int)threadIdx.x) + 3) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0 [...]
- pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((1 <= ((((((int)threadIdx.x) + 8) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 8) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 224) / 27) * 49)) + ((((((int)threadIdx.x) + 8) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0 [...]
- pad_temp_shared[(((int)threadIdx.x) + 256)] = (((((1 <= ((((((int)threadIdx.x) + 13) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 13) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 256) / 27) * 49)) + ((((((int)threadIdx.x) + 13) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] [...]
- pad_temp_shared[(((int)threadIdx.x) + 288)] = (((((1 <= ((((int)blockIdx.x) % 7) + (((((int)threadIdx.x) / 9) + 2) % 3))) && (((((int)blockIdx.x) % 7) + (((((int)threadIdx.x) / 9) + 2) % 3)) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 288) / 27) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + ((((((int)threadIdx.x) / 9) + 2) % 3) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 320)] = (((((1 <= ((((((int)threadIdx.x) + 23) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 23) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 320) / 27) * 49)) + ((((((int)threadIdx.x) + 23) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 5) % 9)) - 8)] [...]
- pad_temp_shared[(((int)threadIdx.x) + 352)] = (((((1 <= ((((((int)threadIdx.x) + 1) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 1) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 352) / 27) * 49)) + ((((((int)threadIdx.x) + 1) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 1) % 9)) - 8)] : 0 [...]
- pad_temp_shared[(((int)threadIdx.x) + 384)] = (((((1 <= ((((((int)threadIdx.x) + 6) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 6) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 384) / 27) * 49)) + ((((((int)threadIdx.x) + 6) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 6) % 9)) - 8)] : 0 [...]
- if (((int)threadIdx.x) < 16) {
- pad_temp_shared[(((int)threadIdx.x) + 416)] = (((((((((int)threadIdx.x) + 11) / 9) + (((int)blockIdx.x) % 7)) < 8) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 416) / 27) * 49)) + (((((int)threadIdx.x) + 11) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[((int)threadIdx.x)] = (((((1 <= (((((int)threadIdx.x) % 27) / 9) + (((int)blockIdx.x) % 7))) && ((((((int)threadIdx.x) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((((rc_outer_outer * 392) + ((((int)threadIdx.x) / 27) * 49)) + (((((int)threadIdx.x) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
+ if (((int)threadIdx.x) < 88) {
+ pad_temp_shared[(((int)threadIdx.x) + 128)] = (((((1 <= ((((((int)threadIdx.x) + 20) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 20) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[((((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 128) / 27) * 49)) + ((((((int)threadIdx.x) + 20) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8) [...]
}
- kernel_shared[((int)threadIdx.x)] = kernel[((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x))];
- kernel_shared[(((int)threadIdx.x) + 32)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 64)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 96)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 128)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 128) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 160)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 160) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 192)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 192) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 224)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 224) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 256)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 256) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 288)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 9216)];
- kernel_shared[(((int)threadIdx.x) + 320)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 320) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 352)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 352) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 384)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 384) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 416)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 416) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 448)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 448) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 480)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 480) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 512)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 512) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 544)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 544) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 576)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 18432)];
- kernel_shared[(((int)threadIdx.x) + 608)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 608) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 640)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 640) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 672)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 672) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 704)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 704) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 736)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 736) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 768)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 768) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 800)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 800) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 832)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 832) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 864)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 27648)];
- kernel_shared[(((int)threadIdx.x) + 896)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 896) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 928)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 928) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 960)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 960) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 992)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 992) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1024)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1024) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1056)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1056) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 1088)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1088) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1120)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1120) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1152)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 36864)];
- kernel_shared[(((int)threadIdx.x) + 1184)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1184) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1216)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1216) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1248)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1248) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 1280)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1280) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1312)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1312) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1344) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 1376)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1376) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1408)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1408) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1440)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 46080)];
- kernel_shared[(((int)threadIdx.x) + 1472)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1472) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1504)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1504) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1536)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1536) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 1568)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1568) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1600)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1600) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1632)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1632) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 1664)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1664) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1696)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1696) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1728)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 55296)];
- kernel_shared[(((int)threadIdx.x) + 1760)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1760) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1792)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1792) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1824)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1824) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 1856)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1856) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1888)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1888) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1920)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1920) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 1952)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1952) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1984)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1984) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2016)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 64512)];
- kernel_shared[(((int)threadIdx.x) + 2048)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2048) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2080)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2080) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2112)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2112) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 2144)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2144) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2176)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2176) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2208)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2208) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 2240)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2240) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2272)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2272) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2304)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 73728)];
- kernel_shared[(((int)threadIdx.x) + 2336)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2336) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2368)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2368) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2400)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2400) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 2432)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2432) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2464)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2464) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2496)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2496) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 2528)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2528) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2560)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2560) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2592)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 82944)];
- kernel_shared[(((int)threadIdx.x) + 2624)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2624) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2656)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2656) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2688)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2688) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 2720)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2720) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2752)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2752) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2784)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2784) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 2816)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2816) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2848)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2848) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2880)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 92160)];
- kernel_shared[(((int)threadIdx.x) + 2912)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2912) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2944)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2944) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2976)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2976) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 3008)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3008) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3040)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3040) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3072)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3072) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 3104)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3104) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3136)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3136) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3168)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 101376)];
- kernel_shared[(((int)threadIdx.x) + 3200)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3200) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3232)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3232) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3264)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3264) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 3296)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3296) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3328)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3328) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3360)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3360) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 3392)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3392) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3424)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3424) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3456)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 110592)];
- kernel_shared[(((int)threadIdx.x) + 3488)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3488) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3520)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3520) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3552)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3552) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 3584)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3584) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3616)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3616) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3648)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3648) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 3680)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3680) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3712)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3712) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3744)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 119808)];
- kernel_shared[(((int)threadIdx.x) + 3776)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3776) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3808)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3808) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3840)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3840) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 3872)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3872) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3904)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3904) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3936)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3936) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 3968)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3968) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4000)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4000) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4032)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 129024)];
- kernel_shared[(((int)threadIdx.x) + 4064)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4064) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4096)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4096) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4128)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4128) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 4160)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4160) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4192)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4192) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4224)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4224) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 4256)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4256) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4288)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4288) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4320)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 138240)];
- kernel_shared[(((int)threadIdx.x) + 4352)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4352) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4384)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4384) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4416)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4416) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 4448)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4448) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4480)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4480) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4512)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4512) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 4544)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4544) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4576)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4576) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((int)threadIdx.x)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 128)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 128) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 256)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 256) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 384)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 384) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 512)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 512) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 640)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 640) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 768)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 768) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 896)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 896) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1024)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1024) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1152)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 73728)];
+ kernel_shared[(((int)threadIdx.x) + 1280)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1280) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1408)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1408) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1536)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1536) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1664)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1664) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 1792)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1792) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1920)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1920) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2048)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2048) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2176)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2176) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2304)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 147456)];
+ kernel_shared[(((int)threadIdx.x) + 2432)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2432) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2560)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2560) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2688)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2688) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2816)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2816) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 2944)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2944) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3072)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3072) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3200)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3200) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3328)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3328) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3456)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 221184)];
+ kernel_shared[(((int)threadIdx.x) + 3584)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3584) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3712)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3712) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3840)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3840) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3968)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3968) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 4096)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4096) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4224)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4224) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4352)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4352) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4480)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4480) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4608)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 294912)];
+ kernel_shared[(((int)threadIdx.x) + 4736)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4736) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4864)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4864) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4992)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4992) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5120)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5120) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 5248)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5248) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5376)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5376) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5504)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5504) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5632)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5632) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5760)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 368640)];
+ kernel_shared[(((int)threadIdx.x) + 5888)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5888) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6016)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6016) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6144)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6144) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6272)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6272) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 6400)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6400) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6528)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6528) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6656)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6656) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6784)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6784) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6912)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 442368)];
+ kernel_shared[(((int)threadIdx.x) + 7040)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7040) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7168)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7168) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7296)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7296) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7424)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7424) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 7552)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7552) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7680)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7680) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7808)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7808) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7936)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7936) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8064)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 516096)];
+ kernel_shared[(((int)threadIdx.x) + 8192)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8192) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8320)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8320) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8448)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8448) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8576)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8576) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 8704)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8704) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8832)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8832) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8960)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8960) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 9088)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 9088) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
__syncthreads();
- for (int ry_outer_inner = 0; ry_outer_inner < 3; ++ry_outer_inner) {
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(ry_outer_inner * 9)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 27)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 28)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 29)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 54)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 55)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 56)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 81)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 82)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 108)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 109)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 110)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 135)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 136)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 137)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 162)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 163)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 216)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 217)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 218)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 243)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 244)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 270)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 271)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 272)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 297)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 298)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 299)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 324)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 325)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 351)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 352)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 353)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 405)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 406)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 1)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 28)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 29)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 30)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 55)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 56)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 57)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 82)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 109)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 110)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 111)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 136)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 137)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 138)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 163)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 217)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 218)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 219)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 244)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 271)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 272)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 273)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 298)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 299)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 300)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 325)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 352)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 353)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 354)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 406)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 2)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 29)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 30)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 31)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 56)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 57)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 58)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 110)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 111)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 112)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 137)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 138)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 139)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 218)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 219)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 220)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 272)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 273)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 274)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 299)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 300)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 301)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 353)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 354)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 355)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 3)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 30)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 31)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 32)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 57)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 58)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 59)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 111)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 112)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 113)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 138)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 139)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 140)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 219)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 220)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 221)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 273)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 274)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 275)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 300)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 301)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 302)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 354)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 355)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 356)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 4)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 31)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 32)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 33)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 58)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 59)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 60)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 112)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 113)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 114)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 139)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 140)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 141)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 220)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 221)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 222)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 274)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 275)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 276)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 301)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 302)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 303)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 355)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 356)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 357)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 5)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 32)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 33)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 34)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 59)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 60)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 61)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 88)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 113)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 114)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 115)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 140)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 141)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 142)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 169)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 221)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 222)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 223)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 250)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 275)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 276)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 277)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 302)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 303)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 304)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 331)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 356)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 357)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 358)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 412)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 6)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 33)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 34)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 35)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 60)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 61)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 62)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 88)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 89)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 114)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 115)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 116)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 141)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 142)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 143)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 169)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 170)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 222)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 223)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 224)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 250)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 251)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 276)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 277)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 278)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 303)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 304)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 305)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 331)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 332)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 357)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 358)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 359)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 412)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 413)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- }
- }
- for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
- compute[(((((((int)blockIdx.x) / 7) * 1568) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + i3_inner)] = max((conv2d_nchw[i3_inner] + bias[(((((int)blockIdx.x) / 7) * 32) + ((int)threadIdx.x))]), 0.000000e+00f);
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[0] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[1] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[2] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[3] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[4] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[5] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[6] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[1] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[8] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[27] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[35] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[54] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[62] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[81] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[82] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[83] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[84] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[85] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[86] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[87] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[82] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[83] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[84] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[85] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[86] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[87] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[88] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[83] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[84] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[85] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[86] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[87] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[88] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[89] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[108] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[109] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[110] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[111] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[112] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[113] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[114] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[109] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[110] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[111] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[112] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[113] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[114] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[115] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[110] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[111] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[112] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[113] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[114] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[115] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[116] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[135] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[136] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[137] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[138] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[139] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[140] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[141] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[136] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[137] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[138] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[139] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[140] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[141] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[142] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[137] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[138] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[139] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[140] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[141] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[142] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[143] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[162] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[163] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[164] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[165] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[166] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[167] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[168] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[163] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[164] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[165] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[166] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[167] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[168] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[169] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[164] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[165] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[166] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[167] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[168] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[169] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[170] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[189] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[190] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[191] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[192] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[193] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[194] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[195] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[190] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[191] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[192] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[193] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[194] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[195] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[196] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[191] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[192] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[193] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[194] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[195] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[196] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[197] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[9] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[17] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[36] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[44] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[63] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[71] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[90] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[91] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[92] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[93] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[94] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[95] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[96] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[91] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[92] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[93] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[94] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[95] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[96] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[97] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[92] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[93] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[94] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[95] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[96] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[97] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[98] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[117] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[118] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[119] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[120] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[121] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[122] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[123] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[118] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[119] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[120] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[121] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[122] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[123] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[124] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[119] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[120] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[121] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[122] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[123] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[124] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[125] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[144] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[145] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[146] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[147] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[148] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[149] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[150] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[145] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[146] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[147] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[148] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[149] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[150] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[151] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[146] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[147] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[148] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[149] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[150] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[151] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[152] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[171] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[172] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[173] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[174] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[175] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[176] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[177] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[172] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[173] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[174] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[175] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[176] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[177] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[178] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[173] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[174] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[175] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[176] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[177] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[178] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[179] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[198] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[199] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[200] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[201] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[202] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[203] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[204] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[199] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[200] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[201] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[202] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[203] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[204] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[205] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[200] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[201] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[202] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[203] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[204] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[205] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[206] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[18] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[26] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[45] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[53] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[72] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[73] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[74] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[75] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[76] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[77] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[78] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[73] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[74] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[75] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[76] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[77] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[78] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[79] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[74] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[75] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[76] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[77] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[78] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[79] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[80] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[99] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[100] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[101] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[102] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[103] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[104] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[105] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[100] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[101] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[102] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[103] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[104] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[105] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[106] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[101] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[102] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[103] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[104] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[105] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[106] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[107] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[126] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[127] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[128] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[129] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[130] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[131] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[132] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[127] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[128] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[129] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[130] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[131] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[132] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[133] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[128] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[129] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[130] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[131] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[132] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[133] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[134] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[153] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[154] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[155] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[156] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[157] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[158] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[159] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[154] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[155] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[156] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[157] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[158] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[159] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[160] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[155] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[156] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[157] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[158] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[159] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[160] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[161] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[180] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[181] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[182] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[183] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[184] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[185] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[186] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[181] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[182] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[183] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[184] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[185] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[186] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[187] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[182] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[183] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[184] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[185] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[186] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[187] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[188] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[207] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[208] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[209] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[210] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[211] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[212] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[213] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[208] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[209] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[210] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[211] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[212] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[213] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[214] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[209] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[210] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[211] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[212] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[213] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[214] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[215] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
}
+ compute[((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7))] = max((conv2d_nchw[0] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 1)] = max((conv2d_nchw[1] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 2)] = max((conv2d_nchw[2] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 3)] = max((conv2d_nchw[3] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 4)] = max((conv2d_nchw[4] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 5)] = max((conv2d_nchw[5] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 6)] = max((conv2d_nchw[6] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
}
@@ -1785,7 +1761,7 @@ In the example below we resume the status and do more 5 trials.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 5 minutes 42.565 seconds)
+ **Total running time of the script:** ( 5 minutes 35.291 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 b6e21daf11..cc2b655e00 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)
- 7.8563 7.8537 7.8646 7.8505 0.0060
+ 7.8773 7.8774 7.8840 7.8704 0.0055
@@ -675,7 +675,7 @@ Other Tips
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 5.839 seconds)
+ **Total running time of the script:** ( 1 minutes 6.161 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 eedc323582..ba93d098d7 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)
- 758.6698 757.9524 760.1585 757.8985 1.0529
+ 749.2774 750.8660 750.9075 746.0588 2.2759
@@ -694,7 +694,7 @@ Other Tips
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 39.693 seconds)
+ **Total running time of the script:** ( 1 minutes 39.443 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 09c11c3630..0f0dc3b45c 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
@@ -389,26 +389,74 @@ layout transformation, parallelization, vectorization, unrolling, and operator f
@T.prim_func
def main(placeholder: T.Buffer((128, 256), "float32"), placeholder_1: T.Buffer((4916, 16, 1), "float32"), placeholder_2: T.Buffer((4916,), "int32"), placeholder_3: T.Buffer((33,), "int32"), placeholder_4: T.Buffer((128, 512), "float32"), compute: T.Buffer((128, 512), "float32")):
T.func_attr({"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True})
- for i0_outer_i1_outer_fused in T.parallel(64):
- compute_1 = T.allocate([1024], "float32", "global")
- compute_2 = T.Buffer((1024,), data=compute_1)
+ for i0_outer_i1_outer_fused in T.parallel(16):
+ compute_1 = T.allocate([4096], "float32", "global")
+ compute_2 = T.Buffer((4096,), data=compute_1)
for i_outer_inner, nb_j_inner in T.grid(16, 2):
- for i_inner_init, j_init in T.grid(2, 16):
- compute_2[i_outer_inner * 64 + i_inner_init * 32 + nb_j_inner * 16 + j_init] = T.float32(0)
- for elem_idx, i_inner, j in T.grid(T.let(cse_var_1, i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner, placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]), 2, 16):
- cse_var_1 = T.var("int32")
+ for i_inner_init in range(8):
+ cse_var_1: T.int32 = i_outer_inner * 256 + 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)
+ compute_2[cse_var_1 + 3] = T.float32(0)
+ compute_2[cse_var_1 + 4] = T.float32(0)
+ compute_2[cse_var_1 + 5] = T.float32(0)
+ compute_2[cse_var_1 + 6] = T.float32(0)
+ compute_2[cse_var_1 + 7] = T.float32(0)
+ compute_2[cse_var_1 + 8] = T.float32(0)
+ compute_2[cse_var_1 + 9] = T.float32(0)
+ compute_2[cse_var_1 + 10] = T.float32(0)
+ compute_2[cse_var_1 + 11] = T.float32(0)
+ compute_2[cse_var_1 + 12] = T.float32(0)
+ 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(cse_var_2, i0_outer_i1_outer_fused * 2 + nb_j_inner, placeholder_5[cse_var_2 + 1] - placeholder_5[cse_var_2]), 8):
+ cse_var_2 = T.var("int32")
placeholder_5 = T.Buffer((33,), "int32", data=placeholder_3.data)
- cse_var_3: T.int32 = i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner
- cse_var_2: T.int32 = i_outer_inner * 64 + i_inner * 32 + nb_j_inner * 16 + j
+ cse_var_21: T.int32 = elem_idx * 16
+ cse_var_20: T.int32 = i0_outer_i1_outer_fused * 2 + nb_j_inner
+ cse_var_19: T.int32 = i_outer_inner * 2048 + i_inner * 256
+ cse_var_18: T.int32 = i_outer_inner * 256 + i_inner * 32 + nb_j_inner * 16
+ cse_var_17: T.int32 = cse_var_18 + 9
+ cse_var_16: T.int32 = cse_var_18 + 8
+ cse_var_15: T.int32 = cse_var_18 + 7
+ cse_var_14: T.int32 = cse_var_18 + 6
+ cse_var_13: T.int32 = cse_var_18 + 5
+ cse_var_12: T.int32 = cse_var_18 + 4
+ cse_var_11: T.int32 = cse_var_18 + 3
+ cse_var_10: T.int32 = cse_var_18 + 2
+ cse_var_9: T.int32 = cse_var_18 + 15
+ cse_var_8: T.int32 = cse_var_18 + 14
+ cse_var_7: T.int32 = cse_var_18 + 13
+ cse_var_6: T.int32 = cse_var_18 + 12
+ cse_var_5: T.int32 = cse_var_18 + 11
+ cse_var_4: T.int32 = cse_var_18 + 10
+ cse_var_3: T.int32 = cse_var_18 + 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_2] = compute_2[cse_var_2] + placeholder_6[placeholder_5[cse_var_3] * 16 + elem_idx * 16 + j] * T.max(placeholder_7[i0_outer_i1_outer_fused // 16 * 8192 + i_outer_inner * 512 + i_inner * 256 + placeholder_8[placeholder_5[cse_var_3] + elem_idx]], T.float32(0))
- for i0_inner in range(32):
- cse_var_4: T.int32 = i0_outer_i1_outer_fused // 16 * 16384 + i0_inner * 512 + i0_outer_i1_outer_fused % 16 * 32
+ compute_2[cse_var_18] = compute_2[cse_var_18] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21] * T.max(placeholder_7[cse_var_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
+ for i0_inner in range(128):
+ cse_var_22: T.int32 = i0_inner * 512 + i0_outer_i1_outer_fused * 32
compute_3 = T.Buffer((65536,), data=compute.data)
placeholder_5 = T.Buffer((65536,), data=placeholder_4.data)
- compute_3[cse_var_4:cse_var_4 + 32] = T.max(compute_2[i0_inner * 32:i0_inner * 32 + 32] + placeholder_5[cse_var_4:cse_var_4 + 32], T.Broadcast(T.float32(0), 32))
+ compute_3[cse_var_22:cse_var_22 + 32] = T.max(compute_2[i0_inner * 32:i0_inner * 32 + 32] + placeholder_5[cse_var_22:cse_var_22 + 32], T.Broadcast(T.float32(0), 32))
@@ -458,7 +506,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 2.117 ms
+ Execution time of this operator: 1.868 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 9e695bce91..b7927e4715 100644
--- a/docs/_sources/how_to/tune_with_autotvm/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/tune_with_autotvm/sg_execution_times.rst.txt
@@ -5,16 +5,16 @@
Computation times
=================
-**00:43.161** total execution time for **how_to_tune_with_autotvm** files:
+**00:42.208** 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``) | 00:43.129 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_conv2d_cuda.py` (``tune_conv2d_cuda.py``) | 00:42.174 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_x86.py` (``tune_relay_x86.py``) | 00:00.019 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_x86.py` (``tune_relay_x86.py``) | 00:00.021 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_cuda.py` (``tune_relay_cuda.py``) | 00:00.005 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_arm.py` (``tune_relay_arm.py``) | 00:00.005 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_arm.py` (``tune_relay_arm.py``) | 00:00.004 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_mobile_gpu.py` (``tune_relay_mobile_gpu.py``) | 00:00.004 | 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 3389b10200..5167246013 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
@@ -390,9 +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, 1, 4, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 8, 64]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8495099
- No: 2 GFLOPS: 90.01/90.01 result: MeasureResult(costs=(0.002572052,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.9165120124816895, timestamp=1674803118.8732517) [('tile_f', [-1, 8, 4, 8]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,41072
- No: 3 GFLOPS: 0.00/90.01 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 128, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2967167
+ 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
@@ -514,8 +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, 1, 8, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 16]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7487659
- No: 4 GFLOPS: 0.00/90.01 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 2, 32]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 32, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9986432
+ 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
@@ -637,8 +636,26 @@ 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, 2, 64]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9392665
- No: 5 GFLOPS: 0.00/90.01 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 64, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 64, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1637115
+ No: 4 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
+ File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 142, in build
+ res = future.result()
+ File "/usr/lib/python3.7/concurrent/futures/_base.py", line 435, in result
+ return self.__get_result()
+ File "/usr/lib/python3.7/concurrent/futures/_base.py", line 384, in __get_result
+ raise self._exception
+ File "/usr/lib/python3.7/concurrent/futures/thread.py", line 57, in run
+ result = self.fn(*self.args, **self.kwargs)
+ File "/workspace/python/tvm/contrib/popen_pool.py", line 432, in <lambda>
+ worker = lambda *args: self._worker_run(*args)
+ File "/workspace/python/tvm/contrib/popen_pool.py", line 401, in _worker_run
+ return proc.recv()
+ File "/workspace/python/tvm/contrib/popen_pool.py", line 309, in recv
+ raise TimeoutError()
+ TimeoutError
+
+ [('tile_f', [-1, 32, 2, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 4]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8784933
+ No: 5 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
@@ -760,8 +777,12 @@ 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, 16, 1]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 512]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2707796
- No: 6 GFLOPS: 0.00/90.01 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 8, 2, 16]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 16, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6792673
+ No: 6 GFLOPS: 38.05/38.05 result: MeasureResult(costs=(0.0060844850526315785,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.9628829956054688, timestamp=1674812669.290498) [('tile_f', [-1, 4, 1, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4942797
+ No: 7 GFLOPS: 279.09/279.09 result: MeasureResult(costs=(0.0008294822422680412,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.5810739994049072, timestamp=1674812671.22481) [('tile_f', [-1, 1, 4, 4]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 8, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,399855
+ No: 8 GFLOPS: 63.50/279.09 result: MeasureResult(costs=(0.0036458056071428566,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.4576036930084229, timestamp=1674812671.9696527) [('tile_f', [-1, 16, 16, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4877438
+ No: 9 GFLOPS: 5.24/279.09 result: MeasureResult(costs=(0.044172023750000004,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.1857759952545166, timestamp=1674812674.3203592) [('tile_f', [-1, 8, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5691650
+ No: 10 GFLOPS: 0.00/279.09 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
@@ -883,27 +904,8 @@ 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, 4, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 8]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2806473
- No: 7 GFLOPS: 164.23/164.23 result: MeasureResult(costs=(0.0014096323918918918,), error_no=MeasureErrorNo.NO_ERROR, all_cost=10.716526985168457, timestamp=1674803132.9284441) [('tile_f', [-1, 2, 2, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 8, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8844285
- No: 8 GFLOPS: 0.00/164.23 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 142, in build
- res = future.result()
- File "/usr/lib/python3.7/concurrent/futures/_base.py", line 435, in result
- return self.__get_result()
- File "/usr/lib/python3.7/concurrent/futures/_base.py", line 384, in __get_result
- raise self._exception
- File "/usr/lib/python3.7/concurrent/futures/thread.py", line 57, in run
- result = self.fn(*self.args, **self.kwargs)
- File "/workspace/python/tvm/contrib/popen_pool.py", line 432, in <lambda>
- worker = lambda *args: self._worker_run(*args)
- File "/workspace/python/tvm/contrib/popen_pool.py", line 401, in _worker_run
- return proc.recv()
- File "/workspace/python/tvm/contrib/popen_pool.py", line 309, in recv
- raise TimeoutError()
- TimeoutError
-
- [('tile_f', [-1, 4, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 64, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9704202
- No: 9 GFLOPS: 0.00/164.23 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 32, 4, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 16, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5629217
+ No: 11 GFLOPS: 0.00/279.09 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
@@ -1025,8 +1027,8 @@ 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, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 128, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7838471
- No: 10 GFLOPS: 0.00/164.23 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 1, 512]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 8]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2808079
+ No: 12 GFLOPS: 0.00/279.09 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
@@ -1148,8 +1150,8 @@ 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, 512, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 32]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,339294
- No: 11 GFLOPS: 0.00/164.23 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 4, 32]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 64, 4]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5899056
+ No: 13 GFLOPS: 0.00/279.09 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
@@ -1271,9 +1273,8 @@ 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, 256]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 16, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10117577
- No: 12 GFLOPS: 316.20/316.20 result: MeasureResult(costs=(0.0007321365147928994,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7440016269683838, timestamp=1674803135.8509939) [('tile_f', [-1, 4, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4070442
- No: 13 GFLOPS: 0.00/316.20 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 32, 1, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 512, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5066825
+ No: 14 GFLOPS: 0.00/279.09 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
@@ -1395,9 +1396,8 @@ 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, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 32, 8]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7471640
- No: 14 GFLOPS: 7.72/316.20 result: MeasureResult(costs=(0.02997332,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.854086637496948, timestamp=1674803141.9115317) [('tile_f', [-1, 4, 2, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 4]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2979825
- No: 15 GFLOPS: 0.00/316.20 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 64, 2, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 64]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1320730
+ No: 15 GFLOPS: 0.00/279.09 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
@@ -1519,11 +1519,8 @@ 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, 8, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 256]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10253255
- No: 16 GFLOPS: 111.35/316.20 result: MeasureResult(costs=(0.0020790924415584416,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.28901743888855, timestamp=1674803142.9225328) [('tile_f', [-1, 2, 4, 8]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7074250
- No: 17 GFLOPS: 58.21/316.20 result: MeasureResult(costs=(0.003976914307692308,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2880961894989014, timestamp=1674803144.4809651) [('tile_f', [-1, 4, 16, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 16]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1668736
- No: 18 GFLOPS: 18.61/316.20 result: MeasureResult(costs=(0.012438851222222222,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.455913782119751, timestamp=1674803145.2864993) [('tile_f', [-1, 8, 1, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 16, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6210218
- No: 19 GFLOPS: 0.00/316.20 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 2, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 256]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2508364
+ No: 16 GFLOPS: 0.00/279.09 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
@@ -1645,8 +1642,11 @@ 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, 2, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 16, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9428211
- No: 20 GFLOPS: 0.00/316.20 result: Traceback (most recent call last):
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 64, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 8, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2208083
+ No: 17 GFLOPS: 98.96/279.09 result: MeasureResult(costs=(0.002339420823529412,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.632169485092163, timestamp=1674812679.9650273) [('tile_f', [-1, 1, 2, 32]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,630
+ No: 18 GFLOPS: 71.38/279.09 result: MeasureResult(costs=(0.003243123677419355,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.0503921508789062, timestamp=1674812680.7025213) [('tile_f', [-1, 4, 32, 4]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8947532
+ No: 19 GFLOPS: 22.79/279.09 result: MeasureResult(costs=(0.010158083299999999,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.957541227340698, timestamp=1674812681.477137) [('tile_f', [-1, 2, 4, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2369576
+ No: 20 GFLOPS: 0.00/279.09 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
@@ -1768,7 +1768,7 @@ 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, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 256]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8705647
+ tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 256, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6425153
@@ -1823,9 +1823,9 @@ and measure running time.
Finish loading 20 records
Best config:
- [('tile_f', [-1, 4, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4070442
+ [('tile_f', [-1, 1, 4, 4]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 8, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,399855
Finish loading 20 records
- Time cost of this operator: 0.001051
+ Time cost of this operator: 0.001200
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 0d8296534b..a986ba0b4c 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
@@ -363,10 +363,10 @@ 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 311.0 98.701 (1, 2, 10, 10, 3) 2 1 [311.0]
- tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.116 0.989 (1, 6, 10, 10) 1 1 [3.116]
- tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.977 0.31 (1, 1, 10, 10, 3) 1 1 [0.977]
- Total_time - 315.092 - - - - -
+ tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 311.5 98.71 (1, 2, 10, 10, 3) 2 1 [311.5]
+ tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.108 0.985 (1, 6, 10, 10) 1 1 [3.108]
+ tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.963 0.305 (1, 1, 10, 10, 3) 1 1 [0.963]
+ Total_time - 315.571 - - - - -
@@ -431,10 +431,10 @@ 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 103.2 97.552 (1, 6, 10, 10, 1) 2 1 [103.2]
- tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.75 1.654 (1, 6, 10, 10) 1 1 [1.75]
- tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.84 0.794 (1, 3, 10, 10, 1) 1 1 [0.84]
- Total_time - 105.79 - - - - -
+ tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 100.5 97.322 (1, 6, 10, 10, 1) 2 1 [100.5]
+ tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.807 1.75 (1, 6, 10, 10) 1 1 [1.807]
+ tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.959 0.929 (1, 1, 10, 10, 3) 1 1 [0.959]
+ Total_time - 103.266 - - - - -
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 e6a9722cc3..a86eef607d 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
@@ -117,7 +117,7 @@ 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]
100%|##########| 3.42M/3.42M [00:00<00:00, 60.7MB/s]
+
0%| | 0.00/3.42M [00:00<?, ?B/s]
100%|##########| 3.42M/3.42M [00:00<00:00, 55.7MB/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.
@@ -322,7 +322,7 @@ Look up prediction top 1 index in 1000 class synset.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 9.814 seconds)
+ **Total running time of the script:** ( 1 minutes 10.103 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 c0cdd41c5c..d3cddb8b7b 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/tmp6krshqz2/images/random'
+ '/tmp/tmpllbfhusq/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: [0.0, 1.0], [1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0]
+ :alt: [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0]
: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/tmp6krshqz2/images/target contains 8144 images
- /tmp/tmp6krshqz2/images/random contains 5000 images
+ /tmp/tmpllbfhusq/images/target contains 8144 images
+ /tmp/tmpllbfhusq/images/random contains 5000 images
@@ -494,13 +494,13 @@ the time on our validation set).
.. code-block:: none
Epoch 1/3
- 328/328 - 47s - loss: 0.2294 - accuracy: 0.9215 - val_loss: 0.1200 - val_accuracy: 0.9596 - 47s/epoch - 144ms/step
+ 328/328 - 47s - loss: 0.2308 - accuracy: 0.9220 - val_loss: 0.1662 - val_accuracy: 0.9437 - 47s/epoch - 143ms/step
Epoch 2/3
- 328/328 - 44s - loss: 0.0962 - accuracy: 0.9643 - val_loss: 0.1094 - val_accuracy: 0.9626 - 44s/epoch - 133ms/step
+ 328/328 - 43s - loss: 0.1072 - accuracy: 0.9613 - val_loss: 0.1142 - val_accuracy: 0.9592 - 43s/epoch - 132ms/step
Epoch 3/3
- 328/328 - 43s - loss: 0.0663 - accuracy: 0.9743 - val_loss: 0.2085 - val_accuracy: 0.9415 - 43s/epoch - 133ms/step
+ 328/328 - 43s - loss: 0.0657 - accuracy: 0.9752 - val_loss: 0.1005 - val_accuracy: 0.9668 - 43s/epoch - 131ms/step
- <keras.callbacks.History object at 0x7fb3666e8e10>
+ <keras.callbacks.History object at 0x7f1830403810>
@@ -857,7 +857,7 @@ 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 22.236 seconds)
+ **Total running time of the script:** ( 4 minutes 33.046 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 5c6308f9c0..2e064b99dd 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
@@ -5,18 +5,18 @@
Computation times
=================
-**06:36.598** total execution time for **how_to_work_with_microtvm** files:
+**06:47.347** total execution time for **how_to_work_with_microtvm** files:
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``) | 04:22.236 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``) | 04:33.046 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_pytorch.py` (``micro_pytorch.py``) | 01:09.814 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_pytorch.py` (``micro_pytorch.py``) | 01:10.103 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``) | 00:51.727 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``) | 00:51.386 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``) | 00:09.034 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``) | 00:09.046 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``) | 00:03.787 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``) | 00:03.765 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tvmc.py` (``micro_tvmc.py``) | 00:00.000 | 0.0 MB |
+---------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt b/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt
index 3977266a7e..c5e4d76f50 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
=================
-**00:45.047** total execution time for **how_to_work_with_relay** files:
+**00:44.381** 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``) | 00:32.702 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_pipeline_executor.py` (``using_pipeline_executor.py``) | 00:32.572 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``) | 00:10.617 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``) | 00:10.261 | 0.0 MB |
+----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``) | 00:01.722 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``) | 00:01.542 | 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 23f36b0dc8..47ac0b8459 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 0x7fb3075403b0>
+ <function my_cuda_math_rule at 0x7f1830b663b0>
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 d9df7ec684..aba50f52a9 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
=================
-**00:06.262** total execution time for **how_to_work_with_schedules** files:
+**00:07.169** total execution time for **how_to_work_with_schedules** files:
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``) | 00:03.654 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``) | 00:04.652 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``) | 00:01.246 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``) | 00:01.154 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``) | 00:00.576 | 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_scan.py` (``scan.py``) | 00:00.559 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``) | 00:00.557 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_extern_op.py` (``extern_op.py``) | 00:00.119 | 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_schedule_primitives.py` (``schedule_primitives.py``) | 00:00.051 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tedd.py` (``tedd.py``) | 00:00.033 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tedd.py` (``tedd.py``) | 00:00.035 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_how_to_work_with_schedules_tuple_inputs.py` (``tuple_inputs.py``) | 00:00.025 | 0.0 MB |
+------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/work_with_schedules/tensorize.rst.txt b/docs/_sources/how_to/work_with_schedules/tensorize.rst.txt
index 86da29d8b2..d7e919ae22 100644
--- a/docs/_sources/how_to/work_with_schedules/tensorize.rst.txt
+++ b/docs/_sources/how_to/work_with_schedules/tensorize.rst.txt
@@ -340,7 +340,7 @@ The importing needs to happen before the tensorized GEMV being executed.
def main(A: T.Buffer((1024, 64), "float32"), B: T.Buffer((512, 64), "float32"), C: T.Buffer((1024, 512), "float32")):
T.func_attr({"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True})
i = T.var("int32")
- T.attr(T.iter_var(i, None, "DataPar", ""), "pragma_import_llvm", "; ModuleID = '/tmp/tmpro9643gx/input0.cc'\nsource_filename = \"/tmp/tmpro9643gx/input0.cc\"\ntarget datalayout = \"e-m:e-i64:64-f80:128-n8:16:32:64-S128\"\ntarget triple = \"x86_64-pc-linux-gnu\"\n\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local i32 @gemv_update(float*, float*, float*, i32, i32, i32) #0 {\n %7 = alloca float*, align 8\n %8 = alloca float*, align 8\n %9 = alloca float*, [...]
+ T.attr(T.iter_var(i, None, "DataPar", ""), "pragma_import_llvm", "; ModuleID = '/tmp/tmp0k_63x2_/input0.cc'\nsource_filename = \"/tmp/tmp0k_63x2_/input0.cc\"\ntarget datalayout = \"e-m:e-i64:64-f80:128-n8:16:32:64-S128\"\ntarget triple = \"x86_64-pc-linux-gnu\"\n\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local i32 @gemv_update(float*, float*, float*, i32, i32, i32) #0 {\n %7 = alloca float*, align 8\n %8 = alloca float*, align 8\n %9 = alloca float*, [...]
for i, j_outer in T.grid(1024, 32):
T.call_extern("int32", "gemv_update", T.tvm_access_ptr(T.type_annotation("float32"), C.data, i * 512 + j_outer * 16, 16, 2), T.tvm_access_ptr(T.type_annotation("float32"), A.data, i * 64, 64, 1), T.tvm_access_ptr(T.type_annotation("float32"), B.data, j_outer * 1024, 1024, 1), 16, 64, 64)
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 f731a317fb..d68c525596 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
=================
-**00:30.483** total execution time for **topic_vta_tutorials_autotvm** files:
+**00:30.464** total execution time for **topic_vta_tutorials_autotvm** files:
+---------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:30.476 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:30.457 | 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 a885c0cd42..beb474e65b 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 32.42s!
+ resnet18_v1 inference graph built in 32.44s!
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 e8eda5f256..36ad8c70e6 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 22.21s!
+ yolov3-tiny inference graph built in 22.16s!
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 6158a274ec..9579aa665e 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
=================
-**01:37.892** total execution time for **topic_vta_tutorials_frontend** files:
+**01:38.678** total execution time for **topic_vta_tutorials_frontend** files:
+------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:49.279 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``) | 00:49.478 | 0.0 MB |
+------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``) | 00:48.613 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:49.200 | 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 67fc38cfa4..c0438e8105 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
=================
-**00:03.145** total execution time for **topic_vta_tutorials_optimize** files:
+**00:03.188** total execution time for **topic_vta_tutorials_optimize** files:
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``) | 00:02.691 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``) | 00:02.727 | 0.0 MB |
+--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.453 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.461 | 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 5c9b49b689..8a01d08f3c 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
=================
-**00:00.816** total execution time for **topic_vta_tutorials** files:
+**00:00.824** total execution time for **topic_vta_tutorials** files:
+---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.432 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.436 | 0.0 MB |
+---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.384 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.388 | 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 8b870d5ef8..c9b2772362 100644
--- a/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
+++ b/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
@@ -318,7 +318,7 @@ We build the binary and check its correctness and performance.
.. code-block:: none
- Execution time of this operator: 95.253 ms
+ Execution time of this operator: 95.895 ms
@@ -436,7 +436,7 @@ operations.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 26.964 seconds)
+ **Total running time of the script:** ( 1 minutes 18.557 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 82a4316e8d..383e5fe37e 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: 4.15/4.15 result: MeasureResult(costs=(0.0646062716,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2819013595581055, timestamp=1674801589.425423) [('tile_y', [-1, 16]), ('tile_x', [-1, 16])],None,44
- No: 2 GFLOPS: 3.32/4.15 result: MeasureResult(costs=(0.080842935,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.536332130432129, timestamp=1674801591.742992) [('tile_y', [-1, 8]), ('tile_x', [-1, 8])],None,33
- No: 3 GFLOPS: 10.64/10.64 result: MeasureResult(costs=(0.0252188724,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6333749294281006, timestamp=1674801592.4050639) [('tile_y', [-1, 1]), ('tile_x', [-1, 512])],None,90
- No: 4 GFLOPS: 0.86/10.64 result: MeasureResult(costs=(0.310955842,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.192100524902344, timestamp=1674801598.3878272) [('tile_y', [-1, 512]), ('tile_x', [-1, 2])],None,19
- No: 5 GFLOPS: 14.64/14.64 result: MeasureResult(costs=(0.0183319464,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.8400697708129883, timestamp=1674801599.3476355) [('tile_y', [-1, 64]), ('tile_x', [-1, 64])],None,66
- No: 6 GFLOPS: 0.51/14.64 result: MeasureResult(costs=(0.52726684,), error_no=MeasureErrorNo.NO_ERROR, all_cost=8.658513307571411, timestamp=1674801608.0261073) [('tile_y', [-1, 256]), ('tile_x', [-1, 1])],None,8
- No: 7 GFLOPS: 1.67/14.64 result: MeasureResult(costs=(0.1606050886,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.800123453140259, timestamp=1674801611.6096852) [('tile_y', [-1, 8]), ('tile_x', [-1, 1])],None,3
- No: 8 GFLOPS: 13.44/14.64 result: MeasureResult(costs=(0.0199796056,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5835709571838379, timestamp=1674801612.1841033) [('tile_y', [-1, 16]), ('tile_x', [-1, 512])],None,94
- No: 9 GFLOPS: 0.90/14.64 result: MeasureResult(costs=(0.29902428719999996,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.989312171936035, timestamp=1674801617.2901492) [('tile_y', [-1, 64]), ('tile_x', [-1, 2])],None,16
- No: 10 GFLOPS: 9.96/14.64 result: MeasureResult(costs=(0.0269493716,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6376514434814453, timestamp=1674801617.9795842) [('tile_y', [-1, 1]), ('tile_x', [-1, 256])],None,80
+ No: 1 GFLOPS: 12.05/12.05 result: MeasureResult(costs=(0.0222695662,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6086859703063965, timestamp=1674811119.2983828) [('tile_y', [-1, 256]), ('tile_x', [-1, 256])],None,88
+ No: 2 GFLOPS: 3.51/12.05 result: MeasureResult(costs=(0.07643162220000001,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.471153974533081, timestamp=1674811121.539481) [('tile_y', [-1, 16]), ('tile_x', [-1, 8])],None,34
+ No: 3 GFLOPS: 8.64/12.05 result: MeasureResult(costs=(0.0310735592,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.9119014739990234, timestamp=1674811123.0475264) [('tile_y', [-1, 16]), ('tile_x', [-1, 64])],None,64
+ No: 4 GFLOPS: 12.78/12.78 result: MeasureResult(costs=(0.021010637399999997,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6143121719360352, timestamp=1674811123.6562173) [('tile_y', [-1, 32]), ('tile_x', [-1, 128])],None,75
+ No: 5 GFLOPS: 1.59/12.78 result: MeasureResult(costs=(0.1692736558,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.92525577545166, timestamp=1674811126.7084243) [('tile_y', [-1, 1]), ('tile_x', [-1, 4])],None,20
+ No: 6 GFLOPS: 10.25/12.78 result: MeasureResult(costs=(0.0261835056,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6508300304412842, timestamp=1674811127.3828309) [('tile_y', [-1, 512]), ('tile_x', [-1, 256])],None,89
+ No: 7 GFLOPS: 1.45/12.78 result: MeasureResult(costs=(0.1850034158,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.1807830333709717, timestamp=1674811131.3510664) [('tile_y', [-1, 4]), ('tile_x', [-1, 1])],None,2
+ No: 8 GFLOPS: 2.05/12.78 result: MeasureResult(costs=(0.13108851959999998,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.325491189956665, timestamp=1674811133.699819) [('tile_y', [-1, 256]), ('tile_x', [-1, 4])],None,28
+ No: 9 GFLOPS: 12.69/12.78 result: MeasureResult(costs=(0.0211580088,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5448176860809326, timestamp=1674811134.3613458) [('tile_y', [-1, 32]), ('tile_x', [-1, 512])],None,95
+ No: 10 GFLOPS: 1.86/12.78 result: MeasureResult(costs=(0.1445250338,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.512148380279541, timestamp=1674811136.9226403) [('tile_y', [-1, 2]), ('tile_x', [-1, 2])],None,11
diff --git a/docs/_sources/tutorial/autotvm_relay_x86.rst.txt b/docs/_sources/tutorial/autotvm_relay_x86.rst.txt
index 4ef718f65a..46f4e12d3e 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': 514.6029234700006, 'median': 515.1770631500028, 'std': 1.6195756103389116}
+ {'mean': 513.7597591200017, 'median': 513.8930147499991, 'std': 1.657267041893733}
@@ -545,30 +545,30 @@ 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: 9.47/ 9.47 GFLOPS | Progress: (4/20) | 8.71 s
[Task 1/25] Current/Best: 12.87/ 22.12 GFLOPS | Progress: (8/20) | 11.93 s
[Task 1/25] Current/Best: 6.39/ 22.12 GFLOPS | Progress: (12/20) | 14.38 s
[Task 1/25] Current/Best: 13.19/ 23.54 GFLOPS | Progress: (16/20) | 17.33 s
[Task 1/25] Current/Best: 17.76/ 23.54 GFLOPS | Progress: (20/20) | 20.02 s Done.
-
[Task 2/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 2/25] Current/Best: 21.62/ 21.62 GFLOPS | Progress: (4/20) | 3.53 s
[Task 2/25] Current/Best: 6.22/ 21.62 GFLOPS | Progress: (8/20) | 4.80 s
[Task 2/25] Current/Best: 16.06/ 21.62 GFLOPS | Progress: (12/20) | 7.87 s
[Task 2/25] Current/Best: 13.57/ 21.62 GFLOPS | Progress: (16/20) | 9.23 s
[Task 2/25] Current/Best: 14.14/ 21.62 GFLOPS | Progress: (20/20) | 10.82 s Done.
-
[Task 3/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 3/25] Current/Best: 18.26/ 18.33 GFLOPS | Progress: (4/20) | 4.59 s
[Task 3/25] Current/Best: 20.19/ 21.03 GFLOPS | Progress: (8/20) | 7.63 s
[Task 3/25] Current/Best: 17.57/ 21.03 GFLOPS | Progress: (12/20) | 9.81 s
[Task 3/25] Current/Best: 20.10/ 21.03 GFLOPS | Progress: (16/20) | 11.95 s
[Task 3/25] Current/Best: 9.06/ 21.03 GFLOPS | Progress: (20/20) | 14.14 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/ 16.00 GFLOPS | Progress: (4/20) | 6.65 s
[Task 4/25] Current/Best: 6.58/ 18.65 GFLOPS | Progress: (8/20) | 8.48 s
[Task 4/25] Current/Best: 11.70/ 18.65 GFLOPS | Progress: (12/20) | 10.53 s
[Task 4/25] Current/Best: 19.67/ 19.67 GFLOPS | Progress: (16/20) | 14.90 s
[Task 4/25] Current/Best: 10.30/ 19.67 GFLOPS | Progress: (20/20) | 18.28 s Done.
-
[Task 5/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 5/25] Current/Best: 4.89/ 10.76 GFLOPS | Progress: (4/20) | 4.77 s
[Task 5/25] Current/Best: 15.42/ 15.42 GFLOPS | Progress: (8/20) | 6.90 s
[Task 5/25] Current/Best: 17.54/ 18.47 GFLOPS | Progress: (12/20) | 8.79 s
[Task 5/25] Current/Best: 9.04/ 18.47 GFLOPS | Progress: (16/20) | 11.18 s
[Task 5/25] Current/Best: 4.44/ 18.47 GFLOPS | Progress: (20/20) | 15.08 s Done.
-
[Task 6/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 6/25] Current/Best: 11.03/ 11.03 GFLOPS | Progress: (4/20) | 5.14 s
[Task 6/25] Current/Best: 14.02/ 18.26 GFLOPS | Progress: (8/20) | 7.53 s
[Task 6/25] Current/Best: 17.79/ 18.26 GFLOPS | Progress: (12/20) | 9.90 s
[Task 6/25] Current/Best: 5.29/ 21.69 GFLOPS | Progress: (16/20) | 13.02 s
[Task 6/25] Current/Best: 12.38/ 21.69 GFLOPS | Progress: (20/20) | 16.20 s Done.
-
[Task 7/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 7/25] Current/Best: 4.63/ 18.55 GFLOPS | Progress: (4/20) | 4.41 s
[Task 7/25] Current/Best: 7.29/ 18.55 GFLOPS | Progress: (8/20) | 7.12 s
[Task 7/25] Current/Best: 5.20/ 18.55 GFLOPS | Progress: (12/20) | 9.94 s
[Task 7/25] Current/Best: 6.03/ 18.67 GFLOPS | Progress: (16/20) | 12.45 s
[Task 7/25] Current/Best: 5.05/ 18.67 GFLOPS | Progress: (20/20) | 15.55 s Done.
-
[Task 8/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 8/25] Current/Best: 7.28/ 15.78 GFLOPS | Progress: (4/20) | 13.86 s
[Task 8/25] Current/Best: 13.39/ 15.78 GFLOPS | Progress: (8/20) | 18.04 s
[Task 8/25] Current/Best: 4.88/ 21.13 GFLOPS | Progress: (12/20) | 23.41 s
[Task 8/25] Current/Best: 17.34/ 21.13 GFLOPS | Progress: (16/20) | 28.29 s
[Task 8/25] Current/Best: 3.07/ 21.13 GFLOPS | Progress: (20/20) | 35.70 s
[Task 9/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 9/25] Current/Best: 20.83/ 20.83 GFLOPS | Progress: (4/20) | 3.65 s
[Task 9/25] Current/Best: 15.94/ 20.83 GFLOPS | Progress: (8/20) | 14.37 s
[Task 9/25] Current/Best: 9.94/ 20.83 GFLOPS | Progress: (12/20) | 17.27 s
[Task 9/25] Current/Best: 8.02/ 20.83 GFLOPS | Progress: (16/20) | 19.03 s
[Task 9/25] Current/Best: 10.14/ 21.01 GFLOPS | Progress: (20
/20) | 22.43 s
[Task 10/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 10/25] Current/Best: 13.82/ 14.09 GFLOPS | Progress: (4/20) | 3.83 s
[Task 10/25] Current/Best: 11.05/ 18.14 GFLOPS | Progress: (8/20) | 6.17 s
[Task 10/25] Current/Best: 15.25/ 18.14 GFLOPS | Progress: (12/20) | 8.63 s
[Task 10/25] Current/Best: 14.14/ 18.14 GFLOPS | Progress: (16/20) | 10.97 s
[Task 10/25] Current/Best: 10.04/ 18.14 GFLOPS | Progress: (20/20) | 13.92 s Done.
-
[Task 11/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 11/25] Current/Best: 12.36/ 19.74 GFLOPS | Progress: (4/20) | 5.42 s
[Task 11/25] Current/Best: 7.01/ 19.74 GFLOPS | Progress: (8/20) | 8.20 s
[Task 11/25] Current/Best: 11.38/ 19.74 GFLOPS | Progress: (12/20) | 10.89 s
[Task 11/25] Current/Best: 15.68/ 19.74 GFLOPS | Progress: (16/20) | 13.15 s
[Task 11/25] Current/Best: 9.64/ 19.74 GFLOPS | Progress: (20/20) | 16.46 s Done.
-
[Task 12/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 12/25] Current/Best: 6.76/ 18.48 GFLOPS | Progress: (4/20) | 4.09 s Done.
+
[Task 1/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 1/25] Current/Best: 9.03/ 15.11 GFLOPS | Progress: (4/20) | 9.17 s
[Task 1/25] Current/Best: 19.09/ 19.09 GFLOPS | Progress: (8/20) | 11.97 s
[Task 1/25] Current/Best: 7.67/ 19.09 GFLOPS | Progress: (12/20) | 15.82 s
[Task 1/25] Current/Best: 12.46/ 19.09 GFLOPS | Progress: (16/20) | 19.35 s
[Task 1/25] Current/Best: 4.25/ 22.04 GFLOPS | Progress: (20/20) | 23.32 s Done.
+
[Task 2/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 2/25] Current/Best: 17.33/ 17.33 GFLOPS | Progress: (4/20) | 3.27 s
[Task 2/25] Current/Best: 14.91/ 17.33 GFLOPS | Progress: (8/20) | 5.28 s
[Task 2/25] Current/Best: 14.91/ 22.47 GFLOPS | Progress: (12/20) | 7.01 s
[Task 2/25] Current/Best: 6.92/ 22.47 GFLOPS | Progress: (16/20) | 8.67 s
[Task 2/25] Current/Best: 17.00/ 22.47 GFLOPS | Progress: (20/20) | 10.30 s Done.
+
[Task 3/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 3/25] Current/Best: 17.60/ 17.60 GFLOPS | Progress: (4/20) | 4.15 s
[Task 3/25] Current/Best: 6.52/ 17.60 GFLOPS | Progress: (8/20) | 7.11 s
[Task 3/25] Current/Best: 3.17/ 19.52 GFLOPS | Progress: (12/20) | 11.14 s
[Task 3/25] Current/Best: 6.20/ 21.76 GFLOPS | Progress: (16/20) | 13.77 s
[Task 3/25] Current/Best: 22.06/ 22.06 GFLOPS | Progress: (20/20) | 16.10 s Done.
+
[Task 4/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 4/25] Current/Best: 4.72/ 12.69 GFLOPS | Progress: (4/20) | 9.41 s
[Task 4/25] Current/Best: 10.90/ 18.37 GFLOPS | Progress: (8/20) | 12.35 s
[Task 4/25] Current/Best: 12.61/ 18.37 GFLOPS | Progress: (12/20) | 18.10 s
[Task 4/25] Current/Best: 14.01/ 18.37 GFLOPS | Progress: (16/20) | 20.15 s
[Task 4/25] Current/Best: 18.69/ 18.69 GFLOPS | Progress: (20/20) | 22.20 s Done.
+
[Task 5/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 5/25] Current/Best: 17.77/ 18.35 GFLOPS | Progress: (4/20) | 4.28 s
[Task 5/25] Current/Best: 17.51/ 20.44 GFLOPS | Progress: (8/20) | 6.13 s
[Task 5/25] Current/Best: 11.63/ 20.44 GFLOPS | Progress: (12/20) | 8.26 s
[Task 5/25] Current/Best: 20.07/ 20.44 GFLOPS | Progress: (16/20) | 10.60 s
[Task 5/25] Current/Best: 9.18/ 20.44 GFLOPS | Progress: (20/20) | 13.11 s Done.
+
[Task 6/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 6/25] Current/Best: 16.26/ 16.26 GFLOPS | Progress: (4/20) | 4.51 s
[Task 6/25] Current/Best: 9.37/ 16.26 GFLOPS | Progress: (8/20) | 7.26 s
[Task 6/25] Current/Best: 10.18/ 17.85 GFLOPS | Progress: (12/20) | 10.54 s
[Task 6/25] Current/Best: 10.95/ 19.47 GFLOPS | Progress: (16/20) | 12.72 s
[Task 6/25] Current/Best: 19.12/ 19.47 GFLOPS | Progress: (20/20) | 16.03 s Done.
+
[Task 7/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 7/25] Current/Best: 19.12/ 19.27 GFLOPS | Progress: (4/20) | 3.95 s
[Task 7/25] Current/Best: 15.51/ 19.27 GFLOPS | Progress: (8/20) | 6.34 s
[Task 7/25] Current/Best: 6.22/ 19.27 GFLOPS | Progress: (12/20) | 8.99 s
[Task 7/25] Current/Best: 15.00/ 19.27 GFLOPS | Progress: (16/20) | 12.10 s
[Task 7/25] Current/Best: 7.90/ 19.27 GFLOPS | Progress: (20/20) | 14.66 s Done.
+
[Task 8/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 8/25] Current/Best: 3.91/ 18.64 GFLOPS | Progress: (4/20) | 7.66 s
[Task 8/25] Current/Best: 4.27/ 18.64 GFLOPS | Progress: (8/20) | 10.15 s
[Task 8/25] Current/Best: 2.75/ 18.64 GFLOPS | Progress: (12/20) | 16.99 s
[Task 8/25] Current/Best: 22.34/ 22.34 GFLOPS | Progress: (16/20) | 20.06 s
[Task 8/25] Current/Best: 10.92/ 22.34 GFLOPS | Progress: (20/20) | 22.97 s Done.
+
[Task 9/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 9/25] Current/Best: 8.72/ 17.81 GFLOPS | Progress: (4/20) | 3.75 s
[Task 9/25] Current/Best: 6.05/ 17.81 GFLOPS | Progress: (8/20) | 14.96 s
[Task 9/25] Current/Best: 17.83/ 19.82 GFLOPS | Progress: (12/20) | 25.94 s
[Task 9/25] Current/Best: 11.50/ 20.51 GFLOPS | Progress: (16/20) | 30.07 s
[Task 9/25] Current/Best: 13.90/ 20.51 GFLOPS | Progress: (20/20) | 34.30 s
[Task 10/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 10/25] Current/Best: 5.22/ 15.25 GFLOPS | Progress: (4/20) | 3.86 s
[Task 10/25] Current/Best: 10.05/ 20.70 GFLOPS | Progress: (8/20) | 5.41 s
[Task 10/25] Current/Best: 4.87/ 20.70 GFLOPS | Progress: (12/20) | 7.90 s
[Task 10/25] Current/Best: 15.83/ 20.70 GFLOPS | Progress: (16/20) | 9.95 s
[Task 10/25] Current/Best: 6.86/ 20.70 GFLOPS | Progress: (20/20)
| 11.96 s Done.
+
[Task 11/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 11/25] Current/Best: 13.27/ 19.62 GFLOPS | Progress: (4/20) | 3.79 s
[Task 11/25] Current/Best: 8.12/ 19.62 GFLOPS | Progress: (8/20) | 6.65 s
[Task 11/25] Current/Best: 7.78/ 22.33 GFLOPS | Progress: (12/20) | 9.10 s
[Task 11/25] Current/Best: 8.76/ 22.33 GFLOPS | Progress: (16/20) | 11.80 s
[Task 11/25] Current/Best: 17.12/ 23.36 GFLOPS | Progress: (20/20) | 13.87 s Done.
+
[Task 12/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 12/25] Current/Best: 14.46/ 15.70 GFLOPS | Progress: (4/20) | 8.24 s
[Task 12/25] Current/Best: 10.63/ 21.69 GFLOPS | Progress: (8/20) | 10.96 s
[Task 12/25] Current/Best: 13.81/ 21.69 GFLOPS | Progress: (12/20) | 13.31 s
[Task 12/25] Current/Best: 14.91/ 21.69 GFLOPS | Progress: (16/20) | 17.36 s
[Task 12/25] Current/Best: 19.02/ 21.69 GFLOPS | Progress: (20/20) | 21.57 s Done.
+
[Task 13/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 13/25] Current/Best: 16.36/ 16.36 GFLOPS | Progress: (4/20) | 6.69 s
[Task 13/25] Current/Best: 17.88/ 20.10 GFLOPS | Progress: (8/20) | 9.13 s
[Task 13/25] Current/Best: 7.62/ 20.10 GFLOPS | Progress: (12/20) | 11.60 s
[Task 13/25] Current/Best: 8.83/ 20.10 GFLOPS | Progress: (16/20) | 15.02 s
[Task 13/25] Current/Best: 10.69/ 20.10 GFLOPS | Progress: (20/20) | 17.86 s Done.
+
[Task 14/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 14/25] Current/Best: 9.24/ 12.43 GFLOPS | Progress: (4/20) | 6.75 s
[Task 14/25] Current/Best: 11.02/ 12.43 GFLOPS | Progress: (8/20) | 8.95 s
[Task 14/25] Current/Best: 2.76/ 14.95 GFLOPS | Progress: (12/20) | 13.23 s Done.
+
[Task 14/25] Current/Best: 9.52/ 14.95 GFLOPS | Progress: (16/20) | 16.59 s
[Task 14/25] Current/Best: 6.73/ 19.94 GFLOPS | Progress: (20/20) | 18.96 s Done.
+
[Task 15/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 15/25] Current/Best: 3.14/ 17.91 GFLOPS | Progress: (4/20) | 4.95 s
[Task 15/25] Current/Best: 8.19/ 17.91 GFLOPS | Progress: (8/20) | 9.37 s
[Task 15/25] Current/Best: 19.86/ 19.86 GFLOPS | Progress: (12/20) | 11.57 s
[Task 15/25] Current/Best: 17.92/ 19.86 GFLOPS | Progress: (16/20) | 17.42 s
[Task 15/25] Current/Best: 14.24/ 22.32 GFLOPS | Progress: (20/20) | 18.92 s
[Task 16/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 16/25] Current/Best: 14.04/ 19.39 GFLOPS | Progress: (4/20) | 4.22 s
[Task 16/25] Current/Best: 10.93/ 20.56 GFLOPS | Progress: (8/20) | 6.73 s
[Task 16/25] Current/Best: 14.38/ 20.56 GFLOPS | Progress: (12/20) | 8.33 s
[Task 16/25] Current/Best: 9.24/ 20.56 GFLOPS | Progress: (16/20) | 10.87 s
[Task 16/25] Current/Best: 6.04/ 20.56 GFLOPS | Progress: (20/20)
| 13.97 s Done.
+
[Task 17/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 17/25] Current/Best: 3.08/ 15.95 GFLOPS | Progress: (4/20) | 5.13 s
[Task 17/25] Current/Best: 8.54/ 23.67 GFLOPS | Progress: (8/20) | 7.38 s
[Task 17/25] Current/Best: 9.71/ 23.67 GFLOPS | Progress: (12/20) | 9.85 s
[Task 17/25] Current/Best: 17.97/ 23.67 GFLOPS | Progress: (16/20) | 13.28 s
[Task 17/25] Current/Best: 14.11/ 23.67 GFLOPS | Progress: (20/20) | 15.61 s Done.
+
[Task 18/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 18/25] Current/Best: 6.41/ 14.58 GFLOPS | Progress: (4/20) | 6.68 s
[Task 18/25] Current/Best: 17.30/ 17.80 GFLOPS | Progress: (8/20) | 14.18 s
[Task 18/25] Current/Best: 15.63/ 17.80 GFLOPS | Progress: (12/20) | 20.63 s
[Task 18/25] Current/Best: 18.86/ 18.86 GFLOPS | Progress: (16/20) | 25.49 s
[Task 18/25] Current/Best: 18.00/ 18.86 GFLOPS | Progress: (20/20) | 27.48 s Done.
+
[Task 19/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 19/25] Current/Best: 2.69/ 9.47 GFLOPS | Progress: (4/20) | 5.45 s
[Task 19/25] Current/Best: 20.56/ 20.56 GFLOPS | Progress: (8/20) | 12.76 s
[Task 19/25] Current/Best: 10.64/ 20.56 GFLOPS | Progress: (12/20) | 15.41 s
[Task 19/25] Current/Best: 23.34/ 23.34 GFLOPS | Progress: (16/20) | 19.04 s
[Task 19/25] Current/Best: 7.86/ 23.34 GFLOPS | Progress: (20/20) | 24.05 s Done.
+
[Task 20/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 20/25] Current/Best: 10.02/ 18.68 GFLOPS | Progress: (4/20) | 3.83 s
[Task 20/25] Current/Best: 19.69/ 19.69 GFLOPS | Progress: (8/20) | 6.04 s
[Task 20/25] Current/Best: 19.14/ 19.69 GFLOPS | Progress: (12/20) | 10.06 s
[Task 20/25] Current/Best: 16.05/ 20.28 GFLOPS | Progress: (16/20) | 12.45 s
[Task 20/25] Current/Best: 4.17/ 20.28 GFLOPS | Progress: (20/20) | 16.37 s
[Task 21/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s Done.
Done.
-
[Task 12/25] Current/Best: 12.97/ 18.48 GFLOPS | Progress: (8/20) | 6.93 s
[Task 12/25] Current/Best: 12.57/ 18.48 GFLOPS | Progress: (12/20) | 11.16 s
[Task 12/25] Current/Best: 9.03/ 18.48 GFLOPS | Progress: (16/20) | 15.62 s
[Task 12/25] Current/Best: 11.79/ 18.48 GFLOPS | Progress: (20/20) | 18.94 s Done.
-
[Task 13/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 13/25] Current/Best: 15.09/ 15.09 GFLOPS | Progress: (4/20) | 4.68 s
[Task 13/25] Current/Best: 3.11/ 21.02 GFLOPS | Progress: (8/20) | 8.20 s
[Task 13/25] Current/Best: 21.98/ 21.98 GFLOPS | Progress: (12/20) | 12.61 s
[Task 13/25] Current/Best: 13.40/ 21.98 GFLOPS | Progress: (16/20) | 15.40 s
[Task 13/25] Current/Best: 3.10/ 21.98 GFLOPS | Progress: (20/20) | 18.85 s Done.
-
[Task 14/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 14/25] Current/Best: 10.42/ 21.12 GFLOPS | Progress: (4/20) | 6.38 s
[Task 14/25] Current/Best: 1.60/ 21.12 GFLOPS | Progress: (8/20) | 9.59 s
[Task 14/25] Current/Best: 3.93/ 22.33 GFLOPS | Progress: (12/20) | 12.29 s
[Task 14/25] Current/Best: 6.97/ 22.33 GFLOPS | Progress: (16/20) | 15.58 s
[Task 14/25] Current/Best: 11.87/ 22.33 GFLOPS | Progress: (20/20) | 20.47 s
[Task 15/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 15/25] Current/Best: 18.53/ 18.53 GFLOPS | Progress: (4/20) | 7.66 s
[Task 15/25] Current/Best: 16.32/ 18.53 GFLOPS | Progress: (8/20) | 9.98 s
[Task 15/25] Current/Best: 11.76/ 18.53 GFLOPS | Progress: (12/20) | 12.19 s Done.
-
[Task 15/25] Current/Best: 18.69/ 18.69 GFLOPS | Progress: (16/20) | 14.25 s
[Task 15/25] Current/Best: 14.78/ 18.69 GFLOPS | Progress: (20/20) | 17.28 s
[Task 16/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 16/25] Current/Best: 16.43/ 18.42 GFLOPS | Progress: (4/20) | 4.25 s
[Task 16/25] Current/Best: 13.88/ 18.42 GFLOPS | Progress: (8/20) | 6.43 s
[Task 16/25] Current/Best: 20.85/ 20.85 GFLOPS | Progress: (12/20) | 9.61 s
[Task 16/25] Current/Best: 10.90/ 20.85 GFLOPS | Progress: (16/20) | 12.30 s
[Task 16/25] Current/Best: 9.58/ 20.85 GFLOPS | Progress: (20/20) | 14.32 s Done.
-
[Task 17/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 17/25] Current/Best: 12.81/ 23.57 GFLOPS | Progress: (4/20) | 3.98 s
[Task 17/25] Current/Best: 19.51/ 23.57 GFLOPS | Progress: (8/20) | 7.47 s
[Task 17/25] Current/Best: 11.28/ 23.57 GFLOPS | Progress: (12/20) | 11.33 s
[Task 17/25] Current/Best: 18.24/ 23.57 GFLOPS | Progress: (16/20) | 13.60 s
[Task 17/25] Current/Best: 14.11/ 23.57 GFLOPS | Progress: (20/20) | 16.58 s Done.
-
[Task 18/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 18/25] Current/Best: 18.96/ 18.96 GFLOPS | Progress: (4/20) | 5.87 s
[Task 18/25] Current/Best: 12.04/ 18.96 GFLOPS | Progress: (8/20) | 8.63 s
[Task 18/25] Current/Best: 6.34/ 19.05 GFLOPS | Progress: (12/20) | 10.90 s
[Task 18/25] Current/Best: 21.23/ 21.23 GFLOPS | Progress: (16/20) | 17.47 s
[Task 18/25] Current/Best: 13.28/ 21.23 GFLOPS | Progress: (20/20) | 19.77 s Done.
-
[Task 19/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 19/25] Current/Best: 10.24/ 12.04 GFLOPS | Progress: (4/20) | 6.07 s
[Task 19/25] Current/Best: 18.29/ 18.29 GFLOPS | Progress: (8/20) | 9.51 s
[Task 19/25] Current/Best: 12.01/ 18.43 GFLOPS | Progress: (12/20) | 11.65 s
[Task 19/25] Current/Best: 18.19/ 19.07 GFLOPS | Progress: (16/20) | 15.42 s
[Task 19/25] Current/Best: 18.05/ 21.79 GFLOPS | Progress: (20/20) | 19.15 s Done.
-
[Task 20/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 20/25] Current/Best: 16.20/ 19.51 GFLOPS | Progress: (4/20) | 4.11 s
[Task 20/25] Current/Best: 12.14/ 20.08 GFLOPS | Progress: (8/20) | 8.48 s
[Task 20/25] Current/Best: 4.06/ 20.08 GFLOPS | Progress: (12/20) | 10.97 s
[Task 20/25] Current/Best: 5.47/ 20.08 GFLOPS | Progress: (16/20) | 14.30 s
[Task 20/25] Current/Best: 10.39/ 20.08 GFLOPS | Progress: (20/20) | 17.48 s
[Task 21/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 21/25] Current/Best: 15.27/ 19.07 GFLOPS | Progress: (4/20) | 3.60 s
[Task 21/25] Current/Best: 16.82/ 19.07 GFLOPS | Progress: (8/20) | 4.92 s
[Task 21/25] Current/Best: 9.18/ 20.99 GFLOPS | Progress: (12/20) | 8.66 s
[Task 21/25] Current/Best: 14.25/ 20.99 GFLOPS | Progress: (16/20) | 10.98 s
[Task 21/25] Current/Best: 10.85/ 20.99 GFLOPS | Progress: (20/20)
| 13.32 s
[Task 22/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s Done.
- Done.
-
[Task 22/25] Current/Best: 18.30/ 18.30 GFLOPS | Progress: (4/20) | 3.69 s
[Task 22/25] Current/Best: 8.91/ 18.30 GFLOPS | Progress: (8/20) | 6.12 s
[Task 22/25] Current/Best: 7.99/ 18.30 GFLOPS | Progress: (12/20) | 8.09 s
[Task 22/25] Current/Best: 12.03/ 18.30 GFLOPS | Progress: (16/20) | 11.39 s
[Task 22/25] Current/Best: 14.29/ 18.30 GFLOPS | Progress: (20/20) | 14.35 s Done.
-
[Task 23/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 23/25] Current/Best: 14.01/ 17.88 GFLOPS | Progress: (4/20) | 8.93 s
[Task 23/25] Current/Best: 9.69/ 17.88 GFLOPS | Progress: (8/20) | 12.02 s
[Task 23/25] Current/Best: 23.63/ 23.63 GFLOPS | Progress: (12/20) | 14.18 s
[Task 23/25] Current/Best: 22.91/ 23.63 GFLOPS | Progress: (16/20) | 16.35 s
[Task 23/25] Current/Best: 19.92/ 23.63 GFLOPS | Progress: (20/20) | 21.68 s Done.
-
[Task 24/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 24/25] Current/Best: 5.25/ 9.19 GFLOPS | Progress: (4/20) | 4.46 s
[Task 24/25] Current/Best: 2.81/ 10.20 GFLOPS | Progress: (8/20) | 15.39 s
[Task 24/25] Current/Best: 2.18/ 10.20 GFLOPS | Progress: (12/20) | 24.52 s
[Task 24/25] Current/Best: 2.27/ 10.20 GFLOPS | Progress: (16/20) | 35.18 s
[Task 24/25] Current/Best: 5.54/ 10.20 GFLOPS | Progress: (20/20) | 46.11 s
[Task 25/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 25/25] Current/Best: 1.52/ 8.88 GFLOPS | Progress: (4/20) | 13.68 s Done.
-
[Task 25/25] Current/Best: 1.52/ 8.88 GFLOPS | Progress: (8/20) | 17.26 s
[Task 25/25] Current/Best: 5.77/ 8.88 GFLOPS | Progress: (12/20) | 27.93 s
[Task 25/25] Current/Best: 1.52/ 8.88 GFLOPS | Progress: (16/20) | 31.27 s
[Task 25/25] Current/Best: 1.55/ 8.88 GFLOPS | Progress: (20/20) | 41.98 s
+
[Task 21/25] Current/Best: 13.41/ 16.23 GFLOPS | Progress: (4/20) | 3.98 s
[Task 21/25] Current/Best: 6.69/ 16.23 GFLOPS | Progress: (8/20) | 9.06 s
[Task 21/25] Current/Best: 5.04/ 19.96 GFLOPS | Progress: (12/20) | 10.74 s
[Task 21/25] Current/Best: 17.80/ 19.96 GFLOPS | Progress: (16/20) | 12.02 s
[Task 21/25] Current/Best: 12.10/ 19.96 GFLOPS | Progress: (20/20) | 15.26 s
[Task 22/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 22/25] Current/Best: 14.88/ 18.60 GFLOPS | Progress: (4/20) | 4.11 s
[Task 22/25] Current/Best: 5.30/ 18.60 GFLOPS | Progress: (8/20) | 6.75 s
[Task 22/25] Current/Best: 9.81/ 18.60 GFLOPS | Progress: (12/20) | 9.64 s
[Task 22/25] Current/Best: 8.68/ 19.80 GFLOPS | Progress: (16/20) | 14.74 s
[Task 22/25] Current/Best: 15.57/ 19.80 GFLOPS | Progress: (20/20) | 16.78 s Done.
+
[Task 23/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 23/25] Current/Best: 13.04/ 23.31 GFLOPS | Progress: (4/20) | 3.91 s
[Task 23/25] Current/Best: 17.82/ 23.31 GFLOPS | Progress: (8/20) | 6.55 s
[Task 23/25] Current/Best: 8.42/ 23.31 GFLOPS | Progress: (12/20) | 9.74 s
[Task 23/25] Current/Best: 17.81/ 23.31 GFLOPS | Progress: (16/20) | 12.21 s
[Task 23/25] Current/Best: 11.95/ 23.31 GFLOPS | Progress: (20/20) | 14.68 s Done.
+
[Task 24/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s
[Task 24/25] Current/Best: 5.78/ 6.45 GFLOPS | Progress: (4/20) | 12.74 s
[Task 24/25] Current/Best: 2.85/ 6.45 GFLOPS | Progress: (8/20) | 23.70 s
[Task 24/25] Current/Best: 4.27/ 6.45 GFLOPS | Progress: (12/20) | 34.64 s
[Task 24/25] Current/Best: 1.44/ 7.08 GFLOPS | Progress: (16/20) | 37.31 s
[Task 24/25] Current/Best: 3.49/ 7.08 GFLOPS | Progress: (20/20) | 47.97 s
[Task 25/25] Current/Best: 0.00/ 0.00 GFLOPS | Progress: (0/20) | 0.00 s Done.
+
[Task 25/25] Current/Best: 6.38/ 8.59 GFLOPS | Progress: (4/20) | 12.78 s
[Task 25/25] Current/Best: 1.55/ 8.59 GFLOPS | Progress: (8/20) | 24.27 s
[Task 25/25] Current/Best: 3.54/ 9.16 GFLOPS | Progress: (12/20) | 26.91 s
[Task 25/25] Current/Best: 3.02/ 9.16 GFLOPS | Progress: (16/20) | 37.83 s
[Task 25/25] Current/Best: 4.21/ 9.16 GFLOPS | Progress: (20/20) | 39.92 s
@@ -664,8 +664,8 @@ Verify that the optimized model runs and produces the same results:
.. code-block:: none
- class='n02123045 tabby, tabby cat' with probability=0.621103
- class='n02123159 tiger cat' with probability=0.356379
+ class='n02123045 tabby, tabby cat' with probability=0.621104
+ class='n02123159 tiger cat' with probability=0.356378
class='n02124075 Egyptian cat' with probability=0.019712
class='n02129604 tiger, Panthera tigris' with probability=0.001215
class='n04040759 radiator' with probability=0.000262
@@ -722,8 +722,8 @@ improvement in comparing the optimized model to the unoptimized model.
.. code-block:: none
- optimized: {'mean': 423.5321302300008, 'median': 422.38251159999436, 'std': 3.261529146558627}
- unoptimized: {'mean': 514.6029234700006, 'median': 515.1770631500028, 'std': 1.6195756103389116}
+ optimized: {'mean': 411.2786791999997, 'median': 410.888242850001, 'std': 0.9985430156413876}
+ unoptimized: {'mean': 513.7597591200017, 'median': 513.8930147499991, 'std': 1.657267041893733}
@@ -746,7 +746,7 @@ profiling/benchmarking.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 12 minutes 6.103 seconds)
+ **Total running time of the script:** ( 12 minutes 13.227 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 f02bc2b63e..246fdf2841 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
- 1.4e-07 secs/op
+ 1.309e-07 secs/op
diff --git a/docs/_sources/tutorial/intro_topi.rst.txt b/docs/_sources/tutorial/intro_topi.rst.txt
index 9c4f041bc4..5343b3a555 100644
--- a/docs/_sources/tutorial/intro_topi.rst.txt
+++ b/docs/_sources/tutorial/intro_topi.rst.txt
@@ -277,7 +277,7 @@ As you can see, scheduled stages of computation have been accumulated and we can
.. code-block:: none
- [stage(a, placeholder(a, 0xd680740)), stage(b, placeholder(b, 0x88632f0)), 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.R [...]
+ [stage(a, placeholder(a, 0x198f3380)), stage(b, placeholder(b, 0x52c9050)), 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 c3ba992fec..2734a1c737 100644
--- a/docs/_sources/tutorial/sg_execution_times.rst.txt
+++ b/docs/_sources/tutorial/sg_execution_times.rst.txt
@@ -5,31 +5,31 @@
Computation times
=================
-**15:46.329** total execution time for **tutorial** files:
+**15:33.177** total execution time for **tutorial** files:
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``) | 12:06.103 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``) | 12:13.227 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 01:26.964 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 01:18.557 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``) | 01:01.053 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``) | 01:00.965 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``) | 00:35.691 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``) | 00:35.453 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``) | 00:34.308 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``) | 00:22.596 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``) | 00:01.230 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``) | 00:01.382 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``) | 00:00.824 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``) | 00:00.826 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.157 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.171 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_tutorial_uma.py` (``uma.py``) | 00:00.000 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tvmc_command_line_driver.py` (``tvmc_command_line_driver.py``) | 00:00.000 | 0.0 MB |
-+------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_tutorial_introduction.py` (``introduction.py``) | 00:00.000 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
+| :ref:`sphx_glr_tutorial_tvmc_command_line_driver.py` (``tvmc_command_line_driver.py``) | 00:00.000 | 0.0 MB |
++------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_tutorial_tvmc_python.py` (``tvmc_python.py``) | 00:00.000 | 0.0 MB |
+------------------------------------------------------------------------------------------+-----------+--------+
| :ref:`sphx_glr_tutorial_install.py` (``install.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 0b88783826..c0c50c6e36 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: 0.000008
+ Numpy running time: 0.000007
naive: 0.000007
@@ -447,7 +447,7 @@ factor to be the number of threads on your CPU.
.. code-block:: none
- vector: 0.000025
+ vector: 0.000030
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -504,10 +504,10 @@ We can now compare the different schedules
.. code-block:: none
Operator Timing Performance
- numpy 7.67368999959217e-06 1.0
- naive 6.672500000000001e-06 0.869529522349042
- parallel 6.9357e-06 0.9038285362542151
- vector 2.45506e-05 3.199321317554498
+ numpy 7.204539999747795e-06 1.0
+ naive 6.655199999999999e-06 0.9237508571307779
+ parallel 6.958100000000001e-06 0.9657937911710642
+ vector 3.03507e-05 4.212718647000707
@@ -928,7 +928,7 @@ matrix multiplication.
.. code-block:: none
- Numpy running time: 0.018616
+ Numpy running time: 0.017946
@@ -986,7 +986,7 @@ optimizations.
.. code-block:: none
- none: 3.410544
+ none: 3.413441
@@ -1086,7 +1086,7 @@ schedule.
.. code-block:: none
- blocking: 0.307069
+ blocking: 0.297418
@@ -1170,7 +1170,7 @@ already cache friendly from our previous optimizations.
.. code-block:: none
- vectorization: 0.336873
+ vectorization: 0.333632
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1236,7 +1236,7 @@ more cache friendly.
.. code-block:: none
- loop permutation: 0.116249
+ loop permutation: 0.116098
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1327,7 +1327,7 @@ optimized schedule.
.. code-block:: none
- array packing: 0.107829
+ array packing: 0.109266
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1410,7 +1410,7 @@ to `C` when all the block results are ready.
.. code-block:: none
- block caching: 0.110483
+ block caching: 0.110672
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1484,7 +1484,7 @@ of thread-level parallelization.
.. code-block:: none
- parallelization: 0.146349
+ parallelization: 0.146325
# from tvm.script import ir as I
# from tvm.script import tir as T
@@ -1554,13 +1554,13 @@ working, we can compare the results.
.. code-block:: none
Operator Timing Performance
- none 3.4105436585 1.0
- blocking 0.3070694962 0.09003535123636361
- vectorization 0.3368725663 0.09877386130519754
- loop permutation 0.1162493442 0.03408528253560839
- array packing 0.10782897310000002 0.031616359119538306
- block caching 0.1104832868 0.032394626154292346
- parallelization 0.14634923719999998 0.042910823567749376
+ none 3.4134410433999998 1.0
+ blocking 0.2974181377 0.08713147053618157
+ vectorization 0.3336320781 0.09774068860661547
+ loop permutation 0.11609818999999999 0.03401206832749599
+ array packing 0.1092662241 0.03201057897609506
+ block caching 0.11067179900000002 0.03242235550368962
+ parallelization 0.146325404 0.04286741799244635
@@ -1602,7 +1602,7 @@ the computation for specific platforms.
.. rst-class:: sphx-glr-timing
- **Total running time of the script:** ( 1 minutes 1.053 seconds)
+ **Total running time of the script:** ( 1 minutes 0.965 seconds)
.. _sphx_glr_download_tutorial_tensor_expr_get_started.py:
diff --git a/docs/commit_hash b/docs/commit_hash
index ccfcbde23c..d3e6be8d16 100644
--- a/docs/commit_hash
+++ b/docs/commit_hash
@@ -1 +1 @@
-18b7dc1dd928bd3e621ff68ad07b354974be13a8
+56771a87d1560f8963cf745ad093ffc3d83f3f6a
diff --git a/docs/how_to/compile_models/from_darknet.html b/docs/how_to/compile_models/from_darknet.html
index c7a9d079a5..14c4942af2 100644
--- a/docs/how_to/compile_models/from_darknet.html
+++ b/docs/how_to/compile_models/from_darknet.html
@@ -585,7 +585,7 @@ 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 14.131 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 17.355 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 09bd4103ee..a19a1d4a7f 100644
--- a/docs/how_to/compile_models/from_keras.html
+++ b/docs/how_to/compile_models/from_keras.html
@@ -506,7 +506,7 @@ 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 947ms/step
+1/1 [==============================] - 1s 954ms/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 6e75c9dc48..f54a81e548 100644
--- a/docs/how_to/compile_models/from_mxnet.html
+++ b/docs/how_to/compile_models/from_mxnet.html
@@ -439,7 +439,7 @@
<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.zip28c6522a-69df-4ac8-ab68-ff22f5a4cbfc 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.zipa0ce24eb-8bfb-45e2-aef5-6e7943372f9a 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 9db54b4c22..0ab0a73d0a 100644
--- a/docs/how_to/compile_models/from_oneflow.html
+++ b/docs/how_to/compile_models/from_oneflow.html
@@ -449,11 +449,12 @@ 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]
- 19%|#9 | 7.99M/41.5M [00:00<00:00, 77.5MB/s]
- 39%|###8 | 16.0M/41.5M [00:00<00:00, 75.0MB/s]
- 58%|#####7 | 24.0M/41.5M [00:00<00:00, 66.7MB/s]
- 90%|########9 | 37.2M/41.5M [00:00<00:00, 85.3MB/s]
-100%|##########| 41.5M/41.5M [00:00<00:00, 78.9MB/s]
+ 19%|#9 | 7.99M/41.5M [00:00<00:00, 35.3MB/s]
+ 39%|###8 | 16.0M/41.5M [00:00<00:00, 40.9MB/s]
+ 58%|#####7 | 24.0M/41.5M [00:00<00:00, 45.4MB/s]
+ 77%|#######7 | 32.0M/41.5M [00:00<00:00, 52.5MB/s]
+ 92%|#########2| 38.3M/41.5M [00:00<00:00, 49.9MB/s]
+100%|##########| 41.5M/41.5M [00:00<00:00, 48.6MB/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 bf244bce64..9ebca65fd9 100644
--- a/docs/how_to/compile_models/from_pytorch.html
+++ b/docs/how_to/compile_models/from_pytorch.html
@@ -432,10 +432,13 @@ 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]
- 27%|##7 | 12.2M/44.7M [00:00<00:00, 128MB/s]
- 55%|#####4 | 24.4M/44.7M [00:00<00:00, 101MB/s]
- 80%|#######9 | 35.7M/44.7M [00:00<00:00, 107MB/s]
-100%|##########| 44.7M/44.7M [00:00<00:00, 106MB/s]
+ 14%|#4 | 6.30M/44.7M [00:00<00:00, 60.9MB/s]
+ 32%|###2 | 14.3M/44.7M [00:00<00:00, 48.1MB/s]
+ 49%|####8 | 21.7M/44.7M [00:00<00:00, 58.3MB/s]
+ 62%|######1 | 27.7M/44.7M [00:00<00:00, 58.6MB/s]
+ 81%|########1 | 36.4M/44.7M [00:00<00:00, 68.9MB/s]
+ 97%|#########6| 43.3M/44.7M [00:00<00:00, 68.7MB/s]
+100%|##########| 44.7M/44.7M [00:00<00:00, 56.8MB/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 8490280628..ef4f56d936 100644
--- a/docs/how_to/compile_models/from_tensorflow.html
+++ b/docs/how_to/compile_models/from_tensorflow.html
@@ -649,7 +649,7 @@ 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 21.545 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 22.479 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 f83dee6ccd..1d8f5a8591 100644
--- a/docs/how_to/compile_models/sg_execution_times.html
+++ b/docs/how_to/compile_models/sg_execution_times.html
@@ -340,7 +340,7 @@
<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>06:19.061</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
+<p><strong>06:25.962</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 81%" />
@@ -349,43 +349,43 @@
</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>01:21.545</p></td>
+<td><p>01:22.479</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>01:14.131</p></td>
+<td><p>01:17.355</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>00:52.717</p></td>
+<td><p>00:53.409</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>00:35.828</p></td>
+<td><p>00:35.917</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>00:30.544</p></td>
+<td><p>00:30.417</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>00:30.531</p></td>
+<td><p>00:30.381</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>00:26.388</p></td>
+<td><p>00:27.487</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>00:24.511</p></td>
+<td><p>00:25.115</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>00:20.244</p></td>
+<td><p>00:20.758</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>00:02.622</p></td>
+<td><p>00:02.644</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 cf7e9611ea..70f1eee9d0 100644
--- a/docs/how_to/deploy_models/deploy_model_on_adreno.html
+++ b/docs/how_to/deploy_models/deploy_model_on_adreno.html
@@ -920,10 +920,10 @@ Top5 predictions:
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 3341.0008 3340.7200 3343.2723 3338.7640 1.4712
+ 3342.1493 3340.8745 3354.0181 3338.0046 4.3589
</pre></div>
</div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 3.163 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 3.134 seconds)</p>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-deploy-models-deploy-model-on-adreno-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../../_downloads/2387d8448da213eb625e6b3d916327d4/deploy_model_on_adreno.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_model_on_adreno.py</span></code></a></p>
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 be4f07e68e..b54dc19b88 100644
--- a/docs/how_to/deploy_models/deploy_model_on_android.html
+++ b/docs/how_to/deploy_models/deploy_model_on_android.html
@@ -662,7 +662,7 @@ to the remote android device.</p>
Evaluate inference time cost...
Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 16.6069 16.7762 17.1110 16.0513 0.3718
+ 15.8913 15.6132 16.6356 15.4781 0.4431
</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 e79d6ac36d..6b82d739e7 100644
--- a/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
+++ b/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
@@ -454,23 +454,26 @@ 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<00:02, 83.3MB/s]
- 13%|#2 | 21.3M/170M [00:00<00:01, 116MB/s]
- 19%|#9 | 32.3M/170M [00:00<00:01, 106MB/s]
- 25%|##5 | 42.6M/170M [00:00<00:01, 105MB/s]
- 31%|### | 52.6M/170M [00:00<00:01, 103MB/s]
- 37%|###6 | 62.5M/170M [00:00<00:01, 102MB/s]
- 43%|####2 | 72.3M/170M [00:00<00:01, 101MB/s]
- 48%|####8 | 81.9M/170M [00:00<00:00, 96.7MB/s]
- 55%|#####4 | 92.9M/170M [00:00<00:00, 102MB/s]
- 60%|###### | 103M/170M [00:01<00:00, 99.3MB/s]
- 66%|######6 | 112M/170M [00:01<00:00, 100MB/s]
- 72%|#######2 | 123M/170M [00:01<00:00, 102MB/s]
- 78%|#######7 | 132M/170M [00:01<00:00, 96.5MB/s]
- 84%|########4 | 143M/170M [00:01<00:00, 102MB/s]
- 90%|######### | 153M/170M [00:01<00:00, 88.6MB/s]
- 99%|#########8| 167M/170M [00:01<00:00, 105MB/s]
-100%|##########| 170M/170M [00:01<00:00, 101MB/s]
+ 5%|4 | 7.99M/170M [00:00<00:04, 35.0MB/s]
+ 9%|9 | 16.0M/170M [00:00<00:03, 44.6MB/s]
+ 14%|#4 | 24.0M/170M [00:00<00:03, 49.0MB/s]
+ 18%|#7 | 30.3M/170M [00:00<00:03, 48.2MB/s]
+ 21%|## | 35.0M/170M [00:00<00:03, 35.7MB/s]
+ 25%|##4 | 42.1M/170M [00:01<00:03, 40.1MB/s]
+ 30%|##9 | 50.2M/170M [00:01<00:02, 50.1MB/s]
+ 35%|###5 | 59.7M/170M [00:01<00:01, 61.7MB/s]
+ 39%|###9 | 66.4M/170M [00:01<00:01, 63.9MB/s]
+ 47%|####7 | 80.0M/170M [00:01<00:01, 73.0MB/s]
+ 52%|#####1 | 88.0M/170M [00:01<00:01, 75.6MB/s]
+ 57%|#####6 | 96.0M/170M [00:01<00:01, 76.4MB/s]
+ 64%|######4 | 109M/170M [00:01<00:00, 93.6MB/s]
+ 70%|######9 | 119M/170M [00:01<00:00, 90.7MB/s]
+ 75%|#######5 | 128M/170M [00:02<00:00, 89.4MB/s]
+ 80%|######## | 136M/170M [00:02<00:00, 73.8MB/s]
+ 85%|########4 | 144M/170M [00:02<00:00, 71.0MB/s]
+ 92%|#########1| 156M/170M [00:02<00:00, 83.6MB/s]
+ 98%|#########7| 166M/170M [00:02<00:00, 83.4MB/s]
+100%|##########| 170M/170M [00:02<00:00, 66.9MB/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=& [...]
@@ -568,7 +571,7 @@ 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 27.786 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes 27.731 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 8c97dcad5f..c16be41934 100644
--- a/docs/how_to/deploy_models/deploy_prequantized.html
+++ b/docs/how_to/deploy_models/deploy_prequantized.html
@@ -495,8 +495,8 @@ 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]
-100%|#########9| 13.5M/13.6M [00:00<00:00, 142MB/s]
-100%|##########| 13.6M/13.6M [00:00<00:00, 141MB/s]
+ 59%|#####8 | 7.99M/13.6M [00:00<00:00, 36.8MB/s]
+100%|##########| 13.6M/13.6M [00:00<00:00, 46.7MB/s]
</pre></div>
</div>
</div>
@@ -587,7 +587,7 @@ 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)
- 90.2892 90.2321 90.9815 90.0854 0.1888
+ 90.6740 90.6287 95.8723 90.2278 0.5534
</pre></div>
</div>
<div class="admonition note">
@@ -626,7 +626,7 @@ 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 14.094 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 13.967 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 b1d7036796..c357134978 100644
--- a/docs/how_to/deploy_models/deploy_prequantized_tflite.html
+++ b/docs/how_to/deploy_models/deploy_prequantized_tflite.html
@@ -580,7 +580,7 @@ TFLite Top-5 labels: [387 102 386 341 349]
</div>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time summary:
mean (ms) median (ms) max (ms) min (ms) std (ms)
- 120.8368 120.7806 123.1834 120.0746 0.4294
+ 119.6954 119.7157 120.3397 118.9367 0.2781
</pre></div>
</div>
<div class="admonition note">
@@ -608,7 +608,7 @@ 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 34.550 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 2 minutes 35.610 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 caa56d99bd..20bba51cb8 100644
--- a/docs/how_to/deploy_models/deploy_quantized.html
+++ b/docs/how_to/deploy_models/deploy_quantized.html
@@ -521,7 +521,7 @@ 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 20.755 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 42.340 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 26f96c9351..8aa76b3fad 100644
--- a/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
+++ b/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
@@ -463,23 +463,30 @@ 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 | 6147/132723 [00:00<00:02, 61464.05KB/s]
- 11%|# | 14511/132723 [00:00<00:01, 74505.85KB/s]
- 17%|#6 | 21962/132723 [00:00<00:01, 59073.28KB/s]
- 21%|##1 | 28175/132723 [00:00<00:01, 59562.02KB/s]
- 27%|##6 | 35273/132723 [00:00<00:01, 63275.70KB/s]
- 33%|###3 | 44041/132723 [00:00<00:01, 71017.19KB/s]
- 40%|###9 | 52694/132723 [00:00<00:01, 75854.81KB/s]
- 46%|####6 | 61385/132723 [00:00<00:00, 79259.51KB/s]
- 53%|#####2 | 70096/132723 [00:00<00:00, 81651.16KB/s]
- 59%|#####9 | 78933/132723 [00:01<00:00, 83691.51KB/s]
- 66%|######5 | 87584/132723 [00:01<00:00, 84539.97KB/s]
- 73%|#######2 | 96432/132723 [00:01<00:00, 85722.14KB/s]
- 79%|#######9 | 105229/132723 [00:01<00:00, 86395.14KB/s]
- 86%|########5 | 114055/132723 [00:01<00:00, 86954.14KB/s]
- 93%|#########2| 122821/132723 [00:01<00:00, 87162.81KB/s]
- 99%|#########9| 131635/132723 [00:01<00:00, 87454.03KB/s]
-100%|##########| 132723/132723 [00:01<00:00, 79248.71KB/s]
+ 3%|2 | 3631/132723 [00:00<00:03, 34719.86KB/s]
+ 6%|5 | 7422/132723 [00:00<00:03, 36451.66KB/s]
+ 9%|8 | 11422/132723 [00:00<00:03, 38043.10KB/s]
+ 12%|#1 | 15700/132723 [00:00<00:02, 39898.17KB/s]
+ 15%|#4 | 19728/132723 [00:00<00:02, 40034.30KB/s]
+ 18%|#8 | 24286/132723 [00:00<00:02, 41898.24KB/s]
+ 22%|##1 | 28587/132723 [00:00<00:02, 42259.41KB/s]
+ 25%|##5 | 33566/132723 [00:00<00:02, 44400.12KB/s]
+ 29%|##9 | 38926/132723 [00:00<00:01, 47037.46KB/s]
+ 33%|###2 | 43628/132723 [00:01<00:01, 44831.40KB/s]
+ 37%|###7 | 49178/132723 [00:01<00:01, 47972.57KB/s]
+ 41%|####1 | 54770/132723 [00:01<00:01, 50324.63KB/s]
+ 45%|####5 | 59918/132723 [00:01<00:01, 50577.75KB/s]
+ 50%|####9 | 65806/132723 [00:01<00:01, 52883.19KB/s]
+ 54%|#####4 | 71886/132723 [00:01<00:01, 55129.48KB/s]
+ 59%|#####8 | 78148/132723 [00:01<00:00, 57359.85KB/s]
+ 64%|######3 | 84764/132723 [00:01<00:00, 59986.53KB/s]
+ 69%|######8 | 91414/132723 [00:01<00:00, 61932.02KB/s]
+ 74%|#######3 | 97947/132723 [00:01<00:00, 62946.66KB/s]
+ 79%|#######9 | 105128/132723 [00:02<00:00, 65599.10KB/s]
+ 85%|########4 | 112169/132723 [00:02<00:00, 67039.19KB/s]
+ 90%|######### | 119673/132723 [00:02<00:00, 69434.24KB/s]
+ 96%|#########5| 127136/132723 [00:02<00:00, 70989.63KB/s]
+100%|##########| 132723/132723 [00:02<00:00, 55008.50KB/s]
</pre></div>
</div>
<p>Create TVM runtime and do inference
@@ -518,7 +525,7 @@ 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 34.842 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 33.963 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 f2ca1a3747..161e704804 100644
--- a/docs/how_to/deploy_models/sg_execution_times.html
+++ b/docs/how_to/deploy_models/sg_execution_times.html
@@ -340,7 +340,7 @@
<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>14:51.380</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
+<p><strong>15:12.069</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 86%" />
@@ -349,39 +349,39 @@
</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>03:34.842</p></td>
+<td><p>03:33.963</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>03:27.786</p></td>
+<td><p>03:27.731</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>02:34.550</p></td>
+<td><p>02:35.610</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>01:20.755</p></td>
+<td><p>01:42.340</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>01:14.094</p></td>
+<td><p>01:13.967</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>01:03.163</p></td>
+<td><p>01:03.134</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><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>00:40.880</p></td>
+<td><p>00:40.390</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><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>00:27.715</p></td>
+<td><p>00:27.686</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><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>00:27.588</p></td>
+<td><p>00:27.241</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><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 ca4803f18c..3745f760e0 100644
--- a/docs/how_to/extend_tvm/bring_your_own_datatypes.html
+++ b/docs/how_to/extend_tvm/bring_your_own_datatypes.html
@@ -619,7 +619,7 @@ 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.zip2e21da86-954a-4776-ad44-0a8bcc9cb300 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.zip338d35f6-a39e-477d-bc04-2e9c1ba5f341 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 73c16ce056..8306188e75 100644
--- a/docs/how_to/extend_tvm/sg_execution_times.html
+++ b/docs/how_to/extend_tvm/sg_execution_times.html
@@ -340,7 +340,7 @@
<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>00:53.112</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
+<p><strong>00:53.023</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -349,15 +349,15 @@
</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>00:49.323</p></td>
+<td><p>00:49.270</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>00:02.708</p></td>
+<td><p>00:02.684</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>00:01.062</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>
diff --git a/docs/how_to/extend_tvm/use_pass_instrument.html b/docs/how_to/extend_tvm/use_pass_instrument.html
index 6aa58b91a7..6533776fd5 100644
--- a/docs/how_to/extend_tvm/use_pass_instrument.html
+++ b/docs/how_to/extend_tvm/use_pass_instrument.html
@@ -526,10 +526,10 @@ 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: 22115us [22115us] (49.77%; 49.77%)
-FoldScaleAxis: 22321us [8us] (50.23%; 50.23%)
- FoldConstant: 22313us [1754us] (50.21%; 99.97%)
- InferType: 20559us [20559us] (46.27%; 92.14%)
+InferType: 21208us [21208us] (48.16%; 48.16%)
+FoldScaleAxis: 22830us [7us] (51.84%; 51.84%)
+ FoldConstant: 22823us [1672us] (51.83%; 99.97%)
+ InferType: 21151us [21151us] (48.03%; 92.67%)
</pre></div>
</div>
</div>
@@ -551,10 +551,10 @@ 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: 20728us [20728us] (48.25%; 48.25%)
-FoldScaleAxis: 22232us [6us] (51.75%; 51.75%)
- FoldConstant: 22227us [1738us] (51.74%; 99.98%)
- InferType: 20489us [20489us] (47.69%; 92.18%)
+InferType: 20662us [20662us] (48.06%; 48.06%)
+FoldScaleAxis: 22327us [5us] (51.94%; 51.94%)
+ FoldConstant: 22322us [1720us] (51.92%; 99.98%)
+ InferType: 20602us [20602us] (47.92%; 92.29%)
</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 5a0f13b0a8..ea3c669310 100644
--- a/docs/how_to/optimize_operators/opt_conv_cuda.html
+++ b/docs/how_to/optimize_operators/opt_conv_cuda.html
@@ -575,7 +575,7 @@ 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: 36.795871 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Convolution: 46.436958 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 5d34b03ca3..276d23c6ed 100644
--- a/docs/how_to/optimize_operators/opt_conv_tensorcore.html
+++ b/docs/how_to/optimize_operators/opt_conv_tensorcore.html
@@ -867,7 +867,7 @@ be able to run on our build server</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"conv2d with tensor core: </span><span class="si">%f</span><span class="s2"> ms"</span> <span class="o">%</span> <span class="p">(</span><span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">w</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span> <span class="o">* [...]
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>conv2d with tensor core: 7.718617 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>conv2d with tensor core: 6.909338 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 0c77cf3165..06349b710b 100644
--- a/docs/how_to/optimize_operators/opt_gemm.html
+++ b/docs/how_to/optimize_operators/opt_gemm.html
@@ -472,8 +472,8 @@ 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: 0.018811
-Baseline: 3.396133
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Numpy running time: 0.018045
+Baseline: 3.425894
</pre></div>
</div>
<p>In TVM, we can always inspect lower level IR to debug or optimize our schedule.
@@ -532,7 +532,7 @@ 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: 0.302338
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt1: 0.293250
</pre></div>
</div>
<p>Here is the generated IR after blocking.</p>
@@ -589,7 +589,7 @@ vastly.</p>
<span class="nb">print</span><span class="p">(</span><span class="s2">"Opt2: </span><span class="si">%f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">)</span>
</pre></div>
</div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt2: 0.332977
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt2: 0.331320
</pre></div>
</div>
<p>Here is the generated IR after vectorization.</p>
@@ -644,7 +644,7 @@ 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: 0.116724
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt3: 0.116212
</pre></div>
</div>
<p>Here is the generated IR after loop permutation.</p>
@@ -721,7 +721,7 @@ 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: 0.109741
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt4: 0.109608
</pre></div>
</div>
<p>Here is the generated IR after array packing.</p>
@@ -799,7 +799,7 @@ 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: 0.110908
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt5: 0.111050
</pre></div>
</div>
<p>Here is the generated IR after blocking.</p>
@@ -879,7 +879,7 @@ 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: 0.146963
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt6: 0.146901
</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 dd9b95c5c3..f303ebf44b 100644
--- a/docs/how_to/optimize_operators/sg_execution_times.html
+++ b/docs/how_to/optimize_operators/sg_execution_times.html
@@ -340,7 +340,7 @@
<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>00:34.845</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
+<p><strong>00:34.690</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -349,15 +349,15 @@
</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>00:32.301</p></td>
+<td><p>00:32.165</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>00:01.475</p></td>
+<td><p>00:01.449</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>00:01.068</p></td>
+<td><p>00:01.075</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 9a714594a6..b8a67bc60c 100644
--- a/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
+++ b/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
@@ -340,7 +340,7 @@
<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>09:23.537</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
+<p><strong>09:16.438</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 85%" />
@@ -349,27 +349,27 @@
</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>05:42.565</p></td>
+<td><p>05:35.291</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>01:39.693</p></td>
+<td><p>01:39.443</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>01:05.839</p></td>
+<td><p>01:06.161</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>00:28.284</p></td>
+<td><p>00:28.467</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>00:14.117</p></td>
+<td><p>00:14.108</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>00:13.038</p></td>
+<td><p>00:12.969</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 02a19c8060..453939b5a4 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
@@ -506,13 +506,13 @@ class Module:
def main(data: T.Buffer((1, 512, 7, 7), "float32"), kernel: T.Buffer((512, 512, 3, 3), "float32"), bias: T.Buffer((1, 512, 1, 1), "float32"), compute: T.Buffer((1, 512, 7, 7), "float32")):
T.func_attr({"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True})
blockIdx_x = T.env_thread("blockIdx.x")
- T.launch_thread(blockIdx_x, 112)
+ T.launch_thread(blockIdx_x, 28)
conv2d_nchw = T.allocate([7], "float32", "local")
- pad_temp_shared = T.allocate([432], "float32", "shared")
- kernel_shared = T.allocate([4608], "float32", "shared")
+ pad_temp_shared = T.allocate([216], "float32", "shared")
+ kernel_shared = T.allocate([9216], "float32", "shared")
threadIdx_x = T.env_thread("threadIdx.x")
- T.launch_thread(threadIdx_x, 32)
- conv2d_nchw_1 = T.Buffer((7,), data=conv2d_nchw, scope="local", align=16)
+ T.launch_thread(threadIdx_x, 128)
+ conv2d_nchw_1 = T.Buffer((1,), data=conv2d_nchw, scope="local", align=4)
conv2d_nchw_1[0] = T.float32(0)
conv2d_nchw_1[1] = T.float32(0)
conv2d_nchw_1[2] = T.float32(0)
@@ -520,786 +520,676 @@ class Module:
conv2d_nchw_1[4] = T.float32(0)
conv2d_nchw_1[5] = T.float32(0)
conv2d_nchw_1[6] = T.float32(0)
- for rc_outer_outer in range(32):
- cse_var_2: T.int32 = rc_outer_outer * 784
- cse_var_1: T.int32 = rc_outer_outer * 144
+ for rc_outer_outer in range(64):
+ cse_var_1: T.int32 = rc_outer_outer * 72
threadIdx_x_1 = T.env_thread("threadIdx.x")
- pad_temp_shared_1 = T.Buffer((432,), data=pad_temp_shared, scope="shared")
+ pad_temp_shared_1 = T.Buffer((216,), data=pad_temp_shared, scope="shared")
data_1 = T.Buffer((25088,), data=data.data)
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(1 <= threadIdx_x_1 % 27 // 9 + blockIdx_x % 7 and threadIdx_x_1 % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_2 + threadIdx_x_1 // 27 * 49 + threadIdx_x_1 % 27 // 9 * 7 + blockIdx_x % 7 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 32] = T.if_then_else(1 <= (threadIdx_x_1 + 5) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 5) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 32) // 27 * 49 + (threadIdx_x_1 + 5) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 64] = T.if_then_else(1 <= (threadIdx_x_1 + 10) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 10) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 64) // 27 * 49 + (threadIdx_x_1 + 10) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 96] = T.if_then_else(1 <= (threadIdx_x_1 + 15) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 15) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 96) // 27 * 49 + (threadIdx_x_1 + 15) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 128] = T.if_then_else(1 <= (threadIdx_x_1 + 20) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 20) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 128) // 27 * 49 + (threadIdx_x_1 + 20) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 160] = T.if_then_else(1 <= (threadIdx_x_1 + 25) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 25) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 160) // 27 * 49 + (threadIdx_x_1 + 25) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 192] = T.if_then_else(1 <= (threadIdx_x_1 + 3) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 3) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 192) // 27 * 49 + (threadIdx_x_1 + 3) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 224] = T.if_then_else(1 <= (threadIdx_x_1 + 8) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 8) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 224) // 27 * 49 + (threadIdx_x_1 + 8) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 256] = T.if_then_else(1 <= (threadIdx_x_1 + 13) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 13) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 256) // 27 * 49 + (threadIdx_x_1 + 13) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 288] = T.if_then_else(1 <= blockIdx_x % 7 + (threadIdx_x_1 // 9 + 2) % 3 and blockIdx_x % 7 + (threadIdx_x_1 // 9 + 2) % 3 < 8 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 288) // 27 * 49 + blockIdx_x % 7 * 7 + (threadIdx_x_1 // 9 + 2) % 3 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 320] = T.if_then_else(1 <= (threadIdx_x_1 + 23) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 23) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 5) % 9 and (threadIdx_x_1 + 5) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 320) // 27 * 49 + (threadIdx_x_1 + 23) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 5) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 352] = T.if_then_else(1 <= (threadIdx_x_1 + 1) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 1) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 1) % 9 and (threadIdx_x_1 + 1) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 352) // 27 * 49 + (threadIdx_x_1 + 1) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 1) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- pad_temp_shared_1[threadIdx_x_1 + 384] = T.if_then_else(1 <= (threadIdx_x_1 + 6) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 6) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 6) % 9 and (threadIdx_x_1 + 6) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 384) // 27 * 49 + (threadIdx_x_1 + 6) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 6) % 9 - 8], T.float32(0))
- with T.launch_thread(threadIdx_x_1, 32):
- if T.likely(threadIdx_x_1 < 16):
- pad_temp_shared_1[threadIdx_x_1 + 416] = T.if_then_else((threadIdx_x_1 + 11) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[cse_var_2 + (threadIdx_x_1 + 416) // 27 * 49 + (threadIdx_x_1 + 11) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
+ with T.launch_thread(threadIdx_x_1, 128):
+ pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(1 <= threadIdx_x_1 % 27 // 9 + blockIdx_x % 7 and threadIdx_x_1 % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[rc_outer_outer * 392 + threadIdx_x_1 // 27 * 49 + threadIdx_x_1 % 27 // 9 * 7 + blockIdx_x % 7 * 7 + threadIdx_x_1 % 9 - 8], T.float32(0))
+ with T.launch_thread(threadIdx_x_1, 128):
+ if T.likely(threadIdx_x_1 < 88):
+ pad_temp_shared_1[threadIdx_x_1 + 128] = T.if_then_else(1 <= (threadIdx_x_1 + 20) % 27 // 9 + blockIdx_x % 7 and (threadIdx_x_1 + 20) % 27 // 9 + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 + 2) % 9 and (threadIdx_x_1 + 2) % 9 < 8, data_1[rc_outer_outer * 392 + (threadIdx_x_1 + 128) // 27 * 49 + (threadIdx_x_1 + 20) % 27 // 9 * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 + 2) % 9 - 8], T.float32(0))
threadIdx_x_2 = T.env_thread("threadIdx.x")
- kernel_shared_1 = T.Buffer((4608,), data=kernel_shared, scope="shared")
+ kernel_shared_1 = T.Buffer((9216,), data=kernel_shared, scope="shared")
kernel_1 = T.Buffer((2359296,), data=kernel.data)
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 32] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + (threadIdx_x_2 + 32) // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 64] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + (threadIdx_x_2 + 64) // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 96] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 96]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 128] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 128) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 160] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 160) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 192] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 192) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 224] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 224) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 256] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 256) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 288] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 9216]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 320] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 320) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 352] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 352) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 384] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 384) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 416] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 416) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 448] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 448) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 480] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 480) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 512] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 512) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 544] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 544) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 576] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 18432]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 608] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 608) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 640] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 640) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 672] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 672) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 704] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 704) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 736] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 736) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 768] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 768) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 800] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 800) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 832] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 832) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 864] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 27648]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 896] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 896) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 928] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 928) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 960] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 960) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 992] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 992) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1024] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1024) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1056] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1056) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1088] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1088) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1120] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1120) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1152] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 36864]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1184] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1184) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1216] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1216) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1248] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1248) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1280] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1280) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1312] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1312) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1344] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1344) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1376] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1376) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1408] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1408) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1440] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 46080]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1472] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1472) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1504] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1504) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1536] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1536) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1568] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1568) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1600] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1600) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1632] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1632) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1664] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1664) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1696] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1696) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1728] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 55296]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1760] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1760) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1792] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1792) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1824] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1824) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1856] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1856) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1888] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1888) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1920] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1920) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1952] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1952) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 1984] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 1984) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2016] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 64512]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2048] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2048) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2080] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2080) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2112] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2112) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2144] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2144) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2176] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2176) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2208] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2208) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2240] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2240) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2272] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2272) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2304] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 73728]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2336] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2336) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2368] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2368) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2400] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2400) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2432] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2432) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2464] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2464) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2496] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2496) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2528] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2528) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2560] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2560) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2592] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 82944]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2624] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2624) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2656] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2656) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2688] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2688) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2720] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2720) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2752] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2752) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2784] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2784) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2816] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2816) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2848] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2848) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2880] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 92160]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2912] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2912) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2944] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2944) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 2976] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 2976) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3008] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3008) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3040] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3040) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3072] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3072) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3104] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3104) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3136] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3136) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3168] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 101376]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3200] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3200) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3232] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3232) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3264] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3264) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3296] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3296) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3328] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3328) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3360] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3360) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3392] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3392) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3424] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3424) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3456] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 110592]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3488] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3488) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3520] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3520) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3552] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3552) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3584] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3584) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3616] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3616) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3648] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3648) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3680] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3680) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3712] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3712) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3744] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 119808]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3776] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3776) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3808] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3808) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3840] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3840) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3872] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3872) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3904] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3904) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3936] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3936) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 3968] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 3968) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4000] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4000) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4032] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 129024]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4064] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4064) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4096] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4096) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4128] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4128) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4160] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4160) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4192] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4192) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4224] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4224) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4256] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4256) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4288] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4288) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4320] = kernel_1[blockIdx_x // 7 * 147456 + cse_var_1 + threadIdx_x_2 + 138240]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4352] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4352) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4384] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4384) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4416] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4416) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 32) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4448] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4448) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 128) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4480] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4480) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4512] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4512) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 // 3 + 16) * 3 + threadIdx_x_2 % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4544] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4544) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 80) % 144 // 3 * 3 + (threadIdx_x_2 + 2) % 3]
- with T.launch_thread(threadIdx_x_2, 32):
- kernel_shared_1[threadIdx_x_2 + 4576] = kernel_1[blockIdx_x // 7 * 147456 + (threadIdx_x_2 + 4576) // 144 * 4608 + cse_var_1 + (threadIdx_x_2 + 112) % 144 // 3 * 3 + (threadIdx_x_2 + 1) % 3]
- for ry_outer_inner in range(3):
- cse_var_115: T.int32 = ry_outer_inner * 9
- cse_var_114: T.int32 = cse_var_115 + 88
- cse_var_113: T.int32 = cse_var_115 + 87
- cse_var_112: T.int32 = cse_var_115 + 86
- cse_var_111: T.int32 = cse_var_115 + 85
- cse_var_110: T.int32 = cse_var_115 + 84
- cse_var_109: T.int32 = cse_var_115 + 83
- cse_var_108: T.int32 = cse_var_115 + 82
- cse_var_107: T.int32 = cse_var_115 + 7
- cse_var_106: T.int32 = cse_var_115 + 61
- cse_var_105: T.int32 = cse_var_115 + 60
- cse_var_104: T.int32 = cse_var_115 + 6
- cse_var_103: T.int32 = cse_var_115 + 59
- cse_var_102: T.int32 = cse_var_115 + 58
- cse_var_101: T.int32 = cse_var_115 + 57
- cse_var_100: T.int32 = cse_var_115 + 56
- cse_var_99: T.int32 = cse_var_115 + 55
- cse_var_98: T.int32 = cse_var_115 + 5
- cse_var_97: T.int32 = cse_var_115 + 412
- cse_var_96: T.int32 = cse_var_115 + 411
- cse_var_95: T.int32 = cse_var_115 + 410
- cse_var_94: T.int32 = cse_var_115 + 409
- cse_var_93: T.int32 = cse_var_115 + 408
- cse_var_92: T.int32 = cse_var_115 + 407
- cse_var_91: T.int32 = cse_var_115 + 406
- cse_var_90: T.int32 = cse_var_115 + 4
- cse_var_89: T.int32 = cse_var_115 + 385
- cse_var_88: T.int32 = cse_var_115 + 384
- cse_var_87: T.int32 = cse_var_115 + 383
- cse_var_86: T.int32 = cse_var_115 + 382
- cse_var_85: T.int32 = cse_var_115 + 381
- cse_var_84: T.int32 = cse_var_115 + 380
- cse_var_83: T.int32 = cse_var_115 + 379
- cse_var_82: T.int32 = cse_var_115 + 358
- cse_var_81: T.int32 = cse_var_115 + 357
- cse_var_80: T.int32 = cse_var_115 + 356
- cse_var_79: T.int32 = cse_var_115 + 355
- cse_var_78: T.int32 = cse_var_115 + 354
- cse_var_77: T.int32 = cse_var_115 + 353
- cse_var_76: T.int32 = cse_var_115 + 352
- cse_var_75: T.int32 = cse_var_115 + 34
- cse_var_74: T.int32 = cse_var_115 + 331
- cse_var_73: T.int32 = cse_var_115 + 330
- cse_var_72: T.int32 = cse_var_115 + 33
- cse_var_71: T.int32 = cse_var_115 + 329
- cse_var_70: T.int32 = cse_var_115 + 328
- cse_var_69: T.int32 = cse_var_115 + 327
- cse_var_68: T.int32 = cse_var_115 + 326
- cse_var_67: T.int32 = cse_var_115 + 325
- cse_var_66: T.int32 = cse_var_115 + 32
- cse_var_65: T.int32 = cse_var_115 + 31
- cse_var_64: T.int32 = cse_var_115 + 304
- cse_var_63: T.int32 = cse_var_115 + 303
- cse_var_62: T.int32 = cse_var_115 + 302
- cse_var_61: T.int32 = cse_var_115 + 301
- cse_var_60: T.int32 = cse_var_115 + 300
- cse_var_59: T.int32 = cse_var_115 + 30
- cse_var_58: T.int32 = cse_var_115 + 3
- cse_var_57: T.int32 = cse_var_115 + 299
- cse_var_56: T.int32 = cse_var_115 + 298
- cse_var_55: T.int32 = cse_var_115 + 29
- cse_var_54: T.int32 = cse_var_115 + 28
- cse_var_53: T.int32 = cse_var_115 + 277
- cse_var_52: T.int32 = cse_var_115 + 276
- cse_var_51: T.int32 = cse_var_115 + 275
- cse_var_50: T.int32 = cse_var_115 + 274
- cse_var_49: T.int32 = cse_var_115 + 273
- cse_var_48: T.int32 = cse_var_115 + 272
- cse_var_47: T.int32 = cse_var_115 + 271
- cse_var_46: T.int32 = cse_var_115 + 250
- cse_var_45: T.int32 = cse_var_115 + 249
- cse_var_44: T.int32 = cse_var_115 + 248
- cse_var_43: T.int32 = cse_var_115 + 247
- cse_var_42: T.int32 = cse_var_115 + 246
- cse_var_41: T.int32 = cse_var_115 + 245
- cse_var_40: T.int32 = cse_var_115 + 244
- cse_var_39: T.int32 = cse_var_115 + 223
- cse_var_38: T.int32 = cse_var_115 + 222
- cse_var_37: T.int32 = cse_var_115 + 221
- cse_var_36: T.int32 = cse_var_115 + 220
- cse_var_35: T.int32 = cse_var_115 + 219
- cse_var_34: T.int32 = cse_var_115 + 218
- cse_var_33: T.int32 = cse_var_115 + 217
- cse_var_32: T.int32 = cse_var_115 + 2
- cse_var_31: T.int32 = cse_var_115 + 196
- cse_var_30: T.int32 = cse_var_115 + 195
- cse_var_29: T.int32 = cse_var_115 + 194
- cse_var_28: T.int32 = cse_var_115 + 193
- cse_var_27: T.int32 = cse_var_115 + 192
- cse_var_26: T.int32 = cse_var_115 + 191
- cse_var_25: T.int32 = cse_var_115 + 190
- cse_var_24: T.int32 = cse_var_115 + 169
- cse_var_23: T.int32 = cse_var_115 + 168
- cse_var_22: T.int32 = cse_var_115 + 167
- cse_var_21: T.int32 = cse_var_115 + 166
- cse_var_20: T.int32 = cse_var_115 + 165
- cse_var_19: T.int32 = cse_var_115 + 164
- cse_var_18: T.int32 = cse_var_115 + 163
- cse_var_17: T.int32 = cse_var_115 + 142
- cse_var_16: T.int32 = cse_var_115 + 141
- cse_var_15: T.int32 = cse_var_115 + 140
- cse_var_14: T.int32 = cse_var_115 + 139
- cse_var_13: T.int32 = cse_var_115 + 138
- cse_var_12: T.int32 = cse_var_115 + 137
- cse_var_11: T.int32 = cse_var_115 + 136
- cse_var_10: T.int32 = cse_var_115 + 115
- cse_var_9: T.int32 = cse_var_115 + 114
- cse_var_8: T.int32 = cse_var_115 + 113
- cse_var_7: T.int32 = cse_var_115 + 112
- cse_var_6: T.int32 = cse_var_115 + 111
- cse_var_5: T.int32 = cse_var_115 + 110
- cse_var_4: T.int32 = cse_var_115 + 109
- cse_var_3: T.int32 = cse_var_115 + 1
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_3] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_32] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 27] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_54] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_55] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 54] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_99] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_100] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 81] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_108] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_109] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 108] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_4] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_5] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 135] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_11] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_12] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 162] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_18] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_19] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 189] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_25] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_26] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 216] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_33] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_34] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 243] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_40] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_41] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 270] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_47] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_48] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 297] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_56] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_57] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 324] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_67] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_68] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 351] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_76] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_77] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 378] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_83] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_84] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_115 + 405] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_91] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[cse_var_92] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_3] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_32] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_58] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_54] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_55] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_59] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_99] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_100] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_101] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_108] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_109] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_110] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_4] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_5] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_6] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_11] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_12] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_13] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_18] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_19] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_20] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_25] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_26] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_27] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_33] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_34] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_35] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_40] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_41] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_42] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_47] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_48] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_49] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_56] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_57] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_60] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_67] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_68] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_69] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_76] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_77] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_78] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_83] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_84] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_85] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_91] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_92] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[cse_var_93] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_32] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_58] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_90] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_55] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_59] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_65] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_100] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_101] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_102] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_109] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_110] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_111] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_5] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_6] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_7] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_12] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_13] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_14] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_19] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_20] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_21] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_26] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_27] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_28] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_34] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_35] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_36] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_41] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_42] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_43] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_48] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_49] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_50] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_57] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_60] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_61] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_68] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_69] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_70] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_77] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_78] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_79] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_84] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_85] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_86] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_92] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_93] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[cse_var_94] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_58] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_90] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_98] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_59] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_65] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_66] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_101] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_102] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_103] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_110] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_111] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_112] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_6] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_7] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_8] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_13] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_14] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_15] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_20] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_21] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_22] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_27] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_28] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_29] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_35] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_36] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_37] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_42] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_43] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_44] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_49] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_50] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_51] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_60] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_61] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_62] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_69] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_70] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_71] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_78] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_79] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_80] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_85] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_86] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_87] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_93] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_94] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[cse_var_95] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_90] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_98] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_104] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_65] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_66] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_72] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_102] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_103] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_105] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_111] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_112] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_113] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_7] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_8] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_9] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_14] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_15] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_16] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_21] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_22] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_23] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_28] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_29] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_30] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_36] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_37] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_38] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_43] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_44] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_45] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_50] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_51] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_52] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_61] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_62] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_63] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_70] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_71] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_73] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_79] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_80] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_81] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_86] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_87] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_88] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_94] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_95] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[cse_var_96] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_98] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_104] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_107] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_66] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_72] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_75] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_103] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_105] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_106] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_112] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_113] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_114] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_8] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_9] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_10] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_15] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_16] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_17] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_22] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_23] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_24] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_29] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_30] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_31] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_37] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_38] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_39] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_44] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_45] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_46] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_51] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_52] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_53] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_62] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_63] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_64] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_71] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_73] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_74] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_80] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_81] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_82] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_87] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_88] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_89] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_95] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_96] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[cse_var_97] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_104] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_107] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 1]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 8] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 2]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_72] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 9]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_75] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 10]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 35] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 11]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_105] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 18]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_106] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 19]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 62] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 20]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_113] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 27]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_114] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 28]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 89] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 29]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_9] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 36]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_10] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 37]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 116] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 38]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_16] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 45]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_17] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 46]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 143] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 47]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_23] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 54]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_24] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 55]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 170] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 56]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_30] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 63]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_31] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 64]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 197] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 65]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_38] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 72]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_39] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 73]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 224] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 74]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_45] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 81]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_46] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 82]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 251] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 83]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_52] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 90]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_53] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 91]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 278] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 92]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_63] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 99]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_64] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 100]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 305] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 101]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_73] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 108]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_74] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 109]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 332] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 110]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_81] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 117]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_82] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 118]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 359] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 119]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_88] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 126]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_89] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 127]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 386] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 128]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_96] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 135]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_97] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 136]
- conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[cse_var_115 + 413] * kernel_shared_1[threadIdx_x * 144 + ry_outer_inner * 3 + 137]
- 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 // 7 * 1568 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + i3_inner] = T.max(conv2d_nchw_1[i3_inner] + bias_1[blockIdx_x // 7 * 32 + threadIdx_x], T.float32(0))
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 128] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 128) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 256] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 256) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 384] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 384) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 512] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 512) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 640] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 640) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 768] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 768) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 896] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 896) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1024] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1024) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1152] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 73728]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1280] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1280) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1408] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1408) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1536] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1536) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1664] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1664) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1792] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1792) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 1920] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 1920) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2048] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2048) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2176] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2176) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2304] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 147456]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2432] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2432) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2560] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2560) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2688] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2688) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2816] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2816) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 2944] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 2944) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3072] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3072) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3200] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3200) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3328] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3328) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3456] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 221184]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3584] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3584) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3712] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3712) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3840] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3840) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 3968] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 3968) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4096] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4096) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4224] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4224) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4352] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4352) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4480] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4480) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4608] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 294912]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4736] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4736) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4864] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4864) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 4992] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 4992) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5120] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5120) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5248] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5248) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5376] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5376) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5504] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5504) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5632] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5632) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5760] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 368640]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 5888] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 5888) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6016] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6016) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6144] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6144) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6272] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6272) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6400] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6400) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6528] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6528) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6656] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6656) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6784] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 6784) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 6912] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 442368]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7040] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7040) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7168] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7168) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7296] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7296) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7424] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7424) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7552] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7552) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7680] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7680) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7808] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7808) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 7936] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 7936) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8064] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_2 // 72 * 4608 + cse_var_1 + threadIdx_x_2 % 72 + 516096]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8192] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8192) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 56) % 72 // 9 * 9 + (threadIdx_x_2 + 2) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8320] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8320) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 40) % 72 // 9 * 9 + (threadIdx_x_2 + 4) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8448] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8448) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 24) % 72 // 9 * 9 + (threadIdx_x_2 + 6) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8576] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8576) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 8) % 72]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8704] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8704) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 64) % 72 // 9 * 9 + (threadIdx_x_2 + 1) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8832] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8832) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 48) % 72 // 9 * 9 + (threadIdx_x_2 + 3) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 8960] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 8960) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 32) % 72 // 9 * 9 + (threadIdx_x_2 + 5) % 9]
+ with T.launch_thread(threadIdx_x_2, 128):
+ kernel_shared_1[threadIdx_x_2 + 9088] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_2 + 9088) // 72 * 4608 + cse_var_1 + (threadIdx_x_2 + 16) % 72 // 9 * 9 + (threadIdx_x_2 + 7) % 9]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[0] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 72]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 72 + 1]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[8] * kernel_shared_1[threadIdx_x * 72 + 2]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[27] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 72 + 9]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 72 + 10]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[35] * kernel_shared_1[threadIdx_x * 72 + 11]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[54] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 72 + 18]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 72 + 19]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[62] * kernel_shared_1[threadIdx_x * 72 + 20]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[81] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[82] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[83] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[84] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[85] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[86] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[87] * kernel_shared_1[threadIdx_x * 72 + 27]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[82] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[83] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[84] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[85] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[86] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[87] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[88] * kernel_shared_1[threadIdx_x * 72 + 28]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[83] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[84] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[85] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[86] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[87] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[88] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[89] * kernel_shared_1[threadIdx_x * 72 + 29]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[108] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[109] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[110] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[111] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[112] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[113] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[114] * kernel_shared_1[threadIdx_x * 72 + 36]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[109] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[110] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[111] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[112] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[113] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[114] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[115] * kernel_shared_1[threadIdx_x * 72 + 37]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[110] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[111] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[112] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[113] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[114] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[115] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[116] * kernel_shared_1[threadIdx_x * 72 + 38]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[135] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[136] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[137] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[138] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[139] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[140] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[141] * kernel_shared_1[threadIdx_x * 72 + 45]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[136] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[137] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[138] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[139] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[140] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[141] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[142] * kernel_shared_1[threadIdx_x * 72 + 46]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[137] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[138] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[139] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[140] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[141] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[142] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[143] * kernel_shared_1[threadIdx_x * 72 + 47]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[162] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[163] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[164] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[165] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[166] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[167] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[168] * kernel_shared_1[threadIdx_x * 72 + 54]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[163] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[164] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[165] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[166] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[167] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[168] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[169] * kernel_shared_1[threadIdx_x * 72 + 55]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[164] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[165] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[166] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[167] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[168] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[169] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[170] * kernel_shared_1[threadIdx_x * 72 + 56]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[189] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[190] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[191] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[192] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[193] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[194] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[195] * kernel_shared_1[threadIdx_x * 72 + 63]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[190] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[191] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[192] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[193] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[194] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[195] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[196] * kernel_shared_1[threadIdx_x * 72 + 64]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[191] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[192] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[193] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[194] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[195] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[196] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[197] * kernel_shared_1[threadIdx_x * 72 + 65]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[9] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 72 + 3]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 72 + 4]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[17] * kernel_shared_1[threadIdx_x * 72 + 5]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[36] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 72 + 12]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 72 + 13]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[44] * kernel_shared_1[threadIdx_x * 72 + 14]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[63] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 72 + 21]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 72 + 22]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[71] * kernel_shared_1[threadIdx_x * 72 + 23]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[90] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[91] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[92] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[93] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[94] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[95] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[96] * kernel_shared_1[threadIdx_x * 72 + 30]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[91] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[92] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[93] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[94] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[95] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[96] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[97] * kernel_shared_1[threadIdx_x * 72 + 31]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[92] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[93] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[94] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[95] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[96] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[97] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[98] * kernel_shared_1[threadIdx_x * 72 + 32]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[117] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[118] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[119] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[120] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[121] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[122] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[123] * kernel_shared_1[threadIdx_x * 72 + 39]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[118] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[119] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[120] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[121] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[122] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[123] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[124] * kernel_shared_1[threadIdx_x * 72 + 40]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[119] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[120] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[121] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[122] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[123] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[124] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[125] * kernel_shared_1[threadIdx_x * 72 + 41]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[144] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[145] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[146] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[147] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[148] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[149] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[150] * kernel_shared_1[threadIdx_x * 72 + 48]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[145] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[146] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[147] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[148] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[149] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[150] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[151] * kernel_shared_1[threadIdx_x * 72 + 49]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[146] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[147] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[148] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[149] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[150] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[151] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[152] * kernel_shared_1[threadIdx_x * 72 + 50]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[171] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[172] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[173] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[174] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[175] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[176] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[177] * kernel_shared_1[threadIdx_x * 72 + 57]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[172] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[173] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[174] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[175] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[176] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[177] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[178] * kernel_shared_1[threadIdx_x * 72 + 58]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[173] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[174] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[175] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[176] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[177] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[178] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[179] * kernel_shared_1[threadIdx_x * 72 + 59]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[198] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[199] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[200] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[201] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[202] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[203] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[204] * kernel_shared_1[threadIdx_x * 72 + 66]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[199] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[200] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[201] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[202] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[203] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[204] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[205] * kernel_shared_1[threadIdx_x * 72 + 67]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[200] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[201] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[202] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[203] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[204] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[205] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[206] * kernel_shared_1[threadIdx_x * 72 + 68]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[18] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 72 + 6]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 72 + 7]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[26] * kernel_shared_1[threadIdx_x * 72 + 8]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[45] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 72 + 15]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 72 + 16]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[53] * kernel_shared_1[threadIdx_x * 72 + 17]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[72] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[73] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[74] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[75] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[76] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[77] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[78] * kernel_shared_1[threadIdx_x * 72 + 24]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[73] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[74] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[75] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[76] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[77] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[78] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[79] * kernel_shared_1[threadIdx_x * 72 + 25]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[74] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[75] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[76] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[77] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[78] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[79] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[80] * kernel_shared_1[threadIdx_x * 72 + 26]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[99] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[100] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[101] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[102] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[103] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[104] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[105] * kernel_shared_1[threadIdx_x * 72 + 33]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[100] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[101] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[102] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[103] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[104] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[105] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[106] * kernel_shared_1[threadIdx_x * 72 + 34]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[101] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[102] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[103] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[104] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[105] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[106] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[107] * kernel_shared_1[threadIdx_x * 72 + 35]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[126] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[127] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[128] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[129] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[130] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[131] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[132] * kernel_shared_1[threadIdx_x * 72 + 42]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[127] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[128] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[129] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[130] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[131] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[132] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[133] * kernel_shared_1[threadIdx_x * 72 + 43]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[128] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[129] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[130] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[131] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[132] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[133] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[134] * kernel_shared_1[threadIdx_x * 72 + 44]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[153] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[154] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[155] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[156] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[157] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[158] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[159] * kernel_shared_1[threadIdx_x * 72 + 51]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[154] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[155] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[156] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[157] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[158] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[159] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[160] * kernel_shared_1[threadIdx_x * 72 + 52]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[155] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[156] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[157] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[158] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[159] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[160] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[161] * kernel_shared_1[threadIdx_x * 72 + 53]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[180] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[181] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[182] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[183] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[184] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[185] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[186] * kernel_shared_1[threadIdx_x * 72 + 60]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[181] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[182] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[183] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[184] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[185] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[186] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[187] * kernel_shared_1[threadIdx_x * 72 + 61]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[182] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[183] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[184] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[185] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[186] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[187] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[188] * kernel_shared_1[threadIdx_x * 72 + 62]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[207] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[208] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[209] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[210] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[211] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[212] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[213] * kernel_shared_1[threadIdx_x * 72 + 69]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[208] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[209] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[210] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[211] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[212] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[213] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[214] * kernel_shared_1[threadIdx_x * 72 + 70]
+ conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[209] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[210] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[211] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[212] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[213] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[214] * kernel_shared_1[threadIdx_x * 72 + 71]
+ conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[215] * kernel_shared_1[threadIdx_x * 72 + 71]
+ compute_1 = T.Buffer((25088,), data=compute.data)
+ bias_1 = T.Buffer((512,), data=bias.data)
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7] = T.max(conv2d_nchw_1[0] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 1] = T.max(conv2d_nchw_1[1] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 2] = T.max(conv2d_nchw_1[2] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 3] = T.max(conv2d_nchw_1[3] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 4] = T.max(conv2d_nchw_1[4] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 5] = T.max(conv2d_nchw_1[5] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
+ compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 49 + blockIdx_x % 7 * 7 + 6] = T.max(conv2d_nchw_1[6] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x], T.float32(0))
</pre></div>
</div>
</div>
@@ -1333,7 +1223,7 @@ 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: 0.251 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 0.216 ms
</pre></div>
</div>
</div>
@@ -1364,17 +1254,17 @@ conv2d_nchw_nn_o_o_o_i, conv2d_nchw_nn_o_o_i = s[conv2d_nchw].split(conv2d_nchw_
conv2d_nchw_nn_o_o_o_o, conv2d_nchw_nn_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_nn_o_o_o_i, factor=1)
conv2d_nchw_ff_o_i, conv2d_nchw_ff_i = s[conv2d_nchw].split(conv2d_nchw_ff, factor=1)
conv2d_nchw_ff_o_o_i, conv2d_nchw_ff_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_i, factor=1)
-conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=32)
+conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=128)
conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=1)
conv2d_nchw_yy_o_i, conv2d_nchw_yy_i = s[conv2d_nchw].split(conv2d_nchw_yy, factor=1)
conv2d_nchw_yy_o_o_i, conv2d_nchw_yy_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_i, factor=1)
conv2d_nchw_yy_o_o_o_i, conv2d_nchw_yy_o_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_o_i, factor=1)
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_o_i, conv2d_nchw_xx_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_i, 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=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=16)
+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=7)
+conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=8)
conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=1)
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)
@@ -1385,14 +1275,14 @@ compute_i0_o_i, compute_i0_i = s[compute].split(compute_i0, factor=1)
compute_i0_o_o_i, compute_i0_o_i = s[compute].split(compute_i0_o_i, factor=1)
compute_i0_o_o_o, compute_i0_o_o_i = s[compute].split(compute_i0_o_o_i, factor=1)
compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=1)
-compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=32)
+compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=128)
compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=1)
compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=1)
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=7)
+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=1)
-compute_i3_o_o_o, compute_i3_o_o_i = s[compute].split(compute_i3_o_o_i, factor=1)
+compute_i3_o_o_o, compute_i3_o_o_i = s[compute].split(compute_i3_o_o_i, factor=7)
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)
kernel_shared = s.cache_read(kernel, "shared", [conv2d_nchw])
@@ -1411,14 +1301,14 @@ s[compute].bind(compute_i0_o_i_i1_o_i_fused_i2_o_i_fused_i3_o_i_fused, te.thread
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[kernel_shared].fuse(kernel_shared_ax0, kernel_shared_ax1, kernel_shared_ax2, kernel_shared_ax3)
kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[kernel_shared].split(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
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=32)
+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=128)
s[kernel_shared].bind(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused = s[pad_temp_shared].fuse(pad_temp_shared_ax0, pad_temp_shared_ax1, pad_temp_shared_ax2, pad_temp_shared_ax3)
pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused, factor=1)
s[pad_temp_shared].vectorize(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_i)
-pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_o, pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i = s[pad_temp_shared].split(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o, factor=32)
+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=128)
s[pad_temp_shared].bind(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis("threadIdx.x"))
-s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 512)
+s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "auto_unroll_max_step", 1024)
s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, "unroll_explicit", True)
CUDA source code:
@@ -1436,10 +1326,10 @@ CUDA source code:
#define int64_t long long
#define uint64_t unsigned long long
#endif
-extern "C" __global__ void __launch_bounds__(32) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
+extern "C" __global__ void __launch_bounds__(128) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
float conv2d_nchw[7];
- __shared__ float pad_temp_shared[432];
- __shared__ float kernel_shared[4608];
+ __shared__ float pad_temp_shared[216];
+ __shared__ float kernel_shared[9216];
conv2d_nchw[0] = 0.000000e+00f;
conv2d_nchw[1] = 0.000000e+00f;
conv2d_nchw[2] = 0.000000e+00f;
@@ -1447,511 +1337,597 @@ extern "C" __global__ void __launch_bounds__(32) default_function_kern
conv2d_nchw[4] = 0.000000e+00f;
conv2d_nchw[5] = 0.000000e+00f;
conv2d_nchw[6] = 0.000000e+00f;
- for (int rc_outer_outer = 0; rc_outer_outer < 32; ++rc_outer_outer) {
+ for (int rc_outer_outer = 0; rc_outer_outer < 64; ++rc_outer_outer) {
__syncthreads();
- pad_temp_shared[((int)threadIdx.x)] = (((((1 <= (((((int)threadIdx.x) % 27) / 9) + (((int)blockIdx.x) % 7))) && ((((((int)threadIdx.x) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + ((((int)threadIdx.x) / 27) * 49)) + (((((int)threadIdx.x) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
- pad_temp_shared[(((int)threadIdx.x) + 32)] = (((((1 <= ((((((int)threadIdx.x) + 5) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 5) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 32) / 27) * 49)) + ((((((int)threadIdx.x) + 5) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)thr [...]
- pad_temp_shared[(((int)threadIdx.x) + 64)] = (((((1 <= ((((((int)threadIdx.x) + 10) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 10) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 64) / 27) * 49)) + ((((((int)threadIdx.x) + 10) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int) [...]
- pad_temp_shared[(((int)threadIdx.x) + 96)] = (((((1 <= ((((((int)threadIdx.x) + 15) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 15) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 96) / 27) * 49)) + ((((((int)threadIdx.x) + 15) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int) [...]
- pad_temp_shared[(((int)threadIdx.x) + 128)] = (((((1 <= ((((((int)threadIdx.x) + 20) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 20) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 128) / 27) * 49)) + ((((((int)threadIdx.x) + 20) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((in [...]
- pad_temp_shared[(((int)threadIdx.x) + 160)] = (((((1 <= ((((((int)threadIdx.x) + 25) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 25) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 160) / 27) * 49)) + ((((((int)threadIdx.x) + 25) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((in [...]
- pad_temp_shared[(((int)threadIdx.x) + 192)] = (((((1 <= ((((((int)threadIdx.x) + 3) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 3) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 192) / 27) * 49)) + ((((((int)threadIdx.x) + 3) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)t [...]
- pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((1 <= ((((((int)threadIdx.x) + 8) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 8) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 224) / 27) * 49)) + ((((((int)threadIdx.x) + 8) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)t [...]
- pad_temp_shared[(((int)threadIdx.x) + 256)] = (((((1 <= ((((((int)threadIdx.x) + 13) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 13) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 256) / 27) * 49)) + ((((((int)threadIdx.x) + 13) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((in [...]
- pad_temp_shared[(((int)threadIdx.x) + 288)] = (((((1 <= ((((int)blockIdx.x) % 7) + (((((int)threadIdx.x) / 9) + 2) % 3))) && (((((int)blockIdx.x) % 7) + (((((int)threadIdx.x) / 9) + 2) % 3)) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 288) / 27) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + ((((((int)threadIdx.x) / 9) + 2) % 3) * 7)) + (((int)threadIdx.x) % 9) [...]
- pad_temp_shared[(((int)threadIdx.x) + 320)] = (((((1 <= ((((((int)threadIdx.x) + 23) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 23) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 5) % 9))) && (((((int)threadIdx.x) + 5) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 320) / 27) * 49)) + ((((((int)threadIdx.x) + 23) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((in [...]
- pad_temp_shared[(((int)threadIdx.x) + 352)] = (((((1 <= ((((((int)threadIdx.x) + 1) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 1) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 1) % 9))) && (((((int)threadIdx.x) + 1) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 352) / 27) * 49)) + ((((((int)threadIdx.x) + 1) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)t [...]
- pad_temp_shared[(((int)threadIdx.x) + 384)] = (((((1 <= ((((((int)threadIdx.x) + 6) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 6) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 6) % 9))) && (((((int)threadIdx.x) + 6) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 384) / 27) * 49)) + ((((((int)threadIdx.x) + 6) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)t [...]
- if (((int)threadIdx.x) < 16) {
- pad_temp_shared[(((int)threadIdx.x) + 416)] = (((((((((int)threadIdx.x) + 11) / 9) + (((int)blockIdx.x) % 7)) < 8) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[((((((rc_outer_outer * 784) + (((((int)threadIdx.x) + 416) / 27) * 49)) + (((((int)threadIdx.x) + 11) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) + 2) % 9)) - 8)] : 0.000000e+00f);
+ pad_temp_shared[((int)threadIdx.x)] = (((((1 <= (((((int)threadIdx.x) % 27) / 9) + (((int)blockIdx.x) % 7))) && ((((((int)threadIdx.x) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[((((((rc_outer_outer * 392) + ((((int)threadIdx.x) / 27) * 49)) + (((((int)threadIdx.x) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
+ if (((int)threadIdx.x) < 88) {
+ pad_temp_shared[(((int)threadIdx.x) + 128)] = (((((1 <= ((((((int)threadIdx.x) + 20) % 27) / 9) + (((int)blockIdx.x) % 7))) && (((((((int)threadIdx.x) + 20) % 27) / 9) + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) + 2) % 9))) && (((((int)threadIdx.x) + 2) % 9) < 8)) ? data[((((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 128) / 27) * 49)) + ((((((int)threadIdx.x) + 20) % 27) / 9) * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((( [...]
}
- kernel_shared[((int)threadIdx.x)] = kernel[((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x))];
- kernel_shared[(((int)threadIdx.x) + 32)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 64)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 96)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 128)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 128) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 160)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 160) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 192)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 192) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 224)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 224) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 256)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 256) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 288)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 9216)];
- kernel_shared[(((int)threadIdx.x) + 320)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 320) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 352)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 352) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 384)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 384) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 416)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 416) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 448)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 448) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 480)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 480) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 512)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 512) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 544)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 544) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 576)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 18432)];
- kernel_shared[(((int)threadIdx.x) + 608)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 608) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 640)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 640) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 672)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 672) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 704)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 704) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 736)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 736) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 768)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 768) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 800)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 800) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 832)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 832) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 864)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 27648)];
- kernel_shared[(((int)threadIdx.x) + 896)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 896) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 928)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 928) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 960)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 960) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 992)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 992) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1024)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1024) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1056)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1056) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 1088)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1088) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1120)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1120) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1152)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 36864)];
- kernel_shared[(((int)threadIdx.x) + 1184)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1184) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1216)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1216) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1248)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1248) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 1280)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1280) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1312)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1312) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1344) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 1376)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1376) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1408)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1408) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1440)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 46080)];
- kernel_shared[(((int)threadIdx.x) + 1472)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1472) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1504)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1504) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1536)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1536) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 1568)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1568) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1600)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1600) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1632)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1632) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 1664)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1664) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1696)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1696) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1728)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 55296)];
- kernel_shared[(((int)threadIdx.x) + 1760)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1760) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1792)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1792) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1824)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1824) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 1856)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1856) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1888)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1888) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1920)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1920) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 1952)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1952) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 1984)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 1984) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2016)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 64512)];
- kernel_shared[(((int)threadIdx.x) + 2048)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2048) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2080)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2080) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2112)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2112) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 2144)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2144) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2176)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2176) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2208)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2208) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 2240)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2240) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2272)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2272) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2304)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 73728)];
- kernel_shared[(((int)threadIdx.x) + 2336)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2336) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2368)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2368) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2400)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2400) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 2432)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2432) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2464)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2464) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2496)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2496) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 2528)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2528) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2560)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2560) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2592)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 82944)];
- kernel_shared[(((int)threadIdx.x) + 2624)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2624) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2656)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2656) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2688)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2688) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 2720)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2720) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2752)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2752) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2784)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2784) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 2816)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2816) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2848)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2848) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2880)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 92160)];
- kernel_shared[(((int)threadIdx.x) + 2912)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2912) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2944)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2944) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 2976)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 2976) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 3008)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3008) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3040)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3040) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3072)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3072) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 3104)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3104) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3136)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3136) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3168)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 101376)];
- kernel_shared[(((int)threadIdx.x) + 3200)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3200) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3232)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3232) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3264)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3264) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 3296)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3296) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3328)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3328) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3360)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3360) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 3392)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3392) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3424)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3424) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3456)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 110592)];
- kernel_shared[(((int)threadIdx.x) + 3488)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3488) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3520)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3520) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3552)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3552) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 3584)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3584) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3616)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3616) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3648)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3648) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 3680)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3680) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3712)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3712) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3744)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 119808)];
- kernel_shared[(((int)threadIdx.x) + 3776)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3776) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3808)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3808) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3840)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3840) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 3872)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3872) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3904)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3904) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 3936)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3936) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 3968)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 3968) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4000)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4000) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4032)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 129024)];
- kernel_shared[(((int)threadIdx.x) + 4064)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4064) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4096)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4096) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4128)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4128) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 4160)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4160) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4192)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4192) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4224)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4224) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 4256)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4256) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4288)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4288) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4320)] = kernel[(((((((int)blockIdx.x) / 7) * 147456) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 138240)];
- kernel_shared[(((int)threadIdx.x) + 4352)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4352) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 32) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4384)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4384) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 64) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4416)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4416) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 96)];
- kernel_shared[(((int)threadIdx.x) + 4448)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4448) / 144) * 4608)) + (rc_outer_outer * 144)) + ((((((int)threadIdx.x) + 128) % 144) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4480)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4480) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 16) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4512)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4512) / 144) * 4608)) + (rc_outer_outer * 144)) + ((int)threadIdx.x)) + 48)];
- kernel_shared[(((int)threadIdx.x) + 4544)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4544) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 80) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))];
- kernel_shared[(((int)threadIdx.x) + 4576)] = kernel[((((((((int)blockIdx.x) / 7) * 147456) + (((((int)threadIdx.x) + 4576) / 144) * 4608)) + (rc_outer_outer * 144)) + (((((int)threadIdx.x) + 112) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+ kernel_shared[((int)threadIdx.x)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 128)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 128) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 256)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 256) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 384)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 384) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 512)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 512) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 640)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 640) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 768)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 768) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 896)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 896) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1024)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1024) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1152)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 73728)];
+ kernel_shared[(((int)threadIdx.x) + 1280)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1280) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1408)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1408) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1536)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1536) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1664)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1664) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 1792)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1792) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 1920)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1920) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2048)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2048) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2176)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2176) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2304)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 147456)];
+ kernel_shared[(((int)threadIdx.x) + 2432)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2432) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2560)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2560) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2688)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2688) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 2816)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2816) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 2944)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2944) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3072)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3072) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3200)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3200) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3328)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3328) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3456)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 221184)];
+ kernel_shared[(((int)threadIdx.x) + 3584)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3584) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3712)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3712) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3840)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3840) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 3968)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3968) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 4096)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4096) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4224)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4224) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4352)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4352) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4480)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4480) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4608)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 294912)];
+ kernel_shared[(((int)threadIdx.x) + 4736)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4736) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4864)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4864) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 4992)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 4992) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5120)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5120) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 5248)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5248) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5376)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5376) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5504)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5504) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5632)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5632) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 5760)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 368640)];
+ kernel_shared[(((int)threadIdx.x) + 5888)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 5888) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6016)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6016) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6144)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6144) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6272)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6272) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 6400)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6400) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6528)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6528) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6656)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6656) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6784)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 6784) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 6912)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 442368)];
+ kernel_shared[(((int)threadIdx.x) + 7040)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7040) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7168)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7168) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7296)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7296) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7424)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7424) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 7552)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7552) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7680)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7680) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7808)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7808) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 7936)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 7936) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8064)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 72) * 4608)) + (rc_outer_outer * 72)) + (((int)threadIdx.x) % 72)) + 516096)];
+ kernel_shared[(((int)threadIdx.x) + 8192)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8192) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 56) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 2) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8320)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8320) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 40) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 4) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8448)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8448) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 24) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 6) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8576)] = kernel[(((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8576) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((int)threadIdx.x) + 8) % 72))];
+ kernel_shared[(((int)threadIdx.x) + 8704)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8704) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 64) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 1) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8832)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8832) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 48) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 3) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 8960)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 8960) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 32) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 5) % 9))];
+ kernel_shared[(((int)threadIdx.x) + 9088)] = kernel[((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 9088) / 72) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 72) / 9) * 9)) + ((((int)threadIdx.x) + 7) % 9))];
__syncthreads();
- for (int ry_outer_inner = 0; ry_outer_inner < 3; ++ry_outer_inner) {
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(ry_outer_inner * 9)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 27)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 28)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 29)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 54)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 55)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 56)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 81)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 82)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 108)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 109)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 110)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 135)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 136)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 137)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 162)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 163)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 216)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 217)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 218)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 243)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 244)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 270)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 271)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 272)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 297)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 298)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 299)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 324)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 325)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 351)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 352)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 353)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 405)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 406)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((ry_outer_inner * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 1)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 28)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 29)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 30)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 55)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 56)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 57)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 82)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 109)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 110)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 111)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 136)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 137)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 138)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 163)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 217)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 218)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 219)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 244)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 271)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 272)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 273)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 298)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 299)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 300)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 325)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 352)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 353)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 354)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 406)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[((ry_outer_inner * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 2)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 29)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 30)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 31)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 56)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 57)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 58)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 83)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 110)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 111)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 112)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 137)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 138)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 139)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 164)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 218)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 219)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 220)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 245)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 272)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 273)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 274)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 299)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 300)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 301)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 326)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 353)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 354)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 355)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 407)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[((ry_outer_inner * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 3)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 30)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 31)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 32)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 57)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 58)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 59)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 84)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 111)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 112)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 113)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 138)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 139)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 140)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 165)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 219)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 220)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 221)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 246)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 273)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 274)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 275)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 300)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 301)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 302)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 327)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 354)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 355)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 356)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 408)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[((ry_outer_inner * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 4)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 31)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 32)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 33)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 58)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 59)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 60)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 85)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 112)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 113)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 114)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 139)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 140)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 141)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 166)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 220)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 221)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 222)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 247)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 274)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 275)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 276)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 301)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 302)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 303)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 328)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 355)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 356)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 357)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 409)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[((ry_outer_inner * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 5)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 32)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 33)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 34)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 59)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 60)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 61)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 86)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 88)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 113)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 114)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 115)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 140)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 141)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 142)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 167)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 169)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 221)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 222)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 223)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 248)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 250)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 275)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 276)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 277)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 302)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 303)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 304)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 329)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 331)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 356)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 357)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 358)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 410)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[((ry_outer_inner * 9) + 412)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 6)] * kernel_shared[((((int)threadIdx.x) * 144) + (ry_outer_inner * 3))]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 1)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 2)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 33)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 9)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 34)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 10)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 35)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 11)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 60)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 18)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 61)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 19)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 62)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 20)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 87)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 27)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 88)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 28)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 89)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 29)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 114)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 36)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 115)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 37)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 116)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 38)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 141)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 45)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 142)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 46)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 143)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 47)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 168)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 54)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 169)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 55)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 170)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 56)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 63)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 64)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 65)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 222)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 72)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 223)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 73)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 224)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 74)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 249)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 81)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 250)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 82)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 251)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 83)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 276)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 90)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 277)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 91)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 278)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 92)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 303)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 99)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 304)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 100)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 305)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 101)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 330)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 108)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 331)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 109)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 332)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 110)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 357)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 117)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 358)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 118)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 359)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 119)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 126)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 127)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 128)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 411)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 135)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 412)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 136)]));
- conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[((ry_outer_inner * 9) + 413)] * kernel_shared[(((((int)threadIdx.x) * 144) + (ry_outer_inner * 3)) + 137)]));
- }
- }
- for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
- compute[(((((((int)blockIdx.x) / 7) * 1568) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + i3_inner)] = max((conv2d_nchw[i3_inner] + bias[(((((int)blockIdx.x) / 7) * 32) + ((int)threadIdx.x))]), 0.000000e+00f);
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[0] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[1] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[2] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[3] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[4] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[5] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[6] * kernel_shared[(((int)threadIdx.x) * 72)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[1] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 72) + 1)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[8] * kernel_shared[((((int)threadIdx.x) * 72) + 2)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[27] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 72) + 9)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 72) + 10)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[35] * kernel_shared[((((int)threadIdx.x) * 72) + 11)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[54] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 72) + 18)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 72) + 19)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[62] * kernel_shared[((((int)threadIdx.x) * 72) + 20)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[81] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[82] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[83] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[84] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[85] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[86] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[87] * kernel_shared[((((int)threadIdx.x) * 72) + 27)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[82] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[83] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[84] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[85] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[86] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[87] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[88] * kernel_shared[((((int)threadIdx.x) * 72) + 28)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[83] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[84] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[85] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[86] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[87] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[88] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[89] * kernel_shared[((((int)threadIdx.x) * 72) + 29)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[108] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[109] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[110] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[111] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[112] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[113] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[114] * kernel_shared[((((int)threadIdx.x) * 72) + 36)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[109] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[110] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[111] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[112] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[113] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[114] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[115] * kernel_shared[((((int)threadIdx.x) * 72) + 37)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[110] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[111] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[112] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[113] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[114] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[115] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[116] * kernel_shared[((((int)threadIdx.x) * 72) + 38)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[135] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[136] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[137] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[138] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[139] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[140] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[141] * kernel_shared[((((int)threadIdx.x) * 72) + 45)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[136] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[137] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[138] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[139] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[140] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[141] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[142] * kernel_shared[((((int)threadIdx.x) * 72) + 46)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[137] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[138] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[139] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[140] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[141] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[142] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[143] * kernel_shared[((((int)threadIdx.x) * 72) + 47)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[162] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[163] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[164] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[165] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[166] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[167] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[168] * kernel_shared[((((int)threadIdx.x) * 72) + 54)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[163] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[164] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[165] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[166] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[167] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[168] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[169] * kernel_shared[((((int)threadIdx.x) * 72) + 55)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[164] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[165] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[166] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[167] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[168] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[169] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[170] * kernel_shared[((((int)threadIdx.x) * 72) + 56)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[189] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[190] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[191] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[192] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[193] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[194] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[195] * kernel_shared[((((int)threadIdx.x) * 72) + 63)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[190] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[191] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[192] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[193] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[194] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[195] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[196] * kernel_shared[((((int)threadIdx.x) * 72) + 64)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[191] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[192] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[193] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[194] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[195] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[196] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[197] * kernel_shared[((((int)threadIdx.x) * 72) + 65)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[9] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 72) + 3)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 72) + 4)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[17] * kernel_shared[((((int)threadIdx.x) * 72) + 5)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[36] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 72) + 12)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 72) + 13)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[44] * kernel_shared[((((int)threadIdx.x) * 72) + 14)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[63] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 72) + 21)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 72) + 22)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[71] * kernel_shared[((((int)threadIdx.x) * 72) + 23)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[90] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[91] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[92] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[93] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[94] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[95] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[96] * kernel_shared[((((int)threadIdx.x) * 72) + 30)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[91] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[92] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[93] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[94] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[95] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[96] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[97] * kernel_shared[((((int)threadIdx.x) * 72) + 31)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[92] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[93] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[94] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[95] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[96] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[97] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[98] * kernel_shared[((((int)threadIdx.x) * 72) + 32)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[117] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[118] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[119] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[120] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[121] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[122] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[123] * kernel_shared[((((int)threadIdx.x) * 72) + 39)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[118] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[119] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[120] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[121] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[122] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[123] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[124] * kernel_shared[((((int)threadIdx.x) * 72) + 40)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[119] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[120] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[121] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[122] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[123] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[124] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[125] * kernel_shared[((((int)threadIdx.x) * 72) + 41)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[144] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[145] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[146] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[147] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[148] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[149] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[150] * kernel_shared[((((int)threadIdx.x) * 72) + 48)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[145] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[146] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[147] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[148] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[149] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[150] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[151] * kernel_shared[((((int)threadIdx.x) * 72) + 49)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[146] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[147] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[148] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[149] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[150] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[151] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[152] * kernel_shared[((((int)threadIdx.x) * 72) + 50)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[171] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[172] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[173] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[174] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[175] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[176] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[177] * kernel_shared[((((int)threadIdx.x) * 72) + 57)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[172] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[173] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[174] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[175] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[176] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[177] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[178] * kernel_shared[((((int)threadIdx.x) * 72) + 58)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[173] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[174] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[175] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[176] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[177] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[178] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[179] * kernel_shared[((((int)threadIdx.x) * 72) + 59)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[198] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[199] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[200] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[201] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[202] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[203] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[204] * kernel_shared[((((int)threadIdx.x) * 72) + 66)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[199] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[200] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[201] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[202] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[203] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[204] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[205] * kernel_shared[((((int)threadIdx.x) * 72) + 67)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[200] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[201] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[202] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[203] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[204] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[205] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[206] * kernel_shared[((((int)threadIdx.x) * 72) + 68)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[18] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 72) + 6)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 72) + 7)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[26] * kernel_shared[((((int)threadIdx.x) * 72) + 8)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[45] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 72) + 15)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 72) + 16)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[53] * kernel_shared[((((int)threadIdx.x) * 72) + 17)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[72] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[73] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[74] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[75] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[76] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[77] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[78] * kernel_shared[((((int)threadIdx.x) * 72) + 24)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[73] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[74] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[75] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[76] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[77] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[78] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[79] * kernel_shared[((((int)threadIdx.x) * 72) + 25)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[74] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[75] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[76] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[77] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[78] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[79] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[80] * kernel_shared[((((int)threadIdx.x) * 72) + 26)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[99] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[100] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[101] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[102] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[103] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[104] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[105] * kernel_shared[((((int)threadIdx.x) * 72) + 33)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[100] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[101] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[102] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[103] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[104] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[105] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[106] * kernel_shared[((((int)threadIdx.x) * 72) + 34)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[101] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[102] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[103] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[104] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[105] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[106] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[107] * kernel_shared[((((int)threadIdx.x) * 72) + 35)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[126] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[127] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[128] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[129] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[130] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[131] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[132] * kernel_shared[((((int)threadIdx.x) * 72) + 42)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[127] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[128] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[129] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[130] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[131] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[132] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[133] * kernel_shared[((((int)threadIdx.x) * 72) + 43)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[128] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[129] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[130] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[131] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[132] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[133] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[134] * kernel_shared[((((int)threadIdx.x) * 72) + 44)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[153] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[154] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[155] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[156] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[157] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[158] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[159] * kernel_shared[((((int)threadIdx.x) * 72) + 51)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[154] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[155] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[156] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[157] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[158] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[159] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[160] * kernel_shared[((((int)threadIdx.x) * 72) + 52)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[155] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[156] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[157] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[158] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[159] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[160] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[161] * kernel_shared[((((int)threadIdx.x) * 72) + 53)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[180] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[181] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[182] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[183] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[184] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[185] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[186] * kernel_shared[((((int)threadIdx.x) * 72) + 60)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[181] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[182] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[183] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[184] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[185] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[186] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[187] * kernel_shared[((((int)threadIdx.x) * 72) + 61)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[182] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[183] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[184] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[185] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[186] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[187] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[188] * kernel_shared[((((int)threadIdx.x) * 72) + 62)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[207] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[208] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[209] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[210] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[211] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[212] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[213] * kernel_shared[((((int)threadIdx.x) * 72) + 69)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[208] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[209] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[210] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[211] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[212] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[213] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[214] * kernel_shared[((((int)threadIdx.x) * 72) + 70)]));
+ conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[209] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[210] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[211] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[212] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[213] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[214] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
+ conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[215] * kernel_shared[((((int)threadIdx.x) * 72) + 71)]));
}
+ compute[((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7))] = max((conv2d_nchw[0] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 1)] = max((conv2d_nchw[1] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 2)] = max((conv2d_nchw[2] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 3)] = max((conv2d_nchw[3] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 4)] = max((conv2d_nchw[4] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 5)] = max((conv2d_nchw[5] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
+ compute[(((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 49)) + ((((int)blockIdx.x) % 7) * 7)) + 6)] = max((conv2d_nchw[6] + bias[(((((int)blockIdx.x) / 7) * 128) + ((int)threadIdx.x))]), 0.000000e+00f);
}
</pre></div>
</div>
@@ -1987,7 +1963,7 @@ 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> ( 5 minutes 42.565 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 5 minutes 35.291 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 110b21f8f1..5a84e1f140 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
@@ -916,7 +916,7 @@ 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)
- 7.8563 7.8537 7.8646 7.8505 0.0060
+ 7.8773 7.8774 7.8840 7.8704 0.0055
</pre></div>
</div>
</div>
@@ -938,7 +938,7 @@ 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 5.839 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 6.161 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 8ca35a11c5..b7c73d7dbd 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
@@ -935,7 +935,7 @@ 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)
- 758.6698 757.9524 760.1585 757.8985 1.0529
+ 749.2774 750.8660 750.9075 746.0588 2.2759
</pre></div>
</div>
</div>
@@ -957,7 +957,7 @@ 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 39.693 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 39.443 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 d6bcd76d6c..047492836d 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
@@ -632,26 +632,74 @@ class Module:
@T.prim_func
def main(placeholder: T.Buffer((128, 256), "float32"), placeholder_1: T.Buffer((4916, 16, 1), "float32"), placeholder_2: T.Buffer((4916,), "int32"), placeholder_3: T.Buffer((33,), "int32"), placeholder_4: T.Buffer((128, 512), "float32"), compute: T.Buffer((128, 512), "float32")):
T.func_attr({"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True})
- for i0_outer_i1_outer_fused in T.parallel(64):
- compute_1 = T.allocate([1024], "float32", "global")
- compute_2 = T.Buffer((1024,), data=compute_1)
+ for i0_outer_i1_outer_fused in T.parallel(16):
+ compute_1 = T.allocate([4096], "float32", "global")
+ compute_2 = T.Buffer((4096,), data=compute_1)
for i_outer_inner, nb_j_inner in T.grid(16, 2):
- for i_inner_init, j_init in T.grid(2, 16):
- compute_2[i_outer_inner * 64 + i_inner_init * 32 + nb_j_inner * 16 + j_init] = T.float32(0)
- for elem_idx, i_inner, j in T.grid(T.let(cse_var_1, i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner, placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]), 2, 16):
- cse_var_1 = T.var("int32")
+ for i_inner_init in range(8):
+ cse_var_1: T.int32 = i_outer_inner * 256 + 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)
+ compute_2[cse_var_1 + 3] = T.float32(0)
+ compute_2[cse_var_1 + 4] = T.float32(0)
+ compute_2[cse_var_1 + 5] = T.float32(0)
+ compute_2[cse_var_1 + 6] = T.float32(0)
+ compute_2[cse_var_1 + 7] = T.float32(0)
+ compute_2[cse_var_1 + 8] = T.float32(0)
+ compute_2[cse_var_1 + 9] = T.float32(0)
+ compute_2[cse_var_1 + 10] = T.float32(0)
+ compute_2[cse_var_1 + 11] = T.float32(0)
+ compute_2[cse_var_1 + 12] = T.float32(0)
+ 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(cse_var_2, i0_outer_i1_outer_fused * 2 + nb_j_inner, placeholder_5[cse_var_2 + 1] - placeholder_5[cse_var_2]), 8):
+ cse_var_2 = T.var("int32")
placeholder_5 = T.Buffer((33,), "int32", data=placeholder_3.data)
- cse_var_3: T.int32 = i0_outer_i1_outer_fused % 16 * 2 + nb_j_inner
- cse_var_2: T.int32 = i_outer_inner * 64 + i_inner * 32 + nb_j_inner * 16 + j
+ cse_var_21: T.int32 = elem_idx * 16
+ cse_var_20: T.int32 = i0_outer_i1_outer_fused * 2 + nb_j_inner
+ cse_var_19: T.int32 = i_outer_inner * 2048 + i_inner * 256
+ cse_var_18: T.int32 = i_outer_inner * 256 + i_inner * 32 + nb_j_inner * 16
+ cse_var_17: T.int32 = cse_var_18 + 9
+ cse_var_16: T.int32 = cse_var_18 + 8
+ cse_var_15: T.int32 = cse_var_18 + 7
+ cse_var_14: T.int32 = cse_var_18 + 6
+ cse_var_13: T.int32 = cse_var_18 + 5
+ cse_var_12: T.int32 = cse_var_18 + 4
+ cse_var_11: T.int32 = cse_var_18 + 3
+ cse_var_10: T.int32 = cse_var_18 + 2
+ cse_var_9: T.int32 = cse_var_18 + 15
+ cse_var_8: T.int32 = cse_var_18 + 14
+ cse_var_7: T.int32 = cse_var_18 + 13
+ cse_var_6: T.int32 = cse_var_18 + 12
+ cse_var_5: T.int32 = cse_var_18 + 11
+ cse_var_4: T.int32 = cse_var_18 + 10
+ cse_var_3: T.int32 = cse_var_18 + 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_2] = compute_2[cse_var_2] + placeholder_6[placeholder_5[cse_var_3] * 16 + elem_idx * 16 + j] * T.max(placeholder_7[i0_outer_i1_outer_fused // 16 * 8192 + i_outer_inner * 512 + i_inner * 256 + placeholder_8[placeholder_5[cse_var_3] + elem_idx]], T.float32(0))
- for i0_inner in range(32):
- cse_var_4: T.int32 = i0_outer_i1_outer_fused // 16 * 16384 + i0_inner * 512 + i0_outer_i1_outer_fused % 16 * 32
+ compute_2[cse_var_18] = compute_2[cse_var_18] + placeholder_6[placeholder_5[cse_var_20] * 16 + cse_var_21] * T.max(placeholder_7[cse_var_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + 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_19 + placeholder_8[placeholder_5[cse_var_20] + elem_idx]], T.float32(0))
+ for i0_inner in range(128):
+ cse_var_22: T.int32 = i0_inner * 512 + i0_outer_i1_outer_fused * 32
compute_3 = T.Buffer((65536,), data=compute.data)
placeholder_5 = T.Buffer((65536,), data=placeholder_4.data)
- compute_3[cse_var_4:cse_var_4 + 32] = T.max(compute_2[i0_inner * 32:i0_inner * 32 + 32] + placeholder_5[cse_var_4:cse_var_4 + 32], T.Broadcast(T.float32(0), 32))
+ compute_3[cse_var_22:cse_var_22 + 32] = T.max(compute_2[i0_inner * 32:i0_inner * 32 + 32] + placeholder_5[cse_var_22:cse_var_22 + 32], T.Broadcast(T.float32(0), 32))
</pre></div>
</div>
</div>
@@ -685,7 +733,7 @@ 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: 2.117 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 1.868 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 8ed446852f..fac7c9c9be 100644
--- a/docs/how_to/tune_with_autotvm/sg_execution_times.html
+++ b/docs/how_to/tune_with_autotvm/sg_execution_times.html
@@ -340,7 +340,7 @@
<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>00:43.161</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
+<p><strong>00:42.208</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -349,11 +349,11 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="tune_conv2d_cuda.html#sphx-glr-how-to-tune-with-autotvm-tune-conv2d-cuda-py"><span class="std std-ref">Tuning High Performance Convolution on NVIDIA GPUs</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_conv2d_cuda.py</span></code>)</p></td>
-<td><p>00:43.129</p></td>
+<td><p>00:42.174</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="tune_relay_x86.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-x86-py"><span class="std std-ref">Auto-tuning a Convolutional Network for x86 CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_x86.py</span></code>)</p></td>
-<td><p>00:00.019</p></td>
+<td><p>00:00.021</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="tune_relay_cuda.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-cuda-py"><span class="std std-ref">Auto-tuning a Convolutional Network for NVIDIA GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_cuda.py</span></code>)</p></td>
@@ -361,7 +361,7 @@
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="tune_relay_arm.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-arm-py"><span class="std std-ref">Auto-tuning a Convolutional Network for ARM CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_arm.py</span></code>)</p></td>
-<td><p>00:00.005</p></td>
+<td><p>00:00.004</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="tune_relay_mobile_gpu.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-mobile-gpu-py"><span class="std std-ref">Auto-tuning a Convolutional Network for Mobile GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_mobile_gpu.py</span></code>)</p></td>
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 bd0ee276ab..b2828c9786 100644
--- a/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
+++ b/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
@@ -690,9 +690,8 @@ 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, 4, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 8, 64]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8495099
-No: 2 GFLOPS: 90.01/90.01 result: MeasureResult(costs=(0.002572052,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.9165120124816895, timestamp=1674803118.8732517) [('tile_f', [-1, 8, 4, 8]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,41072
-No: 3 GFLOPS: 0.00/90.01 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 128, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2967167
+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
@@ -814,8 +813,8 @@ 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, 8, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 16]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7487659
-No: 4 GFLOPS: 0.00/90.01 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 2, 32]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 32, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9986432
+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
@@ -937,8 +936,26 @@ 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, 2, 64]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9392665
-No: 5 GFLOPS: 0.00/90.01 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 64, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 64, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1637115
+No: 4 GFLOPS: 0.00/0.00 result: Traceback (most recent call last):
+ File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 142, in build
+ res = future.result()
+ File "/usr/lib/python3.7/concurrent/futures/_base.py", line 435, in result
+ return self.__get_result()
+ File "/usr/lib/python3.7/concurrent/futures/_base.py", line 384, in __get_result
+ raise self._exception
+ File "/usr/lib/python3.7/concurrent/futures/thread.py", line 57, in run
+ result = self.fn(*self.args, **self.kwargs)
+ File "/workspace/python/tvm/contrib/popen_pool.py", line 432, in <lambda>
+ worker = lambda *args: self._worker_run(*args)
+ File "/workspace/python/tvm/contrib/popen_pool.py", line 401, in _worker_run
+ return proc.recv()
+ File "/workspace/python/tvm/contrib/popen_pool.py", line 309, in recv
+ raise TimeoutError()
+TimeoutError
+
+ [('tile_f', [-1, 32, 2, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 4]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8784933
+No: 5 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
@@ -1060,8 +1077,12 @@ 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, 16, 1]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 512]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2707796
-No: 6 GFLOPS: 0.00/90.01 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 8, 2, 16]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 16, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6792673
+No: 6 GFLOPS: 38.05/38.05 result: MeasureResult(costs=(0.0060844850526315785,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.9628829956054688, timestamp=1674812669.290498) [('tile_f', [-1, 4, 1, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4942797
+No: 7 GFLOPS: 279.09/279.09 result: MeasureResult(costs=(0.0008294822422680412,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.5810739994049072, timestamp=1674812671.22481) [('tile_f', [-1, 1, 4, 4]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 8, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,399855
+No: 8 GFLOPS: 63.50/279.09 result: MeasureResult(costs=(0.0036458056071428566,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.4576036930084229, timestamp=1674812671.9696527) [('tile_f', [-1, 16, 16, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4877438
+No: 9 GFLOPS: 5.24/279.09 result: MeasureResult(costs=(0.044172023750000004,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.1857759952545166, timestamp=1674812674.3203592) [('tile_f', [-1, 8, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5691650
+No: 10 GFLOPS: 0.00/279.09 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
@@ -1183,27 +1204,8 @@ 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, 4, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 8]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2806473
-No: 7 GFLOPS: 164.23/164.23 result: MeasureResult(costs=(0.0014096323918918918,), error_no=MeasureErrorNo.NO_ERROR, all_cost=10.716526985168457, timestamp=1674803132.9284441) [('tile_f', [-1, 2, 2, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 8, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8844285
-No: 8 GFLOPS: 0.00/164.23 result: Traceback (most recent call last):
- File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 142, in build
- res = future.result()
- File "/usr/lib/python3.7/concurrent/futures/_base.py", line 435, in result
- return self.__get_result()
- File "/usr/lib/python3.7/concurrent/futures/_base.py", line 384, in __get_result
- raise self._exception
- File "/usr/lib/python3.7/concurrent/futures/thread.py", line 57, in run
- result = self.fn(*self.args, **self.kwargs)
- File "/workspace/python/tvm/contrib/popen_pool.py", line 432, in <lambda>
- worker = lambda *args: self._worker_run(*args)
- File "/workspace/python/tvm/contrib/popen_pool.py", line 401, in _worker_run
- return proc.recv()
- File "/workspace/python/tvm/contrib/popen_pool.py", line 309, in recv
- raise TimeoutError()
-TimeoutError
-
- [('tile_f', [-1, 4, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 64, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9704202
-No: 9 GFLOPS: 0.00/164.23 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 32, 4, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 16, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5629217
+No: 11 GFLOPS: 0.00/279.09 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
@@ -1325,8 +1327,8 @@ 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, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 128, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7838471
-No: 10 GFLOPS: 0.00/164.23 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 1, 512]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 8]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2808079
+No: 12 GFLOPS: 0.00/279.09 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
@@ -1448,8 +1450,8 @@ 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, 512, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 32]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,339294
-No: 11 GFLOPS: 0.00/164.23 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 4, 4, 32]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 64, 4]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5899056
+No: 13 GFLOPS: 0.00/279.09 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
@@ -1571,9 +1573,8 @@ 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, 256]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 16, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10117577
-No: 12 GFLOPS: 316.20/316.20 result: MeasureResult(costs=(0.0007321365147928994,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7440016269683838, timestamp=1674803135.8509939) [('tile_f', [-1, 4, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4070442
-No: 13 GFLOPS: 0.00/316.20 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 32, 1, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 512, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,5066825
+No: 14 GFLOPS: 0.00/279.09 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
@@ -1695,9 +1696,8 @@ 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, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 32, 8]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7471640
-No: 14 GFLOPS: 7.72/316.20 result: MeasureResult(costs=(0.02997332,), error_no=MeasureErrorNo.NO_ERROR, all_cost=5.854086637496948, timestamp=1674803141.9115317) [('tile_f', [-1, 4, 2, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 4]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2979825
-No: 15 GFLOPS: 0.00/316.20 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 64, 2, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 64]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1320730
+No: 15 GFLOPS: 0.00/279.09 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
@@ -1819,11 +1819,8 @@ 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, 8, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 256]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10253255
-No: 16 GFLOPS: 111.35/316.20 result: MeasureResult(costs=(0.0020790924415584416,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.28901743888855, timestamp=1674803142.9225328) [('tile_f', [-1, 2, 4, 8]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7074250
-No: 17 GFLOPS: 58.21/316.20 result: MeasureResult(costs=(0.003976914307692308,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2880961894989014, timestamp=1674803144.4809651) [('tile_f', [-1, 4, 16, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 16]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1668736
-No: 18 GFLOPS: 18.61/316.20 result: MeasureResult(costs=(0.012438851222222222,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.455913782119751, timestamp=1674803145.2864993) [('tile_f', [-1, 8, 1, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 16, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6210218
-No: 19 GFLOPS: 0.00/316.20 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 2, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 1, 256]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2508364
+No: 16 GFLOPS: 0.00/279.09 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
@@ -1945,8 +1942,11 @@ 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, 2, 4]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 16, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9428211
-No: 20 GFLOPS: 0.00/316.20 result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 1, 64, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 8, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2208083
+No: 17 GFLOPS: 98.96/279.09 result: MeasureResult(costs=(0.002339420823529412,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.632169485092163, timestamp=1674812679.9650273) [('tile_f', [-1, 1, 2, 32]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 1, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,630
+No: 18 GFLOPS: 71.38/279.09 result: MeasureResult(costs=(0.003243123677419355,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.0503921508789062, timestamp=1674812680.7025213) [('tile_f', [-1, 4, 32, 4]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8947532
+No: 19 GFLOPS: 22.79/279.09 result: MeasureResult(costs=(0.010158083299999999,), error_no=MeasureErrorNo.NO_ERROR, all_cost=4.957541227340698, timestamp=1674812681.477137) [('tile_f', [-1, 2, 4, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2369576
+No: 20 GFLOPS: 0.00/279.09 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
@@ -2068,7 +2068,7 @@ 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, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 256]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8705647
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel [('tile_f', [-1, 2, 256, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6425153
</pre></div>
</div>
<p>Finally we can inspect the best config from log file, check correctness,
@@ -2107,9 +2107,9 @@ 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, 1, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4070442
+[('tile_f', [-1, 1, 4, 4]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 8, 1]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,399855
Finish loading 20 records
-Time cost of this operator: 0.001051
+Time cost of this operator: 0.001200
</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 b32cdd6524..a6bd9ca53c 100644
--- a/docs/how_to/work_with_microtvm/micro_autotune.html
+++ b/docs/how_to/work_with_microtvm/micro_autotune.html
@@ -647,10 +647,10 @@ 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 311.0 98.701 (1, 2, 10, 10, 3) 2 1 [311.0]
-tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.116 0.989 (1, 6, 10, 10) 1 1 [3.116]
-tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.977 0.31 (1, 1, 10, 10, 3) 1 1 [0.977]
-Total_time - 315.092 - - - - -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 311.5 98.71 (1, 2, 10, 10, 3) 2 1 [311.5]
+tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 3.108 0.985 (1, 6, 10, 10) 1 1 [3.108]
+tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.963 0.305 (1, 1, 10, 10, 3) 1 1 [0.963]
+Total_time - 315.571 - - - - -
</pre></div>
</div>
</div>
@@ -702,10 +702,10 @@ 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 103.2 97.552 (1, 6, 10, 10, 1) 2 1 [103.2]
-tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.75 1.654 (1, 6, 10, 10) 1 1 [1.75]
-tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.84 0.794 (1, 3, 10, 10, 1) 1 1 [0.84]
-Total_time - 105.79 - - - - -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc tvmgen_default_fused_nn_contrib_conv2d_NCHWc 100.5 97.322 (1, 6, 10, 10, 1) 2 1 [100.5]
+tvmgen_default_fused_layout_transform_1 tvmgen_default_fused_layout_transform_1 1.807 1.75 (1, 6, 10, 10) 1 1 [1.807]
+tvmgen_default_fused_layout_transform tvmgen_default_fused_layout_transform 0.959 0.929 (1, 1, 10, 10, 3) 1 1 [0.959]
+Total_time - 103.266 - - - - -
</pre></div>
</div>
<div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-work-with-microtvm-micro-autotune-py">
diff --git a/docs/how_to/work_with_microtvm/micro_pytorch.html b/docs/how_to/work_with_microtvm/micro_pytorch.html
index e8771a72f7..9bdaefe979 100644
--- a/docs/how_to/work_with_microtvm/micro_pytorch.html
+++ b/docs/how_to/work_with_microtvm/micro_pytorch.html
@@ -454,7 +454,7 @@ 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]
-100%|##########| 3.42M/3.42M [00:00<00:00, 60.7MB/s]
+100%|##########| 3.42M/3.42M [00:00<00:00, 55.7MB/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.
@@ -578,7 +578,7 @@ 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 9.814 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes 10.103 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 819e6f4524..12931170b1 100644
--- a/docs/how_to/work_with_microtvm/micro_train.html
+++ b/docs/how_to/work_with_microtvm/micro_train.html
@@ -524,7 +524,7 @@ 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/tmp6krshqz2/images/random'
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>'/tmp/tmpllbfhusq/images/random'
</pre></div>
</div>
</div>
@@ -584,8 +584,8 @@ 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="[0.0, 1.0], [1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0]" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>/tmp/tmp6krshqz2/images/target contains 8144 images
-/tmp/tmp6krshqz2/images/random contains 5000 images
+<img src="../../_images/sphx_glr_micro_train_001.png" srcset="../../_images/sphx_glr_micro_train_001.png" alt="[0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [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/tmpllbfhusq/images/target contains 8144 images
+/tmp/tmpllbfhusq/images/random contains 5000 images
</pre></div>
</div>
</div>
@@ -697,13 +697,13 @@ 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 - 47s - loss: 0.2294 - accuracy: 0.9215 - val_loss: 0.1200 - val_accuracy: 0.9596 - 47s/epoch - 144ms/step
+328/328 - 47s - loss: 0.2308 - accuracy: 0.9220 - val_loss: 0.1662 - val_accuracy: 0.9437 - 47s/epoch - 143ms/step
Epoch 2/3
-328/328 - 44s - loss: 0.0962 - accuracy: 0.9643 - val_loss: 0.1094 - val_accuracy: 0.9626 - 44s/epoch - 133ms/step
+328/328 - 43s - loss: 0.1072 - accuracy: 0.9613 - val_loss: 0.1142 - val_accuracy: 0.9592 - 43s/epoch - 132ms/step
Epoch 3/3
-328/328 - 43s - loss: 0.0663 - accuracy: 0.9743 - val_loss: 0.2085 - val_accuracy: 0.9415 - 43s/epoch - 133ms/step
+328/328 - 43s - loss: 0.0657 - accuracy: 0.9752 - val_loss: 0.1005 - val_accuracy: 0.9668 - 43s/epoch - 131ms/step
-<keras.callbacks.History object at 0x7fb3666e8e10>
+<keras.callbacks.History object at 0x7f1830403810>
</pre></div>
</div>
</div>
@@ -963,7 +963,7 @@ 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 22.236 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 4 minutes 33.046 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 c8188906a9..a1852415e7 100644
--- a/docs/how_to/work_with_microtvm/sg_execution_times.html
+++ b/docs/how_to/work_with_microtvm/sg_execution_times.html
@@ -340,7 +340,7 @@
<div class="section" id="computation-times">
<span id="sphx-glr-how-to-work-with-microtvm-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>06:36.598</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
+<p><strong>06:47.347</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -349,23 +349,23 @@
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="micro_train.html#sphx-glr-how-to-work-with-microtvm-micro-train-py"><span class="std std-ref">Training Vision Models for microTVM on Arduino</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_train.py</span></code>)</p></td>
-<td><p>04:22.236</p></td>
+<td><p>04:33.046</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><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">microTVM PyTorch Tutorial</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_pytorch.py</span></code>)</p></td>
-<td><p>01:09.814</p></td>
+<td><p>01:10.103</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="micro_autotune.html#sphx-glr-how-to-work-with-microtvm-micro-autotune-py"><span class="std std-ref">Autotuning with microTVM</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_autotune.py</span></code>)</p></td>
-<td><p>00:51.727</p></td>
+<td><p>00:51.386</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">microTVM Host-Driven AoT</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_aot.py</span></code>)</p></td>
-<td><p>00:09.034</p></td>
+<td><p>00:09.046</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="micro_tflite.html#sphx-glr-how-to-work-with-microtvm-micro-tflite-py"><span class="std std-ref">microTVM with TFLite Models</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_tflite.py</span></code>)</p></td>
-<td><p>00:03.787</p></td>
+<td><p>00:03.765</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="micro_tvmc.html#sphx-glr-how-to-work-with-microtvm-micro-tvmc-py"><span class="std std-ref">Executing a Tiny Model with TVMC Micro</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_tvmc.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 916d6e2c83..f713a1c149 100644
--- a/docs/how_to/work_with_relay/sg_execution_times.html
+++ b/docs/how_to/work_with_relay/sg_execution_times.html
@@ -340,7 +340,7 @@
<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>00:45.047</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
+<p><strong>00:44.381</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 84%" />
@@ -349,15 +349,15 @@
</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>00:32.702</p></td>
+<td><p>00:32.572</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>00:10.617</p></td>
+<td><p>00:10.261</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>00:01.722</p></td>
+<td><p>00:01.542</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 b988e43109..bf962a57d4 100644
--- a/docs/how_to/work_with_schedules/intrin_math.html
+++ b/docs/how_to/work_with_schedules/intrin_math.html
@@ -535,7 +535,7 @@ 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 0x7fb3075403b0>
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span><function my_cuda_math_rule at 0x7f1830b663b0>
</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 e0be25cd18..d40d572c0e 100644
--- a/docs/how_to/work_with_schedules/sg_execution_times.html
+++ b/docs/how_to/work_with_schedules/sg_execution_times.html
@@ -340,7 +340,7 @@
<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>00:06.262</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
+<p><strong>00:07.169</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 83%" />
@@ -349,23 +349,23 @@
</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>00:03.654</p></td>
+<td><p>00:04.652</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>00:01.246</p></td>
+<td><p>00:01.154</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>00:00.576</p></td>
+<td><p>00:00.579</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>00:00.559</p></td>
+<td><p>00:00.557</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>00:00.119</p></td>
+<td><p>00:00.116</p></td>
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="schedule_primitives.html#sphx-glr-how-to-work-with-schedules-schedule-primitives-py"><span class="std std-ref">Schedule Primitives in TVM</span></a> (<code class="docutils literal notranslate"><span class="pre">schedule_primitives.py</span></code>)</p></td>
@@ -373,7 +373,7 @@
<td><p>0.0 MB</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="tedd.html#sphx-glr-how-to-work-with-schedules-tedd-py"><span class="std std-ref">Use Tensor Expression Debug Display (TEDD) for Visualization</span></a> (<code class="docutils literal notranslate"><span class="pre">tedd.py</span></code>)</p></td>
-<td><p>00:00.033</p></td>
+<td><p>00:00.035</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>
diff --git a/docs/how_to/work_with_schedules/tensorize.html b/docs/how_to/work_with_schedules/tensorize.html
index 9dd31c9218..04c62ba759 100644
--- a/docs/how_to/work_with_schedules/tensorize.html
+++ b/docs/how_to/work_with_schedules/tensorize.html
@@ -586,7 +586,7 @@ class Module:
def main(A: T.Buffer((1024, 64), "float32"), B: T.Buffer((512, 64), "float32"), C: T.Buffer((1024, 512), "float32")):
T.func_attr({"from_legacy_te_schedule": True, "global_symbol": "main", "tir.noalias": True})
i = T.var("int32")
- T.attr(T.iter_var(i, None, "DataPar", ""), "pragma_import_llvm", "; ModuleID = '/tmp/tmpro9643gx/input0.cc'\nsource_filename = \"/tmp/tmpro9643gx/input0.cc\"\ntarget datalayout = \"e-m:e-i64:64-f80:128-n8:16:32:64-S128\"\ntarget triple = \"x86_64-pc-linux-gnu\"\n\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local i32 @gemv_update(float*, float*, float*, i32, i32, i32) #0 {\n %7 = alloca [...]
+ T.attr(T.iter_var(i, None, "DataPar", ""), "pragma_import_llvm", "; ModuleID = '/tmp/tmp0k_63x2_/input0.cc'\nsource_filename = \"/tmp/tmp0k_63x2_/input0.cc\"\ntarget datalayout = \"e-m:e-i64:64-f80:128-n8:16:32:64-S128\"\ntarget triple = \"x86_64-pc-linux-gnu\"\n\n; Function Attrs: noinline nounwind optnone uwtable\ndefine dso_local i32 @gemv_update(float*, float*, float*, i32, i32, i32) #0 {\n %7 = alloca [...]
for i, j_outer in T.grid(1024, 32):
T.call_extern("int32", "gemv_update", T.tvm_access_ptr(T.type_annotation("float32"), C.data, i * 512 + j_outer * 16, 16, 2), T.tvm_access_ptr(T.type_annotation("float32"), A.data, i * 64, 64, 1), T.tvm_access_ptr(T.type_annotation("float32"), B.data, j_outer * 1024, 1024, 1), 16, 64, 64)
</pre></div>
diff --git a/docs/reference/api/python/auto_scheduler.html b/docs/reference/api/python/auto_scheduler.html
index d67bca2514..f05f565e40 100644
--- a/docs/reference/api/python/auto_scheduler.html
+++ b/docs/reference/api/python/auto_scheduler.html
@@ -1615,7 +1615,7 @@ 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>
@@ -1899,7 +1899,7 @@ 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 308dc54933..b170dc7935 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/18b7dc1dd/web/src/rpc_server.ts#L43">rpc_server.ts:43</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L43">rpc_server.ts:43</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/18b7dc1dd/web/src/rpc_server.ts#L43">rpc_server.ts:43</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L43">rpc_server.ts:43</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/18b7dc1dd/web/src/rpc_server.ts#L42">rpc_server.ts:42</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L42">rpc_server.ts:42</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/18b7dc1dd/web/src/rpc_server.ts#L63">rpc_server.ts:63</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L63">rpc_server.ts:63</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/18b7dc1dd/web/src/rpc_server.ts#L49">rpc_server.ts:49</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L49">rpc_server.ts:49</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/18b7dc1dd/web/src/rpc_server.ts#L57">rpc_server.ts:57</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L57">rpc_server.ts:57</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 5b198b68cd..d003f7789d 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/18b7dc1dd/web/src/memory.ts#L223">memory.ts:223</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L208">memory.ts:208</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L312">memory.ts:312</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L284">memory.ts:284</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L388">memory.ts:388</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L376">memory.ts:376</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L267">memory.ts:267</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L243">memory.ts:243</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L321">memory.ts:321</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L252">memory.ts:252</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L359">memory.ts:359</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L342">memory.ts:342</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L350">memory.ts:350</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L326">memory.ts:326</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L363">memory.ts:363</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L346">memory.ts:346</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L334">memory.ts:334</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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 99917893fd..83aadedbcd 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/18b7dc1dd/web/src/runtime.ts#L262">runtime.ts:262</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L262">runtime.ts:262</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/18b7dc1dd/web/src/runtime.ts#L260">runtime.ts:260</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L260">runtime.ts:260</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/18b7dc1dd/web/src/runtime.ts#L258">runtime.ts:258</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L258">runtime.ts:258</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/18b7dc1dd/web/src/runtime.ts#L262">runtime.ts:262</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L262">runtime.ts:262</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/18b7dc1dd/web/src/runtime.ts#L279">runtime.ts:279</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L279">runtime.ts:279</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/18b7dc1dd/web/src/runtime.ts#L270">runtime.ts:270</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L270">runtime.ts:270</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 4f6c6f3761..7cfa6d73a4 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/18b7dc1dd/web/src/runtime.ts#L202">runtime.ts:202</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L202">runtime.ts:202</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/18b7dc1dd/web/src/runtime.ts#L200">runtime.ts:200</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L200">runtime.ts:200</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/18b7dc1dd/web/src/runtime.ts#L198">runtime.ts:198</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L198">runtime.ts:198</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/18b7dc1dd/web/src/runtime.ts#L223">runtime.ts:223</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L223">runtime.ts:223</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/18b7dc1dd/web/src/runtime.ts#L230">runtime.ts:230</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L230">runtime.ts:230</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 7095ec3134..d4dcdc25d5 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/18b7dc1dd/web/src/environment.ts#L86">environment.ts:86</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/environment.ts#L70">environment.ts:70</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/environment.ts#L69">environment.ts:69</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/environment.ts#L78">environment.ts:78</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/environment.ts#L84">environment.ts:84</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/environment.ts#L105">environment.ts:105</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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 8e584151d9..6911f5b560 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/18b7dc1dd/web/src/runtime.ts#L49">runtime.ts:49</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L49">runtime.ts:49</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/18b7dc1dd/web/src/runtime.ts#L46">runtime.ts:46</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L46">runtime.ts:46</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/18b7dc1dd/web/src/runtime.ts#L45">runtime.ts:45</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L45">runtime.ts:45</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/18b7dc1dd/web/src/runtime.ts#L44">runtime.ts:44</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L44">runtime.ts:44</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/18b7dc1dd/web/src/runtime.ts#L47">runtime.ts:47</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L47">runtime.ts:47</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/18b7dc1dd/web/src/runtime.ts#L76">runtime.ts:76</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L76">runtime.ts:76</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/18b7dc1dd/web/src/runtime.ts#L66">runtime.ts:66</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L66">runtime.ts:66</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/18b7dc1dd/web/src/runtime.ts#L84">runtime.ts:84</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L84">runtime.ts:84</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/18b7dc1dd/web/src/runtime.ts#L95">runtime.ts:95</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L95">runtime.ts:95</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/18b7dc1dd/web/src/runtime.ts#L72">runtime.ts:72</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L72">runtime.ts:72</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/graphexecutor.html b/docs/reference/api/typedoc/classes/graphexecutor.html
index 72bc0348c1..c4510e866c 100644
--- a/docs/reference/api/typedoc/classes/graphexecutor.html
+++ b/docs/reference/api/typedoc/classes/graphexecutor.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/18b7dc1dd/web/src/runtime.ts#L583">runtime.ts:583</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L583">runtime.ts:583</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -162,7 +162,7 @@
<div class="tsd-signature tsd-kind-icon">module<span class="tsd-signature-symbol">:</span> <a href="module.html" class="tsd-signature-type">Module</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L579">runtime.ts:579</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L579">runtime.ts:579</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/18b7dc1dd/web/src/runtime.ts#L654">runtime.ts:654</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L654">runtime.ts:654</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -224,7 +224,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L597">runtime.ts:597</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L597">runtime.ts:597</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -241,7 +241,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L631">runtime.ts:631</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L631">runtime.ts:631</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -279,7 +279,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L644">runtime.ts:644</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L644">runtime.ts:644</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -310,7 +310,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L621">runtime.ts:621</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L621">runtime.ts:621</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -332,7 +332,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L609">runtime.ts:609</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L609">runtime.ts:609</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
diff --git a/docs/reference/api/typedoc/classes/instance.html b/docs/reference/api/typedoc/classes/instance.html
index 36f0311231..2ca239ba77 100644
--- a/docs/reference/api/typedoc/classes/instance.html
+++ b/docs/reference/api/typedoc/classes/instance.html
@@ -139,7 +139,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L692">runtime.ts:692</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L692">runtime.ts:692</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -202,7 +202,7 @@
<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/18b7dc1dd/web/src/runtime.ts#L684">runtime.ts:684</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L684">runtime.ts:684</a></li>
</ul>
</aside>
</section>
@@ -212,7 +212,7 @@
<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/18b7dc1dd/web/src/runtime.ts#L683">runtime.ts:683</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L683">runtime.ts:683</a></li>
</ul>
</aside>
</section>
@@ -229,7 +229,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L932">runtime.ts:932</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L932">runtime.ts:932</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -260,7 +260,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L994">runtime.ts:994</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L994">runtime.ts:994</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -303,7 +303,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L924">runtime.ts:924</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L924">runtime.ts:924</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -341,7 +341,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L732">runtime.ts:732</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L732">runtime.ts:732</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -358,7 +358,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L952">runtime.ts:952</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L952">runtime.ts:952</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -402,7 +402,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L816">runtime.ts:816</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L816">runtime.ts:816</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -434,7 +434,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L1033">runtime.ts:1033</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L1033">runtime.ts:1033</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -465,7 +465,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L846">runtime.ts:846</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L846">runtime.ts:846</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -497,7 +497,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L750">runtime.ts:750</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L750">runtime.ts:750</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -520,7 +520,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L1013">runtime.ts:1013</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L1013">runtime.ts:1013</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -568,7 +568,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L789">runtime.ts:789</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L789">runtime.ts:789</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -608,7 +608,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L914">runtime.ts:914</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L914">runtime.ts:914</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -646,7 +646,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L1145">runtime.ts:1145</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L1145">runtime.ts:1145</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -698,7 +698,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L740">runtime.ts:740</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L740">runtime.ts:740</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -722,7 +722,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L868">runtime.ts:868</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L868">runtime.ts:868</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -754,7 +754,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L857">runtime.ts:857</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L857">runtime.ts:857</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -786,7 +786,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L940">runtime.ts:940</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L940">runtime.ts:940</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 07eba0d859..13d36ea4cd 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/18b7dc1dd/web/src/memory.ts#L40">memory.ts:40</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L32">memory.ts:32</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L33">memory.ts:33</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L154">memory.ts:154</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L90">memory.ts:90</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L97">memory.ts:97</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L74">memory.ts:74</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L81">memory.ts:81</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L104">memory.ts:104</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L132">memory.ts:132</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L145">memory.ts:145</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L60">memory.ts:60</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L67">memory.ts:67</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L53">memory.ts:53</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L114">memory.ts:114</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L124">memory.ts:124</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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/18b7dc1dd/web/src/memory.ts#L175">memory.ts:175</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/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 04aa189795..8649aa037a 100644
--- a/docs/reference/api/typedoc/classes/module.html
+++ b/docs/reference/api/typedoc/classes/module.html
@@ -124,7 +124,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L504">runtime.ts:504</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L504">runtime.ts:504</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -170,7 +170,7 @@
<div class="tsd-signature tsd-kind-icon">handle<span class="tsd-signature-symbol">:</span> <a href="../index.html#pointer" class="tsd-signature-type">Pointer</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L502">runtime.ts:502</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L502">runtime.ts:502</a></li>
</ul>
</aside>
</section>
@@ -187,7 +187,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L516">runtime.ts:516</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L516">runtime.ts:516</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -204,7 +204,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L530">runtime.ts:530</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L530">runtime.ts:530</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -236,7 +236,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L561">runtime.ts:561</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L561">runtime.ts:561</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 5d3d8b7f0d..2d2ce375fd 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/18b7dc1dd/web/src/runtime.ts#L304">runtime.ts:304</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L304">runtime.ts:304</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/18b7dc1dd/web/src/runtime.ts#L297">runtime.ts:297</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L297">runtime.ts:297</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/18b7dc1dd/web/src/runtime.ts#L293">runtime.ts:293</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L293">runtime.ts:293</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -188,7 +188,7 @@
<div class="tsd-signature tsd-kind-icon">handle<span class="tsd-signature-symbol">:</span> <a href="../index.html#pointer" class="tsd-signature-type">Pointer</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L289">runtime.ts:289</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L289">runtime.ts:289</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -203,7 +203,7 @@
<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/18b7dc1dd/web/src/runtime.ts#L291">runtime.ts:291</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L291">runtime.ts:291</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -218,7 +218,7 @@
<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/18b7dc1dd/web/src/runtime.ts#L295">runtime.ts:295</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L295">runtime.ts:295</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -240,7 +240,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L370">runtime.ts:370</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L370">runtime.ts:370</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -273,7 +273,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L414">runtime.ts:414</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L414">runtime.ts:414</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -305,7 +305,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L355">runtime.ts:355</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L355">runtime.ts:355</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -322,7 +322,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L474">runtime.ts:474</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L474">runtime.ts:474</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -346,7 +346,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L443">runtime.ts:443</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L443">runtime.ts:443</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 c3488f7f70..a45e27fcd6 100644
--- a/docs/reference/api/typedoc/classes/packedfunccell.html
+++ b/docs/reference/api/typedoc/classes/packedfunccell.html
@@ -122,7 +122,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L158">runtime.ts:158</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L158">runtime.ts:158</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -147,7 +147,7 @@
<div class="tsd-signature tsd-kind-icon">handle<span class="tsd-signature-symbol">:</span> <a href="../index.html#pointer" class="tsd-signature-type">Pointer</a></div>
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L157">runtime.ts:157</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L157">runtime.ts:157</a></li>
</ul>
</aside>
</section>
@@ -164,7 +164,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L165">runtime.ts:165</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L165">runtime.ts:165</a></li>
</ul>
</aside>
<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
diff --git a/docs/reference/api/typedoc/classes/rpcserver.html b/docs/reference/api/typedoc/classes/rpcserver.html
index d62ad1c71f..61f4b6686f 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/18b7dc1dd/web/src/rpc_server.ts#L92">rpc_server.ts:92</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L92">rpc_server.ts:92</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/18b7dc1dd/web/src/rpc_server.ts#L82">rpc_server.ts:82</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L82">rpc_server.ts:82</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/18b7dc1dd/web/src/rpc_server.ts#L78">rpc_server.ts:78</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L78">rpc_server.ts:78</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/18b7dc1dd/web/src/rpc_server.ts#L81">rpc_server.ts:81</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L81">rpc_server.ts:81</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/18b7dc1dd/web/src/rpc_server.ts#L79">rpc_server.ts:79</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L79">rpc_server.ts:79</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/18b7dc1dd/web/src/rpc_server.ts#L80">rpc_server.ts:80</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L80">rpc_server.ts:80</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/18b7dc1dd/web/src/rpc_server.ts#L77">rpc_server.ts:77</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/rpc_server.ts#L77">rpc_server.ts:77</a></li>
</ul>
</aside>
</section>
diff --git a/docs/reference/api/typedoc/classes/scalar.html b/docs/reference/api/typedoc/classes/scalar.html
index d700dffb93..2b7ac5285d 100644
--- a/docs/reference/api/typedoc/classes/scalar.html
+++ b/docs/reference/api/typedoc/classes/scalar.html
@@ -112,7 +112,7 @@
<li class="tsd-description">
<aside class="tsd-sources">
<ul>
- <li>Defined in <a href="https://github.com/apache/tvm/blob/18b7dc1dd/web/src/runtime.ts#L145">runtime.ts:145</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L145">runtime.ts:145</a></li>
</ul>
</aside>
<h4 class="tsd-parameters-title">Parameters</h4>
@@ -137,7 +137,7 @@
<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/18b7dc1dd/web/src/runtime.ts#L145">runtime.ts:145</a></li>
+ <li>Defined in <a href="https://github.com/apache/tvm/blob/56771a87d/web/src/runtime.ts#L145">runtime.ts:145</a></li>
</ul>
</aside>
<div class="tsd-comment tsd-typography">
@@ -152,7 +152,7 @@
<div class="tsd-signature tsd-kind-icon">value<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
<aside class="tsd-sources">
<ul>
... 1765 lines suppressed ...