You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by tq...@apache.org on 2023/03/21 09:20:05 UTC

[tvm-site] branch asf-site updated: deploying docs (apache/tvm@4819300803fd2477180d3b38972e24eb080f1c08)

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 b8ebba9964 deploying docs (apache/tvm@4819300803fd2477180d3b38972e24eb080f1c08)
b8ebba9964 is described below

commit b8ebba9964e1fb2d61547c882e99f78ad214254b
Author: tvm-bot <95...@users.noreply.github.com>
AuthorDate: Tue Mar 21 09:19:56 2023 +0000

    deploying docs (apache/tvm@4819300803fd2477180d3b38972e24eb080f1c08)
---
 docs/_images/sphx_glr_micro_train_001.png          |  Bin 307018 -> 346025 bytes
 docs/_images/sphx_glr_micro_train_thumb.png        |  Bin 22335 -> 24054 bytes
 .../how_to/compile_models/from_darknet.rst.txt     |    2 +-
 .../how_to/compile_models/from_keras.rst.txt       |    2 +-
 .../how_to/compile_models/from_mxnet.rst.txt       |    2 +-
 .../how_to/compile_models/from_oneflow.rst.txt     |    2 +-
 .../how_to/compile_models/from_pytorch.rst.txt     |    2 +-
 .../how_to/compile_models/from_tensorflow.rst.txt  |    2 +-
 .../compile_models/sg_execution_times.rst.txt      |   22 +-
 .../deploy_models/deploy_model_on_adreno.rst.txt   |    2 +-
 .../deploy_model_on_adreno_tvmc.rst.txt            |    2 +-
 .../deploy_models/deploy_model_on_android.rst.txt  |    2 +-
 .../deploy_object_detection_pytorch.rst.txt        |    4 +-
 .../deploy_models/deploy_prequantized.rst.txt      |    6 +-
 .../deploy_prequantized_tflite.rst.txt             |    4 +-
 .../how_to/deploy_models/deploy_quantized.rst.txt  |    2 +-
 .../deploy_models/deploy_ssd_gluoncv.rst.txt       |    4 +-
 .../deploy_models/sg_execution_times.rst.txt       |   22 +-
 .../extend_tvm/bring_your_own_datatypes.rst.txt    |    2 +-
 .../how_to/extend_tvm/sg_execution_times.rst.txt   |    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                 | 2332 ++++++++------------
 .../tune_network_cuda.rst.txt                      |    4 +-
 .../tune_network_x86.rst.txt                       |    4 +-
 .../tune_sparse_x86.rst.txt                        |  160 +-
 .../tune_with_autotvm/sg_execution_times.rst.txt   |   10 +-
 .../tune_with_autotvm/tune_conv2d_cuda.rst.txt     |  331 ++-
 .../work_with_microtvm/micro_autotune.rst.txt      |   18 +-
 .../work_with_microtvm/micro_pytorch.rst.txt       |    4 +-
 .../how_to/work_with_microtvm/micro_train.rst.txt  |   18 +-
 .../work_with_microtvm/sg_execution_times.rst.txt  |   14 +-
 .../work_with_relay/sg_execution_times.rst.txt     |    8 +-
 .../how_to/work_with_schedules/intrin_math.rst.txt |    2 +-
 .../work_with_schedules/sg_execution_times.rst.txt |   14 +-
 .../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 |    4 +-
 .../tutorial/auto_scheduler_matmul_x86.rst.txt     |   13 +-
 docs/_sources/tutorial/autotvm_matmul_x86.rst.txt  |   20 +-
 docs/_sources/tutorial/autotvm_relay_x86.rst.txt   |   56 +-
 .../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       |    8 +-
 docs/how_to/compile_models/from_tensorflow.html    |    2 +-
 docs/how_to/compile_models/sg_execution_times.html |   22 +-
 .../deploy_models/deploy_model_on_adreno.html      |    2 +-
 .../deploy_models/deploy_model_on_adreno_tvmc.html |   30 +-
 .../deploy_models/deploy_model_on_android.html     |    2 +-
 .../deploy_object_detection_pytorch.html           |   42 +-
 docs/how_to/deploy_models/deploy_prequantized.html |    9 +-
 .../deploy_models/deploy_prequantized_tflite.html  |    4 +-
 docs/how_to/deploy_models/deploy_quantized.html    |    2 +-
 docs/how_to/deploy_models/deploy_ssd_gluoncv.html  |   38 +-
 docs/how_to/deploy_models/sg_execution_times.html  |   22 +-
 .../extend_tvm/bring_your_own_datatypes.html       |    2 +-
 docs/how_to/extend_tvm/sg_execution_times.html     |    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                    | 2332 ++++++++------------
 .../tune_with_autoscheduler/tune_network_cuda.html |    4 +-
 .../tune_with_autoscheduler/tune_network_x86.html  |    4 +-
 .../tune_with_autoscheduler/tune_sparse_x86.html   |  160 +-
 .../tune_with_autotvm/sg_execution_times.html      |   12 +-
 .../how_to/tune_with_autotvm/tune_conv2d_cuda.html |  331 ++-
 docs/how_to/work_with_microtvm/micro_autotune.html |   18 +-
 docs/how_to/work_with_microtvm/micro_pytorch.html  |    5 +-
 docs/how_to/work_with_microtvm/micro_train.html    |   16 +-
 .../work_with_microtvm/sg_execution_times.html     |   14 +-
 .../how_to/work_with_relay/sg_execution_times.html |    8 +-
 docs/how_to/work_with_schedules/intrin_math.html   |    2 +-
 .../work_with_schedules/sg_execution_times.html    |   14 +-
 docs/install/nnpack.html                           |   12 +-
 docs/reference/api/python/auto_scheduler.html      |    4 +-
 .../api/typedoc/classes/bytestreamreader.html      |   12 +-
 .../api/typedoc/classes/cachedcallstack.html       |   34 +-
 docs/reference/api/typedoc/classes/dldatatype.html |   12 +-
 docs/reference/api/typedoc/classes/dldevice.html   |   10 +-
 .../reference/api/typedoc/classes/environment.html |   12 +-
 docs/reference/api/typedoc/classes/ffilibrary.html |   20 +-
 docs/reference/api/typedoc/classes/instance.html   |   58 +-
 docs/reference/api/typedoc/classes/memory.html     |   34 +-
 docs/reference/api/typedoc/classes/module.html     |   10 +-
 docs/reference/api/typedoc/classes/ndarray.html    |   22 +-
 .../api/typedoc/classes/packedfunccell.html        |    6 +-
 docs/reference/api/typedoc/classes/rpcserver.html  |   14 +-
 .../api/typedoc/classes/runtimecontext.html        |   22 +-
 docs/reference/api/typedoc/classes/scalar.html     |    6 +-
 docs/reference/api/typedoc/classes/tvmarray.html   |   16 +-
 docs/reference/api/typedoc/classes/tvmobject.html  |   12 +-
 .../api/typedoc/classes/webgpucontext.html         |   12 +-
 docs/reference/api/typedoc/enums/argtypecode.html  |   30 +-
 .../api/typedoc/enums/aynccallbackcode.html        |    4 +-
 .../api/typedoc/enums/dldatatypecode.html          |    8 +-
 .../api/typedoc/enums/rpcserverstate.html          |   12 +-
 docs/reference/api/typedoc/enums/sizeof.html       |   18 +-
 docs/reference/api/typedoc/index.html              |  124 +-
 .../api/typedoc/interfaces/disposable.html         |    2 +-
 .../api/typedoc/interfaces/functioninfo.html       |    6 +-
 .../api/typedoc/interfaces/libraryprovider.html    |    4 +-
 docs/searchindex.js                                |    2 +-
 .../vta/tutorials/autotvm/sg_execution_times.html  |    4 +-
 .../tutorials/frontend/deploy_classification.html  |    2 +-
 .../vta/tutorials/frontend/deploy_detection.html   |    2 +-
 .../vta/tutorials/frontend/sg_execution_times.html |    6 +-
 .../vta/tutorials/optimize/sg_execution_times.html |    6 +-
 docs/topic/vta/tutorials/sg_execution_times.html   |    4 +-
 docs/tutorial/auto_scheduler_matmul_x86.html       |    8 +-
 docs/tutorial/autotvm_matmul_x86.html              |   20 +-
 docs/tutorial/autotvm_relay_x86.html               |  268 +--
 docs/tutorial/cross_compilation_and_rpc.html       |    2 +-
 docs/tutorial/intro_topi.html                      |    2 +-
 docs/tutorial/sg_execution_times.html              |   22 +-
 docs/tutorial/tensor_expr_get_started.html         |   44 +-
 132 files changed, 3423 insertions(+), 3947 deletions(-)

diff --git a/docs/_images/sphx_glr_micro_train_001.png b/docs/_images/sphx_glr_micro_train_001.png
index 117ca26f92..b216b6b215 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 f942f0e17d..ebb8af5037 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 fc493530e8..ae78c824ab 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  17.627 seconds)
+   **Total running time of the script:** ( 1 minutes  20.524 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 9705b23dc4..abd25658d0 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 946ms/step
+
    1/1 [==============================] - ETA: 0s
    1/1 [==============================] - 1s 953ms/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 1b1775a96c..d745b8372c 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.zipff49ba47-aac5-4505-a39d-7b52e8e5d528 from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet18_v1-a0666292.zip...
+    Downloading /workspace/.mxnet/models/resnet18_v1-a0666292.zip7edf74da-5053-4323-a546-fa38db9222a0 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 1be9a4e38c..178d54d3ba 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, 80.8MB/s]
     39%|###8      | 16.1M/41.5M [00:00<00:00, 71.3MB/s]
     55%|#####5    | 23.0M/41.5M [00:00<00:00, 66.1MB/s]
     77%|#######7  | 32.1M/41.5M [00:00<00:00, 73.8MB/s]
    100%|##########| 41.5M/41.5M [00:00<00:00, 82.3MB/s]
+
      0%|          | 0.00/41.5M [00:00<?, ?B/s]
     22%|##1       | 9.04M/41.5M [00:00<00:00, 94.8MB/s]
     44%|####3     | 18.1M/41.5M [00:00<00:00, 78.3MB/s]
     62%|######1   | 25.7M/41.5M [00:00<00:00, 77.1MB/s]
     80%|#######9  | 33.2M/41.5M [00:00<00:00, 73.6MB/s]
     97%|#########6| 40.2M/41.5M [00:00<00:00, 62.9MB/s]
    100%|##########| 41.5M/41.5M [00:00<00:00, 70.8MB/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 bd3a0c22aa..d758a67056 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]
     28%|##7       | 12.3M/44.7M [00:00<00:00, 128MB/s]
     55%|#####4    | 24.5M/44.7M [00:00<00:00, 110MB/s]
     79%|#######8  | 35.1M/44.7M [00:00<00:00, 106MB/s]
    100%|##########| 44.7M/44.7M [00:00<00:00, 106MB/s]
+
      0%|          | 0.00/44.7M [00:00<?, ?B/s]
     26%|##6       | 11.6M/44.7M [00:00<00:00, 122MB/s]
     54%|#####3    | 24.0M/44.7M [00:00<00:00, 107MB/s]
     77%|#######7  | 34.5M/44.7M [00:00<00:00, 109MB/s]
    100%|##########| 44.7M/44.7M [00:00<00:00, 71.2MB/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 5baa85b9a1..0f3f997127 100644
--- a/docs/_sources/how_to/compile_models/from_tensorflow.rst.txt
+++ b/docs/_sources/how_to/compile_models/from_tensorflow.rst.txt
@@ -424,7 +424,7 @@ Run the corresponding model on tensorflow
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 1 minutes  23.887 seconds)
+   **Total running time of the script:** ( 1 minutes  23.477 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 baf7384cbc..1640d2fdb9 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
 =================
-**06:37.466** total execution time for **how_to_compile_models** files:
+**06:39.807** total execution time for **how_to_compile_models** files:
 
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:23.887 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tensorflow.py` (``from_tensorflow.py``) | 01:23.477 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``)       | 01:17.627 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_darknet.py` (``from_darknet.py``)       | 01:20.524 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``)         | 00:56.604 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_paddle.py` (``from_paddle.py``)         | 00:56.418 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``)       | 00:37.764 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_oneflow.py` (``from_oneflow.py``)       | 00:37.570 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``)         | 00:32.531 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_coreml.py` (``from_coreml.py``)         | 00:32.478 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``)           | 00:31.432 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_mxnet.py` (``from_mxnet.py``)           | 00:31.420 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``)         | 00:26.843 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_tflite.py` (``from_tflite.py``)         | 00:26.647 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``)       | 00:26.083 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_pytorch.py` (``from_pytorch.py``)       | 00:26.337 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``)           | 00:21.991 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_keras.py` (``from_keras.py``)           | 00:22.175 | 0.0 MB |
 +-----------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``)             | 00:02.704 | 0.0 MB |
+| :ref:`sphx_glr_how_to_compile_models_from_onnx.py` (``from_onnx.py``)             | 00:02.761 | 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 aaaab98d9f..42764e311a 100644
--- a/docs/_sources/how_to/deploy_models/deploy_model_on_adreno.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_model_on_adreno.rst.txt
@@ -637,7 +637,7 @@ well as provides information about the model's performance
     Evaluate inference time cost...
     Execution time summary:
      mean (ms)   median (ms)    max (ms)     min (ms)     std (ms)  
-     2535.3280    2534.9226    2538.4680    2533.7433      1.4229   
+     2538.1084    2537.6719    2541.9637    2536.2990      1.6237   
                
 
 
diff --git a/docs/_sources/how_to/deploy_models/deploy_model_on_adreno_tvmc.rst.txt b/docs/_sources/how_to/deploy_models/deploy_model_on_adreno_tvmc.rst.txt
index 547ed1d870..633d760156 100644
--- a/docs/_sources/how_to/deploy_models/deploy_model_on_adreno_tvmc.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_model_on_adreno_tvmc.rst.txt
@@ -127,7 +127,7 @@ Make a Keras Resnet50 Model
  .. code-block:: none
 
     Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels.h5
-
         8192/102967424 [..............................] - ETA: 0s
      6995968/102967424 [=>............................] - ETA: 0s
     12099584/102967424 [==>...........................] - ETA: 0s
     16769024/102967424 [===>..........................] - ETA: 0s
     22142976/102967424 [=====>........................] - ETA: 0s
     27222016/102967424 [======>.......................] - ETA: 0s
     32186368/102967424 [========>.....................] - ETA: 0s
     37208064/102967424 [=========>....................] - ETA: 0s
 
     41934848/102967424 [===========>..................] - ETA: 0s
     48603136/102967424 [=============>................] - ETA: 0s
     53714944/102967424 [==============>...............] - ETA: 0s
     58712064/102967424 [================>.............] - ETA: 0s
     64036864/102967424 [=================>............] - ETA: 0s
     69025792/102967424 [===================>..........] - ETA: 0s
     74006528/102967424 [====================>.........] - ETA: 0s
     79011840/102967424 [======================>.......] -
  ETA: 0s
     84803584/102967424 [=======================>......] - ETA: 0s
     89776128/102967424 [=========================>....] - ETA: 0s
     94330880/102967424 [==========================>...] - ETA: 0s
    100114432/102967424 [============================>.] - ETA: 0s
    102967424/102967424 [==============================] - 1s 0us/step
+
         8192/102967424 [..............................] - ETA: 0s
      8380416/102967424 [=>............................] - ETA: 1s
     16769024/102967424 [===>..........................] - ETA: 0s
     26320896/102967424 [======>.......................] - ETA: 0s
     33546240/102967424 [========>.....................] - ETA: 0s
     39624704/102967424 [==========>...................] - ETA: 0s
     42065920/102967424 [===========>..................] - ETA: 0s
     53239808/102967424 [==============>...............] - ETA: 0s
 
     58712064/102967424 [================>.............] - ETA: 0s
     65732608/102967424 [==================>...........] - ETA: 0s
     75169792/102967424 [====================>.........] - ETA: 0s
     75628544/102967424 [=====================>........] - ETA: 0s
     77684736/102967424 [=====================>........] - ETA: 0s
     83877888/102967424 [=======================>......] - ETA: 0s
     92405760/102967424 [=========================>....] - ETA: 0s
    100646912/102967424 [============================>.] -
  ETA: 0s
    102967424/102967424 [==============================] - 1s 0us/step
 
 
 
diff --git a/docs/_sources/how_to/deploy_models/deploy_model_on_android.rst.txt b/docs/_sources/how_to/deploy_models/deploy_model_on_android.rst.txt
index 82a1652710..909f80db2d 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)  
-      16.0745      16.0362      16.3460      15.9037       0.1235   
+      16.1313      16.0852      16.4169      15.9376       0.1567   
                
 
 
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 1990762583..a0cf070c8a 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, 69.9MB/s]
     10%|9         | 16.6M/170M [00:00<00:01, 80.9MB/s]
     15%|#5        | 25.6M/170M [00:00<00:01, 86.6MB/s]
     20%|#9        | 33.9M/170M [00:00<00:01, 80.9MB/s]
     26%|##6       | 44.5M/170M [00:00<00:01, 91.3MB/s]
     31%|###1      | 53.4M/170M [00:00<00:01, 79.6MB/s]
     38%|###7      | 64.0M/170M [00:00<00:01, 88.2MB/s]
     43%|####2     | 72.7M/170M [00:00<00:01, 80.4MB/s]
     52%|#####1    | 88.0M/170M [00:01<00:00, 94.9MB/s]
     57%|#####7    | 97.2M/170M [00:01<00:00, 91.1MB/s]
     66%|######5   | 112M/170M [00:01<00:00, 106MB/s]  
     72%|#######1  | 122M/170M [00:01<00:00, 96.3MB/s]
     80%|########  | 136M/170M [00:01<00:00, 98.9MB/s]
     89%|########9 | 152M/170M [00:01<00:00, 103MB/s] 
     95%|#########5| 162M/170M [00:02<00:00, 70.0MB/s]
    100%|##########| 170M/170M [00:02<00:00, 86.6MB/s]
+
      0%|          | 0.00/170M [00:00<?, ?B/s]
      4%|3         | 6.30M/170M [00:00<00:06, 28.0MB/s]
      5%|5         | 8.98M/170M [00:00<00:06, 26.6MB/s]
      9%|9         | 16.0M/170M [00:00<00:04, 32.5MB/s]
     13%|#3        | 22.3M/170M [00:00<00:03, 41.5MB/s]
     18%|#7        | 30.3M/170M [00:00<00:02, 50.3MB/s]
     23%|##2       | 38.5M/170M [00:00<00:02, 59.9MB/s]
     26%|##6       | 44.6M/170M [00:01<00:02, 54.8MB/s]
     31%|###       | 52.2M/170M [00:01<00:02, 61.2MB/s]
     37%|###6      | 62.3M/170M [00:01<00:01, 69.3MB/s]
     41%|####      | 69.1M/170M [00:01<00:01, 63.3MB/s]
     44%|####4     | 75.4M/170M [00:01<00:01, 63.3MB/s]
     48%|####8     | 81.6M/170M [00:01<00:01, 61.1MB/s]
     56%|#####5    | 94.6M/170M [00:01<00:00, 81.1MB/s]
     60%|######    | 103M/170M [00:01<00:00, 74.6MB/s] 
     65%|######4   | 110M/170M [00:01<00:00, 72.0MB/s]
     69%|######8   | 117M/170M [00:02<00:00, 66.5MB/s]
     73%|#######2  | 124M/170M [00:02<00:00, 64.2MB/s]
      76%|#######6  | 130M/170M [00:02<00:00, 56.6MB/s]
     80%|########  | 136M/170M [00:02<00:00, 54.2MB/s]
     85%|########4 | 144M/170M [00:02<00:00, 60.0MB/s]
     89%|########9 | 152M/170M [00:02<00:00, 60.7MB/s]
     94%|#########4| 160M/170M [00:02<00:00, 59.0MB/s]
     99%|#########8| 168M/170M [00:02<00:00, 64.3MB/s]
    100%|##########| 170M/170M [00:03<00:00, 58.9MB/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  39.868 seconds)
+   **Total running time of the script:** ( 3 minutes  38.121 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 830009a6a7..f9bf5caef1 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]
     59%|#####8    | 7.99M/13.6M [00:00<00:00, 63.1MB/s]
    100%|##########| 13.6M/13.6M [00:00<00:00, 78.1MB/s]
+
      0%|          | 0.00/13.6M [00:00<?, ?B/s]
     59%|#####8    | 7.99M/13.6M [00:00<00:00, 53.1MB/s]
     96%|#########6| 13.1M/13.6M [00:00<00:00, 51.2MB/s]
    100%|##########| 13.6M/13.6M [00:00<00:00, 52.9MB/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)  
-      90.4523      90.4264      91.4663      90.0434       0.2777   
+      90.4230      90.2571      96.1396      90.0245       0.7508   
                
 
 
@@ -458,7 +458,7 @@ TODO
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 1 minutes  18.206 seconds)
+   **Total running time of the script:** ( 1 minutes  17.766 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 e15944ca4c..55d12260ec 100644
--- a/docs/_sources/how_to/deploy_models/deploy_prequantized_tflite.rst.txt
+++ b/docs/_sources/how_to/deploy_models/deploy_prequantized_tflite.rst.txt
@@ -423,7 +423,7 @@ Here we give an example of how to measure performance of TVM compiled models.
 
     Execution time summary:
      mean (ms)   median (ms)    max (ms)     min (ms)     std (ms)  
-      119.5424     119.5118     120.5062     118.7352      0.3464   
+      118.1251     118.0449     124.9242     117.3214      0.7493   
                
 
 
@@ -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  30.194 seconds)
+   **Total running time of the script:** ( 2 minutes  26.632 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 3b3ac329ae..b660266149 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  37.937 seconds)
+   **Total running time of the script:** ( 1 minutes  39.156 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 732a854e51..44b3110742 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]
      1%|          | 1081/132723 [00:00<00:12, 10793.34KB/s]
      3%|2         | 3398/132723 [00:00<00:07, 18067.79KB/s]
      6%|6         | 8046/132723 [00:00<00:04, 30979.96KB/s]
     12%|#2        | 16404/132723 [00:00<00:02, 51716.28KB/s]
     19%|#8        | 25204/132723 [00:00<00:01, 64787.28KB/s]
     26%|##5       | 34017/132723 [00:00<00:01, 72715.63KB/s]
     32%|###2      | 42881/132723 [00:00<00:01, 77917.45KB/s]
     39%|###8      | 51683/132723 [00:00<00:00, 81131.35KB/s]
     46%|####5     | 60503/132723 [00:00<00:00, 83339.27KB/s]
     52%|#####2    | 69308/132723 [00:01<00:00, 84792.17KB/s]
     59%|#####8    | 78180/132723 [00:01<00:00, 85991.28KB/s]
     66%|######5   | 87068/132723 [00:01<00:00, 86867.72KB/s]
     72%|#######2  | 96009/132723 [00:01<00:00, 87634.97KB/s]
     79%|#######9  | 104959/132723 [00:01<00:00, 88195.92KB/s]
     86%|########5 | 113779/132723 [00:01<00:00, 87914.72KB/s]
     92%|#########2|
  122571/132723 [00:01<00:00, 87870.97KB/s]
     99%|#########9| 131485/132723 [00:01<00:00, 88248.87KB/s]
    100%|##########| 132723/132723 [00:01<00:00, 77254.29KB/s]
+
      0%|          | 0/132723 [00:00<?, ?KB/s]
      1%|          | 1233/132723 [00:00<00:10, 12289.05KB/s]
      3%|2         | 3719/132723 [00:00<00:06, 19669.36KB/s]
      7%|7         | 9384/132723 [00:00<00:03, 36539.49KB/s]
     13%|#3        | 17902/132723 [00:00<00:02, 55729.40KB/s]
     20%|#9        | 26501/132723 [00:00<00:01, 66634.48KB/s]
     26%|##6       | 35066/132723 [00:00<00:01, 73095.74KB/s]
     33%|###2      | 43717/132723 [00:00<00:01, 77476.55KB/s]
     39%|###9      | 52365/132723 [00:00<00:01, 80338.75KB/s]
     46%|####5     | 61016/132723 [00:00<00:00, 82266.36KB/s]
     52%|#####2    | 69667/132723 [00:01<00:00, 83573.07KB/s]
     59%|#####9    | 78313/132723 [00:01<00:00, 84453.78KB/s]
     66%|######5   | 86977/132723 [00:01<00:00, 85114.80KB/s]
     72%|#######2  | 95662/132723 [00:01<00:00, 85638.38KB/s]
     79%|#######8  | 104311/132723 [00:01<00:00, 85888.13KB/s]
     85%|########5 | 112983/132723 [00:01<00:00, 86136.53KB/s]
     92%|#########1|
  121599/132723 [00:01<00:00, 86140.40KB/s]
     98%|#########8| 130254/132723 [00:01<00:00, 86262.16KB/s]
    100%|##########| 132723/132723 [00:01<00:00, 76651.91KB/s]
 
 
 
@@ -246,7 +246,7 @@ Display result
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 3 minutes  48.374 seconds)
+   **Total running time of the script:** ( 3 minutes  48.677 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 de0fc79761..6308abc4ee 100644
--- a/docs/_sources/how_to/deploy_models/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/deploy_models/sg_execution_times.rst.txt
@@ -5,28 +5,28 @@
 
 Computation times
 =================
-**16:20.971** total execution time for **how_to_deploy_models** files:
+**16:16.916** total execution time for **how_to_deploy_models** files:
 
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``)                           | 03:48.374 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_ssd_gluoncv.py` (``deploy_ssd_gluoncv.py``)                           | 03:48.677 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 03:39.868 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_object_detection_pytorch.py` (``deploy_object_detection_pytorch.py``) | 03:38.121 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``)           | 02:30.194 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized_tflite.py` (``deploy_prequantized_tflite.py``)           | 02:26.632 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``)                               | 01:37.937 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_quantized.py` (``deploy_quantized.py``)                               | 01:39.156 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``)                         | 01:18.206 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_prequantized.py` (``deploy_prequantized.py``)                         | 01:17.766 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno.py` (``deploy_model_on_adreno.py``)                   | 00:55.133 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno.py` (``deploy_model_on_adreno.py``)                   | 00:55.145 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno_tvmc.py` (``deploy_model_on_adreno_tvmc.py``)         | 00:51.213 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_adreno_tvmc.py` (``deploy_model_on_adreno_tvmc.py``)         | 00:51.542 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``)                 | 00:43.500 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_android.py` (``deploy_model_on_android.py``)                 | 00:43.172 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``)                       | 00:28.498 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_nano.py` (``deploy_model_on_nano.py``)                       | 00:28.682 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``)                       | 00:28.043 | 0.0 MB |
+| :ref:`sphx_glr_how_to_deploy_models_deploy_model_on_rasp.py` (``deploy_model_on_rasp.py``)                       | 00:28.017 | 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 36a3e41b8a..04c894f3c0 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.zip76c23e2b-be0e-45c2-b357-b52ba13382ed from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/mobilenet0.25-9f83e440.zip...
+    Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zipd1bde805-15e1-4a94-8ac8-775af63ae770 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 c9465cc1b2..3be56528f2 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
 =================
-**00:55.422** total execution time for **how_to_extend_tvm** files:
+**00:55.248** 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``) | 00:51.500 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_bring_your_own_datatypes.py` (``bring_your_own_datatypes.py``) | 00:51.342 | 0.0 MB |
 +-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``)           | 00:02.813 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_instrument.py` (``use_pass_instrument.py``)           | 00:02.802 | 0.0 MB |
 +-------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_extend_tvm_use_pass_infra.py` (``use_pass_infra.py``)                     | 00:01.102 | 0.0 MB |
+| :ref:`sphx_glr_how_to_extend_tvm_use_pass_infra.py` (``use_pass_infra.py``)                     | 00:01.097 | 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 bfadb6a2cf..ec083e1e1d 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: 22343us [22343us] (47.86%; 47.86%)
-    FoldScaleAxis: 24339us [8us] (52.14%; 52.14%)
-            FoldConstant: 24330us [1883us] (52.12%; 99.97%)
-                    InferType: 22447us [22447us] (48.09%; 92.26%)
+    InferType: 22041us [22041us] (48.58%; 48.58%)
+    FoldScaleAxis: 23332us [7us] (51.42%; 51.42%)
+            FoldConstant: 23325us [1701us] (51.41%; 99.97%)
+                    InferType: 21625us [21625us] (47.66%; 92.71%)
 
 
 
@@ -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: 22191us [22191us] (48.42%; 48.42%)
-    FoldScaleAxis: 23642us [7us] (51.58%; 51.58%)
-            FoldConstant: 23634us [1796us] (51.57%; 99.97%)
-                    InferType: 21839us [21839us] (47.65%; 92.40%)
+    InferType: 21885us [21885us] (48.26%; 48.26%)
+    FoldScaleAxis: 23467us [6us] (51.74%; 51.74%)
+            FoldConstant: 23461us [1748us] (51.73%; 99.97%)
+                    InferType: 21713us [21713us] (47.88%; 92.55%)
 
 
 
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 4758863ab2..1b1c2c99ad 100644
--- a/docs/_sources/how_to/optimize_operators/opt_conv_cuda.rst.txt
+++ b/docs/_sources/how_to/optimize_operators/opt_conv_cuda.rst.txt
@@ -331,7 +331,7 @@ latency of convolution.
 
  .. code-block:: none
 
-    Convolution: 54.226974 ms
+    Convolution: 40.327167 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 6b2ecc61d4..a639b63afc 100644
--- a/docs/_sources/how_to/optimize_operators/opt_conv_tensorcore.rst.txt
+++ b/docs/_sources/how_to/optimize_operators/opt_conv_tensorcore.rst.txt
@@ -598,7 +598,7 @@ be able to run on our build server
 
  .. code-block:: none
 
-    conv2d with tensor core: 13.375407 ms
+    conv2d with tensor core: 13.350711 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 cbce8f07c1..a04601ffa9 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: 0.018962
-    Baseline: 3.307704
+    Numpy running time: 0.018989
+    Baseline: 3.314976
 
 
 
@@ -227,7 +227,7 @@ fill 32 * 32 * sizeof(float) which is 4KB in the cache whose total size is 32KB
 
  .. code-block:: none
 
-    Opt1: 0.303799
+    Opt1: 0.309745
 
 
 
@@ -318,7 +318,7 @@ In this tutorial, we chose to vectorize the inner loop row data since it is cach
 
  .. code-block:: none
 
-    Opt2: 0.339524
+    Opt2: 0.341284
 
 
 
@@ -406,7 +406,7 @@ the access pattern for A matrix is more cache friendly.
 
  .. code-block:: none
 
-    Opt3: 0.117044
+    Opt3: 0.116433
 
 
 
@@ -523,7 +523,7 @@ flattening.
 
  .. code-block:: none
 
-    Opt4: 0.114789
+    Opt4: 0.109233
 
 
 
@@ -635,7 +635,7 @@ write to C when all the block results are ready.
 
  .. code-block:: none
 
-    Opt5: 0.110691
+    Opt5: 0.110824
 
 
 
@@ -748,7 +748,7 @@ Furthermore, we can also utilize multi-core processors to do the thread-level pa
 
  .. code-block:: none
 
-    Opt6: 0.146671
+    Opt6: 0.147740
 
 
 
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 5cb90df8c4..0c7ac50646 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
 =================
-**00:35.049** total execution time for **how_to_optimize_operators** files:
+**00:35.022** total execution time for **how_to_optimize_operators** files:
 
 +-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``)                       | 00:32.194 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_gemm.py` (``opt_gemm.py``)                       | 00:32.313 | 0.0 MB |
 +-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.707 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_tensorcore.py` (``opt_conv_tensorcore.py``) | 00:01.627 | 0.0 MB |
 +-----------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``)             | 00:01.147 | 0.0 MB |
+| :ref:`sphx_glr_how_to_optimize_operators_opt_conv_cuda.py` (``opt_conv_cuda.py``)             | 00:01.083 | 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 b0829aecba..da0fb56cb7 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
 =================
-**10:02.671** total execution time for **how_to_tune_with_autoscheduler** files:
+**10:06.943** total execution time for **how_to_tune_with_autoscheduler** files:
 
 +----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_conv2d_layer_cuda.py` (``tune_conv2d_layer_cuda.py``) | 06:10.917 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_conv2d_layer_cuda.py` (``tune_conv2d_layer_cuda.py``) | 06:12.035 | 0.0 MB |
 +----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``)             | 01:42.867 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_x86.py` (``tune_network_x86.py``)             | 01:42.823 | 0.0 MB |
 +----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``)           | 01:07.853 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_cuda.py` (``tune_network_cuda.py``)           | 01:08.262 | 0.0 MB |
 +----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``)               | 00:32.813 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_sparse_x86.py` (``tune_sparse_x86.py``)               | 00:35.774 | 0.0 MB |
 +----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``)             | 00:14.364 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_arm.py` (``tune_network_arm.py``)             | 00:14.335 | 0.0 MB |
 +----------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``)           | 00:13.856 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autoscheduler_tune_network_mali.py` (``tune_network_mali.py``)           | 00:13.715 | 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 9ddbde93a2..baf8fb21f3 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
@@ -243,769 +243,479 @@ cooperative fetching, unrolling and operator fusion.
         @T.prim_func
         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.launch_thread("blockIdx.x", 8)
-            conv2d_nchw = T.allocate([28], "float32", "local")
-            pad_temp_shared = T.allocate([504], "float32", "shared")
-            kernel_shared = T.allocate([1536], "float32", "shared")
-            threadIdx_x = T.launch_thread("threadIdx.x", 112)
-            conv2d_nchw_1 = T.Buffer((196,), data=conv2d_nchw, scope="local", align=32)
+            blockIdx_x = T.launch_thread("blockIdx.x", 28)
+            conv2d_nchw = T.allocate([14], "float32", "local")
+            pad_temp_shared = T.allocate([72], "float32", "shared")
+            kernel_shared = T.allocate([3072], "float32", "shared")
+            threadIdx_x = T.launch_thread("threadIdx.x", 64)
+            conv2d_nchw_1 = T.Buffer((14,), data=conv2d_nchw, scope="local", align=32)
             conv2d_nchw_1[0] = T.float32(0)
-            conv2d_nchw_1[14] = T.float32(0)
             conv2d_nchw_1[1] = T.float32(0)
-            conv2d_nchw_1[15] = T.float32(0)
             conv2d_nchw_1[2] = T.float32(0)
-            conv2d_nchw_1[16] = T.float32(0)
             conv2d_nchw_1[3] = T.float32(0)
-            conv2d_nchw_1[17] = T.float32(0)
             conv2d_nchw_1[4] = T.float32(0)
-            conv2d_nchw_1[18] = T.float32(0)
             conv2d_nchw_1[5] = T.float32(0)
-            conv2d_nchw_1[19] = T.float32(0)
             conv2d_nchw_1[6] = T.float32(0)
-            conv2d_nchw_1[20] = T.float32(0)
             conv2d_nchw_1[7] = T.float32(0)
-            conv2d_nchw_1[21] = T.float32(0)
             conv2d_nchw_1[8] = T.float32(0)
-            conv2d_nchw_1[22] = T.float32(0)
             conv2d_nchw_1[9] = T.float32(0)
-            conv2d_nchw_1[23] = T.float32(0)
             conv2d_nchw_1[10] = T.float32(0)
-            conv2d_nchw_1[24] = T.float32(0)
             conv2d_nchw_1[11] = T.float32(0)
-            conv2d_nchw_1[25] = T.float32(0)
             conv2d_nchw_1[12] = T.float32(0)
-            conv2d_nchw_1[26] = T.float32(0)
             conv2d_nchw_1[13] = T.float32(0)
-            conv2d_nchw_1[27] = T.float32(0)
             for rc_outer_outer, ry_outer_outer in T.grid(64, 3):
-                cse_var_4: T.int32 = rc_outer_outer * 392
-                cse_var_3: T.int32 = ry_outer_outer * 7
                 cse_var_2: T.int32 = rc_outer_outer * 72
                 cse_var_1: T.int32 = ry_outer_outer * 3
+                pad_temp_shared_1 = T.Buffer((72,), data=pad_temp_shared, scope="shared")
+                with T.launch_thread("threadIdx.x", 64) as threadIdx_x_1:
+                    data_1 = T.Buffer((25088,), data=data.data)
+                    if T.likely(threadIdx_x_1 < 18):
+                        pad_temp_shared_1[threadIdx_x_1 * 4] = T.if_then_else(1 <= ry_outer_outer + blockIdx_x % 7 and ry_outer_outer + blockIdx_x % 7 < 8 and 1 <= threadIdx_x_1 * 4 % 9 and threadIdx_x_1 * 4 % 9 < 8, data_1[rc_outer_outer * 392 + threadIdx_x_1 * 4 // 9 * 49 + ry_outer_outer * 7 + blockIdx_x % 7 * 7 + threadIdx_x_1 * 4 % 9 - 8], T.float32(0))
+                    if T.likely(threadIdx_x_1 < 18):
+                        pad_temp_shared_1[threadIdx_x_1 * 4 + 1] = T.if_then_else(1 <= ry_outer_outer + blockIdx_x % 7 and ry_outer_outer + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 * 4 + 1) % 9 and (threadIdx_x_1 * 4 + 1) % 9 < 8, data_1[rc_outer_outer * 392 + (threadIdx_x_1 * 4 + 1) // 9 * 49 + ry_outer_outer * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 * 4 + 1) % 9 - 8], T.float32(0))
+                    if T.likely(threadIdx_x_1 < 18):
+                        pad_temp_shared_1[threadIdx_x_1 * 4 + 2] = T.if_then_else(1 <= ry_outer_outer + blockIdx_x % 7 and ry_outer_outer + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 * 4 + 2) % 9 and (threadIdx_x_1 * 4 + 2) % 9 < 8, data_1[rc_outer_outer * 392 + (threadIdx_x_1 * 4 + 2) // 9 * 49 + ry_outer_outer * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 * 4 + 2) % 9 - 8], T.float32(0))
+                    if T.likely(threadIdx_x_1 < 18):
+                        pad_temp_shared_1[threadIdx_x_1 * 4 + 3] = T.if_then_else(1 <= ry_outer_outer + blockIdx_x % 7 and ry_outer_outer + blockIdx_x % 7 < 8 and 1 <= (threadIdx_x_1 * 4 + 3) % 9 and (threadIdx_x_1 * 4 + 3) % 9 < 8, data_1[rc_outer_outer * 392 + (threadIdx_x_1 * 4 + 3) // 9 * 49 + ry_outer_outer * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 * 4 + 3) % 9 - 8], T.float32(0))
                 threadIdx_x_1 = T.env_thread("threadIdx.x")
-                pad_temp_shared_1 = T.Buffer((504,), data=pad_temp_shared, scope="shared")
-                data_1 = T.Buffer((25088,), data=data.data)
-                with T.launch_thread(threadIdx_x_1, 112):
-                    pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(1 <= threadIdx_x_1 % 63 // 9 + ry_outer_outer and threadIdx_x_1 % 63 // 9 + ry_outer_outer < 8 and 1 <= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 < 8, data_1[cse_var_4 + threadIdx_x_1 // 9 * 7 + cse_var_3 + threadIdx_x_1 % 9 - 8], T.float32(0))
-                with T.launch_thread(threadIdx_x_1, 112):
-                    pad_temp_shared_1[threadIdx_x_1 + 112] = T.if_then_else(1 <= (threadIdx_x_1 + 49) % 63 // 9 + ry_outer_outer and (threadIdx_x_1 + 49) % 63 // 9 + ry_outer_outer < 8 and 1 <= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 < 8, data_1[cse_var_4 + (threadIdx_x_1 + 112) // 9 * 7 + cse_var_3 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
-                with T.launch_thread(threadIdx_x_1, 112):
-                    pad_temp_shared_1[threadIdx_x_1 + 224] = T.if_then_else(1 <= (threadIdx_x_1 + 35) % 63 // 9 + ry_outer_outer and (threadIdx_x_1 + 35) % 63 // 9 + ry_outer_outer < 8 and 1 <= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 < 8, data_1[cse_var_4 + (threadIdx_x_1 + 224) // 9 * 7 + cse_var_3 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
-                with T.launch_thread(threadIdx_x_1, 112):
-                    pad_temp_shared_1[threadIdx_x_1 + 336] = T.if_then_else(1 <= (threadIdx_x_1 + 21) % 63 // 9 + ry_outer_outer and (threadIdx_x_1 + 21) % 63 // 9 + ry_outer_outer < 8 and 1 <= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 < 8, data_1[cse_var_4 + (threadIdx_x_1 + 336) // 9 * 7 + cse_var_3 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
-                with T.launch_thread(threadIdx_x_1, 112):
-                    if T.likely(threadIdx_x_1 < 56):
-                        pad_temp_shared_1[threadIdx_x_1 + 448] = T.if_then_else(1 <= (threadIdx_x_1 + 7) // 9 + ry_outer_outer and (threadIdx_x_1 + 7) // 9 + ry_outer_outer < 8 and 1 <= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 < 8, data_1[cse_var_4 + (threadIdx_x_1 + 448) // 9 * 7 + cse_var_3 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
-                threadIdx_x_2 = T.env_thread("threadIdx.x")
-                kernel_shared_1 = T.Buffer((1536,), data=kernel_shared, scope="shared")
+                kernel_shared_1 = T.Buffer((3072,), data=kernel_shared, scope="shared")
                 kernel_1 = T.Buffer((2359296,), data=kernel.data)
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[threadIdx_x_2] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[(threadIdx_x_2 + 112) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 112) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[(threadIdx_x_2 + 224) // 24 * 24 + (threadIdx_x_2 + 8) % 24 // 3 * 3 + (threadIdx_x_2 + 2) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 224) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 2) % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[threadIdx_x_2 + 336] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3 + 64512]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[(threadIdx_x_2 + 448) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 448) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[(threadIdx_x_2 + 560) // 24 * 24 + (threadIdx_x_2 + 8) % 24 // 3 * 3 + (threadIdx_x_2 + 2) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 560) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 2) % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[threadIdx_x_2 + 672] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3 + 129024]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[(threadIdx_x_2 + 784) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 784) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[(threadIdx_x_2 + 896) // 24 * 24 + (threadIdx_x_2 + 8) % 24 // 3 * 3 + (threadIdx_x_2 + 2) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 896) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 2) % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[threadIdx_x_2 + 1008] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3 + 193536]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[(threadIdx_x_2 + 1120) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 1120) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[(threadIdx_x_2 + 1232) // 24 * 24 + (threadIdx_x_2 + 8) % 24 // 3 * 3 + (threadIdx_x_2 + 2) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 1232) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 2) % 3]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    kernel_shared_1[threadIdx_x_2 + 1344] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3 + 258048]
-                with T.launch_thread(threadIdx_x_2, 112):
-                    if T.likely(threadIdx_x_2 < 80):
-                        kernel_shared_1[(threadIdx_x_2 + 1456) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 1456) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9] * kernel_shared_1[threadIdx_x // 7 * 48]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 63] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 63] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 126] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 126] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 189] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 189] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 63] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 63] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 126] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 126] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 189] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 189] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 8] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 8] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 71] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 71] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 134] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 134] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 197] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 197] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 8] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 8] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 71] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 71] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 134] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 134] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 197] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 197] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 252] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 252] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 315] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 315] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 378] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 378] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 441] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 441] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 252] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 252] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 315] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 315] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 378] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 378] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 441] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 441] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-                conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-                conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-                conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-                conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-                conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-                conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 260] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 260] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 323] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 323] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 386] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 386] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 449] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-                conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 449] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-                conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-                conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-                conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-                conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-                conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-                conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 260] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 260] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 323] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 323] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 386] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 386] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 449] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-                conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 449] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 64) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 64) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 128) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 128) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 192] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 36864]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 256) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 256) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 320) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 320) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 384] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 73728]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 448) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 448) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 512) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 512) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 576] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 110592]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 640) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 640) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 704) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 704) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 768] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 147456]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 832) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 832) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 896) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 896) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 960] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 184320]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1024) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1024) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1088) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1088) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 1152] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 221184]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1216) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1216) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1280) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1280) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 1344] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 258048]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1408) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1408) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1472) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1472) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 1536] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 294912]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1600) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1600) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1664) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1664) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 1728] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 331776]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1792) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1792) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1856) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1856) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 1920] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 368640]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 1984) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1984) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2048) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2048) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 2112] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 405504]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2176) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2176) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2240) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2240) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 2304] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 442368]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2368) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2368) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2432) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2432) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 2496] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 479232]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2560) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2560) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2624) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2624) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 2688] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 516096]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2752) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2752) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2816) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2816) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[threadIdx_x_1 + 2880] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 552960]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 2944) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2944) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+                with T.launch_thread(threadIdx_x_1, 64):
+                    kernel_shared_1[(threadIdx_x_1 + 3008) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 3008) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[0] * kernel_shared_1[threadIdx_x * 48]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[9] * kernel_shared_1[threadIdx_x * 48 + 3]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 48]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 48 + 3]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 3]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 3]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 3]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 3]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 3]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[0] * kernel_shared_1[threadIdx_x * 48 + 24]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[9] * kernel_shared_1[threadIdx_x * 48 + 27]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 48 + 24]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 48 + 27]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 24]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 27]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 24]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 27]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 24]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 27]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 24]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 27]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 24]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 27]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 48 + 1]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 48 + 4]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 1]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 4]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 1]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 4]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 1]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 4]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 1]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 4]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 1]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 4]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 48 + 1]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 48 + 4]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 48 + 25]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 48 + 28]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 25]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 28]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 25]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 28]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 25]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 28]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 25]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 28]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 25]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 28]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 48 + 25]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 48 + 28]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 2]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 5]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 2]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 5]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 2]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 5]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 2]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 5]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 2]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 5]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 48 + 2]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 48 + 5]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[8] * kernel_shared_1[threadIdx_x * 48 + 2]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[17] * kernel_shared_1[threadIdx_x * 48 + 5]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 26]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 29]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 26]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 29]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 26]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 29]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 26]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 29]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 26]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 29]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 48 + 26]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 48 + 29]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[8] * kernel_shared_1[threadIdx_x * 48 + 26]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[17] * kernel_shared_1[threadIdx_x * 48 + 29]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[18] * kernel_shared_1[threadIdx_x * 48 + 6]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[27] * kernel_shared_1[threadIdx_x * 48 + 9]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 48 + 6]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 48 + 9]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 6]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 9]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 6]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 9]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 6]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 9]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 6]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 9]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 6]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 9]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[18] * kernel_shared_1[threadIdx_x * 48 + 30]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[27] * kernel_shared_1[threadIdx_x * 48 + 33]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 48 + 30]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 48 + 33]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 30]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 33]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 30]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 33]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 30]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 33]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 30]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 33]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 30]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 33]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 48 + 7]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 48 + 10]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 7]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 10]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 7]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 10]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 7]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 10]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 7]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 10]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 7]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 10]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 48 + 7]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 48 + 10]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 48 + 31]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 48 + 34]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 31]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 34]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 31]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 34]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 31]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 34]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 31]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 34]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 31]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 34]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 48 + 31]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 48 + 34]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 8]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 11]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 8]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 11]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 8]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 11]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 8]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 11]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 8]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 11]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 48 + 8]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 48 + 11]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[26] * kernel_shared_1[threadIdx_x * 48 + 8]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[35] * kernel_shared_1[threadIdx_x * 48 + 11]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 32]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 35]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 32]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 35]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 32]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 35]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 32]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 35]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 32]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 35]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 48 + 32]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 48 + 35]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[26] * kernel_shared_1[threadIdx_x * 48 + 32]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[35] * kernel_shared_1[threadIdx_x * 48 + 35]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[36] * kernel_shared_1[threadIdx_x * 48 + 12]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[45] * kernel_shared_1[threadIdx_x * 48 + 15]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 48 + 12]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 48 + 15]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 12]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 15]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 12]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 15]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 12]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 15]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 12]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 15]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 12]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 15]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[36] * kernel_shared_1[threadIdx_x * 48 + 36]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[45] * kernel_shared_1[threadIdx_x * 48 + 39]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 48 + 36]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 48 + 39]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 36]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 39]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 36]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 39]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 36]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 39]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 36]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 39]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 36]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 39]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 48 + 13]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 48 + 16]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 13]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 16]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 13]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 16]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 13]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 16]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 13]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 16]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 13]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 16]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 48 + 13]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 48 + 16]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 48 + 37]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 48 + 40]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 37]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 40]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 37]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 40]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 37]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 40]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 37]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 40]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 37]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 40]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 48 + 37]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 48 + 40]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 14]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 17]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 14]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 17]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 14]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 17]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 14]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 17]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 14]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 17]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 48 + 14]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 48 + 17]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[44] * kernel_shared_1[threadIdx_x * 48 + 14]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[53] * kernel_shared_1[threadIdx_x * 48 + 17]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 38]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 41]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 38]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 41]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 38]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 41]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 38]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 41]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 38]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 41]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 48 + 38]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 48 + 41]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[44] * kernel_shared_1[threadIdx_x * 48 + 38]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[53] * kernel_shared_1[threadIdx_x * 48 + 41]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[54] * kernel_shared_1[threadIdx_x * 48 + 18]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[63] * kernel_shared_1[threadIdx_x * 48 + 21]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 48 + 18]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 48 + 21]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 18]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 21]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 18]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 21]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 18]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 21]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 18]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 21]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 18]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 21]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[54] * kernel_shared_1[threadIdx_x * 48 + 42]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[63] * kernel_shared_1[threadIdx_x * 48 + 45]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 48 + 42]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 48 + 45]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 42]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 45]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 42]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 45]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 42]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 45]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 42]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 45]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 42]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 45]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 48 + 19]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 48 + 22]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 19]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 22]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 19]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 22]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 19]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 22]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 19]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 22]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 19]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 22]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 48 + 19]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 48 + 22]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 48 + 43]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 48 + 46]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 43]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 46]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 43]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 46]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 43]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 46]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 43]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 46]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 43]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 46]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 48 + 43]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 48 + 46]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 20]
+                conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 23]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 20]
+                conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 23]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 20]
+                conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 23]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 20]
+                conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 23]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 20]
+                conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 23]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 48 + 20]
+                conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 48 + 23]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[62] * kernel_shared_1[threadIdx_x * 48 + 20]
+                conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[71] * kernel_shared_1[threadIdx_x * 48 + 23]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 44]
+                conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 47]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 44]
+                conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 47]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 44]
+                conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 47]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 44]
+                conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 47]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 44]
+                conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 47]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 48 + 44]
+                conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 48 + 47]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[62] * kernel_shared_1[threadIdx_x * 48 + 44]
+                conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[71] * kernel_shared_1[threadIdx_x * 48 + 47]
             for i1_inner, i3_inner in T.grid(2, 7):
-                cse_var_5: T.int32 = i1_inner * 7 + i3_inner
                 compute_1 = T.Buffer((25088,), data=compute.data)
                 bias_1 = T.Buffer((512,), data=bias.data)
-                compute_1[blockIdx_x * 3136 + threadIdx_x // 7 * 98 + i1_inner * 49 + threadIdx_x % 7 * 7 + i3_inner] = T.max(conv2d_nchw_1[cse_var_5] + bias_1[blockIdx_x * 64 + threadIdx_x // 7 * 2 + i1_inner], T.float32(0))
-                compute_1[blockIdx_x * 3136 + threadIdx_x // 7 * 98 + i1_inner * 49 + threadIdx_x % 7 * 7 + i3_inner + 1568] = T.max(conv2d_nchw_1[cse_var_5 + 14] + bias_1[blockIdx_x * 64 + threadIdx_x // 7 * 2 + i1_inner + 32], T.float32(0))
+                compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 98 + i1_inner * 49 + blockIdx_x % 7 * 7 + i3_inner] = T.max(conv2d_nchw_1[i1_inner * 7 + i3_inner] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x * 2 + i1_inner], T.float32(0))
 
 
 
@@ -1055,7 +765,7 @@ We build the binary and check its correctness and performance.
 
  .. code-block:: none
 
-    Execution time of this operator: 0.381 ms
+    Execution time of this operator: 0.362 ms
 
 
 
@@ -1105,18 +815,18 @@ 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=2)
-    conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=16)
-    conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=2)
+    conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=64)
+    conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=1)
     conv2d_nchw_yy_o_i, conv2d_nchw_yy_i = s[conv2d_nchw].split(conv2d_nchw_yy, factor=1)
     conv2d_nchw_yy_o_o_i, conv2d_nchw_yy_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_i, factor=1)
-    conv2d_nchw_yy_o_o_o_i, conv2d_nchw_yy_o_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_o_i, factor=7)
+    conv2d_nchw_yy_o_o_o_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_o_i, conv2d_nchw_xx_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_i, factor=1)
     conv2d_nchw_xx_o_o_o_o, conv2d_nchw_xx_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_o_i, factor=1)
-    conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=4)
-    conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=2)
+    conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=2)
+    conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=4)
     conv2d_nchw_ry_o_i, conv2d_nchw_ry_i = s[conv2d_nchw].split(conv2d_nchw_ry, factor=1)
     conv2d_nchw_ry_o_o, conv2d_nchw_ry_o_i = s[conv2d_nchw].split(conv2d_nchw_ry_o_i, factor=1)
     conv2d_nchw_rx_o_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=1)
@@ -1126,10 +836,10 @@ 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=2)
-    compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=16)
-    compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=2)
+    compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=64)
+    compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=1)
     compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
-    compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=7)
+    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_o_i, compute_i3_o_i = s[compute].split(compute_i3_o_i, factor=1)
@@ -1152,14 +862,14 @@ 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=112)
+    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=64)
     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)
+    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=4)
     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=112)
+    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=64)
     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", 1024)
+    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, "unroll_explicit", True)
 
     CUDA source code:
@@ -1177,743 +887,429 @@ 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__(112) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
-      float conv2d_nchw[28];
-      __shared__ float pad_temp_shared[504];
-      __shared__ float kernel_shared[1536];
+    extern "C" __global__ void __launch_bounds__(64) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
+      float conv2d_nchw[14];
+      __shared__ float pad_temp_shared[72];
+      __shared__ float kernel_shared[3072];
       conv2d_nchw[0] = 0.000000e+00f;
-      conv2d_nchw[14] = 0.000000e+00f;
       conv2d_nchw[1] = 0.000000e+00f;
-      conv2d_nchw[15] = 0.000000e+00f;
       conv2d_nchw[2] = 0.000000e+00f;
-      conv2d_nchw[16] = 0.000000e+00f;
       conv2d_nchw[3] = 0.000000e+00f;
-      conv2d_nchw[17] = 0.000000e+00f;
       conv2d_nchw[4] = 0.000000e+00f;
-      conv2d_nchw[18] = 0.000000e+00f;
       conv2d_nchw[5] = 0.000000e+00f;
-      conv2d_nchw[19] = 0.000000e+00f;
       conv2d_nchw[6] = 0.000000e+00f;
-      conv2d_nchw[20] = 0.000000e+00f;
       conv2d_nchw[7] = 0.000000e+00f;
-      conv2d_nchw[21] = 0.000000e+00f;
       conv2d_nchw[8] = 0.000000e+00f;
-      conv2d_nchw[22] = 0.000000e+00f;
       conv2d_nchw[9] = 0.000000e+00f;
-      conv2d_nchw[23] = 0.000000e+00f;
       conv2d_nchw[10] = 0.000000e+00f;
-      conv2d_nchw[24] = 0.000000e+00f;
       conv2d_nchw[11] = 0.000000e+00f;
-      conv2d_nchw[25] = 0.000000e+00f;
       conv2d_nchw[12] = 0.000000e+00f;
-      conv2d_nchw[26] = 0.000000e+00f;
       conv2d_nchw[13] = 0.000000e+00f;
-      conv2d_nchw[27] = 0.000000e+00f;
       for (int rc_outer_outer = 0; rc_outer_outer < 64; ++rc_outer_outer) {
         for (int ry_outer_outer = 0; ry_outer_outer < 3; ++ry_outer_outer) {
           __syncthreads();
-          pad_temp_shared[((int)threadIdx.x)] = (((((1 <= (((((int)threadIdx.x) % 63) / 9) + ry_outer_outer)) && ((((((int)threadIdx.x) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= (((int)threadIdx.x) % 9))) && ((((int)threadIdx.x) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + ((((int)threadIdx.x) / 9) * 7)) + (ry_outer_outer * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
-          pad_temp_shared[(((int)threadIdx.x) + 112)] = (((((1 <= ((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 4) % 9))) && (((((int)threadIdx.x) + 4) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 112) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
-          pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((1 <= ((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 8) % 9))) && (((((int)threadIdx.x) + 8) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 224) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
-          pad_temp_shared[(((int)threadIdx.x) + 336)] = (((((1 <= ((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer)) && (((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 3) % 9))) && (((((int)threadIdx.x) + 3) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 336) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
-          if (((int)threadIdx.x) < 56) {
-            pad_temp_shared[(((int)threadIdx.x) + 448)] = (((((1 <= (((((int)threadIdx.x) + 7) / 9) + ry_outer_outer)) && ((((((int)threadIdx.x) + 7) / 9) + ry_outer_outer) < 8)) && (1 <= ((((int)threadIdx.x) + 7) % 9))) && (((((int)threadIdx.x) + 7) % 9) < 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 448) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
+          if (((int)threadIdx.x) < 18) {
+            pad_temp_shared[(((int)threadIdx.x) * 4)] = (((((1 <= (ry_outer_outer + (((int)blockIdx.x) % 7))) && ((ry_outer_outer + (((int)blockIdx.x) % 7)) < 8)) && (1 <= ((((int)threadIdx.x) * 4) % 9))) && (((((int)threadIdx.x) * 4) % 9) < 8)) ? data[((((((rc_outer_outer * 392) + (((((int)threadIdx.x) * 4) / 9) * 49)) + (ry_outer_outer * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) * 4) % 9)) - 8)] : 0.000000e+00f);
           }
-          kernel_shared[((int)threadIdx.x)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
-          kernel_shared[(((((((int)threadIdx.x) + 112) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
-          kernel_shared[(((((((int)threadIdx.x) + 224) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
-          kernel_shared[(((int)threadIdx.x) + 336)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 64512)];
-          kernel_shared[(((((((int)threadIdx.x) + 448) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
-          kernel_shared[(((((((int)threadIdx.x) + 560) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
-          kernel_shared[(((int)threadIdx.x) + 672)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 129024)];
-          kernel_shared[(((((((int)threadIdx.x) + 784) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
-          kernel_shared[(((((((int)threadIdx.x) + 896) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
-          kernel_shared[(((int)threadIdx.x) + 1008)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 193536)];
-          kernel_shared[(((((((int)threadIdx.x) + 1120) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
-          kernel_shared[(((((((int)threadIdx.x) + 1232) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
-          kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 258048)];
-          if (((int)threadIdx.x) < 80) {
-            kernel_shared[(((((((int)threadIdx.x) + 1456) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          if (((int)threadIdx.x) < 18) {
+            pad_temp_shared[((((int)threadIdx.x) * 4) + 1)] = (((((1 <= (ry_outer_outer + (((int)blockIdx.x) % 7))) && ((ry_outer_outer + (((int)blockIdx.x) % 7)) < 8)) && (1 <= (((((int)threadIdx.x) * 4) + 1) % 9))) && ((((((int)threadIdx.x) * 4) + 1) % 9) < 8)) ? data[((((((rc_outer_outer * 392) + ((((((int)threadIdx.x) * 4) + 1) / 9) * 49)) + (ry_outer_outer * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((((int)threadIdx.x) * 4) + 1) % 9)) - 8)] : 0.000000e+00f);
           }
+          if (((int)threadIdx.x) < 18) {
+            pad_temp_shared[((((int)threadIdx.x) * 4) + 2)] = (((((1 <= (ry_outer_outer + (((int)blockIdx.x) % 7))) && ((ry_outer_outer + (((int)blockIdx.x) % 7)) < 8)) && (1 <= (((((int)threadIdx.x) * 4) + 2) % 9))) && ((((((int)threadIdx.x) * 4) + 2) % 9) < 8)) ? data[((((((rc_outer_outer * 392) + ((((((int)threadIdx.x) * 4) + 2) / 9) * 49)) + (ry_outer_outer * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((((int)threadIdx.x) * 4) + 2) % 9)) - 8)] : 0.000000e+00f);
+          }
+          if (((int)threadIdx.x) < 18) {
+            pad_temp_shared[((((int)threadIdx.x) * 4) + 3)] = (((((1 <= (ry_outer_outer + (((int)blockIdx.x) % 7))) && ((ry_outer_outer + (((int)blockIdx.x) % 7)) < 8)) && (1 <= (((((int)threadIdx.x) * 4) + 3) % 9))) && ((((((int)threadIdx.x) * 4) + 3) % 9) < 8)) ? data[((((((rc_outer_outer * 392) + ((((((int)threadIdx.x) * 4) + 3) / 9) * 49)) + (ry_outer_outer * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((((int)threadIdx.x) * 4) + 3) % 9)) - 8)] : 0.000000e+00f);
+          }
+          kernel_shared[((int)threadIdx.x)] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 64) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 64) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 128) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 128) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 192)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 36864)];
+          kernel_shared[(((((((int)threadIdx.x) + 256) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 256) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 320) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 320) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 384)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 73728)];
+          kernel_shared[(((((((int)threadIdx.x) + 448) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 512) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 512) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 576)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 110592)];
+          kernel_shared[(((((((int)threadIdx.x) + 640) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 640) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 704) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 704) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 768)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 147456)];
+          kernel_shared[(((((((int)threadIdx.x) + 832) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 832) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 896) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 960)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 184320)];
+          kernel_shared[(((((((int)threadIdx.x) + 1024) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1024) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 1088) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1088) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 1152)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 221184)];
+          kernel_shared[(((((((int)threadIdx.x) + 1216) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1216) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 1280) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1280) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 258048)];
+          kernel_shared[(((((((int)threadIdx.x) + 1408) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1408) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 1472) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1472) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 1536)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 294912)];
+          kernel_shared[(((((((int)threadIdx.x) + 1600) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1600) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 1664) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1664) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 1728)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 331776)];
+          kernel_shared[(((((((int)threadIdx.x) + 1792) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1792) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 1856) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1856) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 1920)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 368640)];
+          kernel_shared[(((((((int)threadIdx.x) + 1984) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1984) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 2048) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2048) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 2112)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 405504)];
+          kernel_shared[(((((((int)threadIdx.x) + 2176) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2176) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 2240) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2240) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 2304)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 442368)];
+          kernel_shared[(((((((int)threadIdx.x) + 2368) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2368) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 2432) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2432) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 2496)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 479232)];
+          kernel_shared[(((((((int)threadIdx.x) + 2560) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2560) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 2624) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2624) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 2688)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 516096)];
+          kernel_shared[(((((((int)threadIdx.x) + 2752) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2752) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 2816) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2816) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+          kernel_shared[(((int)threadIdx.x) + 2880)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 552960)];
+          kernel_shared[(((((((int)threadIdx.x) + 2944) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2944) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+          kernel_shared[(((((((int)threadIdx.x) + 3008) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3008) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
           __syncthreads();
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 63)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 63)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 126)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 126)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 63)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 63)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 126)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 126)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 71)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 71)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 134)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 134)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 71)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 71)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 134)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 134)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 315)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 315)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 441)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 441)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 315)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 315)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 441)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 441)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-          conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-          conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-          conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-          conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-          conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-          conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 323)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 323)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 449)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-          conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 449)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-          conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-          conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-          conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-          conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-          conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-          conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 323)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 323)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 449)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-          conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 449)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[0] * kernel_shared[(((int)threadIdx.x) * 48)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[9] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[1] * kernel_shared[(((int)threadIdx.x) * 48)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[2] * kernel_shared[(((int)threadIdx.x) * 48)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[3] * kernel_shared[(((int)threadIdx.x) * 48)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[4] * kernel_shared[(((int)threadIdx.x) * 48)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[5] * kernel_shared[(((int)threadIdx.x) * 48)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[6] * kernel_shared[(((int)threadIdx.x) * 48)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[0] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[9] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[1] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[1] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[1] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[8] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[17] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[8] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[17] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[18] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[27] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[18] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[27] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[26] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[35] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[26] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[35] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[36] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[45] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[36] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[45] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[44] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[53] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[44] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[53] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[54] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[63] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[54] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[63] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+          conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+          conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+          conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+          conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+          conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+          conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[62] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+          conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[71] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+          conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+          conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+          conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+          conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+          conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+          conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[62] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+          conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[71] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
         }
       }
       for (int i1_inner = 0; i1_inner < 2; ++i1_inner) {
         for (int i3_inner = 0; i3_inner < 7; ++i3_inner) {
-          compute[(((((((int)blockIdx.x) * 3136) + ((((int)threadIdx.x) / 7) * 98)) + (i1_inner * 49)) + ((((int)threadIdx.x) % 7) * 7)) + i3_inner)] = max((conv2d_nchw[((i1_inner * 7) + i3_inner)] + bias[(((((int)blockIdx.x) * 64) + ((((int)threadIdx.x) / 7) * 2)) + i1_inner)]), 0.000000e+00f);
-          compute[((((((((int)blockIdx.x) * 3136) + ((((int)threadIdx.x) / 7) * 98)) + (i1_inner * 49)) + ((((int)threadIdx.x) % 7) * 7)) + i3_inner) + 1568)] = max((conv2d_nchw[(((i1_inner * 7) + i3_inner) + 14)] + bias[((((((int)blockIdx.x) * 64) + ((((int)threadIdx.x) / 7) * 2)) + i1_inner) + 32)]), 0.000000e+00f);
+          compute[((((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 98)) + (i1_inner * 49)) + ((((int)blockIdx.x) % 7) * 7)) + i3_inner)] = max((conv2d_nchw[((i1_inner * 7) + i3_inner)] + bias[((((((int)blockIdx.x) / 7) * 128) + (((int)threadIdx.x) * 2)) + i1_inner)]), 0.000000e+00f);
         }
       }
     }
@@ -1974,7 +1370,7 @@ In the example below we resume the status and do more 5 trials.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 6 minutes  10.917 seconds)
+   **Total running time of the script:** ( 6 minutes  12.035 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 872fdf1da5..d304b127fc 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)  
-       7.8953       7.8927       7.9128       7.8804       0.0134   
+       7.8844       7.8839       7.8919       7.8774       0.0060   
                
 
 
@@ -675,7 +675,7 @@ Other Tips
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 1 minutes  7.853 seconds)
+   **Total running time of the script:** ( 1 minutes  8.262 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 2ba845901c..07d7f4ee87 100644
--- a/docs/_sources/how_to/tune_with_autoscheduler/tune_network_x86.rst.txt
+++ b/docs/_sources/how_to/tune_with_autoscheduler/tune_network_x86.rst.txt
@@ -666,7 +666,7 @@ so we can read the log file and load the best schedules.
     Evaluate inference time cost...
     Execution time summary:
      mean (ms)   median (ms)    max (ms)     min (ms)     std (ms)  
-      754.7261     756.4411     756.4888     751.2483      2.4592   
+      752.8096     752.8915     753.3811     752.1563      0.5033   
                
 
 
@@ -694,7 +694,7 @@ Other Tips
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 1 minutes  42.867 seconds)
+   **Total running time of the script:** ( 1 minutes  42.823 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 d7db3d4e80..58a450a876 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
@@ -389,27 +389,149 @@ 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 i_outer_inner in range(8):
-                    for i_inner_init, j_init in T.grid(8, 16):
-                        compute_2[i_outer_inner * 128 + i_inner_init * 16 + j_init] = T.float32(0)
-                    for elem_idx, i_inner, j in T.grid(T.Let(placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1], where={cse_var_1: i0_outer_i1_outer_fused % 32}), 8, 16):
-                        cse_var_1 = T.int32()
+            for i0_outer_i1_outer_fused in T.parallel(512):
+                compute_1 = T.allocate([128], "float32", "global")
+                compute_2 = T.Buffer((128,), data=compute_1)
+                for i_outer_inner in range(4):
+                    cse_var_2: T.int32 = i_outer_inner * 32
+                    cse_var_1: T.int32 = i0_outer_i1_outer_fused % 64 // 2
+                    compute_2[cse_var_2] = T.float32(0)
+                    compute_2[cse_var_2 + 1] = T.float32(0)
+                    compute_2[cse_var_2 + 2] = T.float32(0)
+                    compute_2[cse_var_2 + 3] = T.float32(0)
+                    compute_2[cse_var_2 + 4] = T.float32(0)
+                    compute_2[cse_var_2 + 5] = T.float32(0)
+                    compute_2[cse_var_2 + 6] = T.float32(0)
+                    compute_2[cse_var_2 + 7] = T.float32(0)
+                    compute_2[cse_var_2 + 8] = T.float32(0)
+                    compute_2[cse_var_2 + 9] = T.float32(0)
+                    compute_2[cse_var_2 + 10] = T.float32(0)
+                    compute_2[cse_var_2 + 11] = T.float32(0)
+                    compute_2[cse_var_2 + 12] = T.float32(0)
+                    compute_2[cse_var_2 + 13] = T.float32(0)
+                    compute_2[cse_var_2 + 14] = T.float32(0)
+                    compute_2[cse_var_2 + 15] = T.float32(0)
+                    compute_2[cse_var_2 + 16] = T.float32(0)
+                    compute_2[cse_var_2 + 17] = T.float32(0)
+                    compute_2[cse_var_2 + 18] = T.float32(0)
+                    compute_2[cse_var_2 + 19] = T.float32(0)
+                    compute_2[cse_var_2 + 20] = T.float32(0)
+                    compute_2[cse_var_2 + 21] = T.float32(0)
+                    compute_2[cse_var_2 + 22] = T.float32(0)
+                    compute_2[cse_var_2 + 23] = T.float32(0)
+                    compute_2[cse_var_2 + 24] = T.float32(0)
+                    compute_2[cse_var_2 + 25] = T.float32(0)
+                    compute_2[cse_var_2 + 26] = T.float32(0)
+                    compute_2[cse_var_2 + 27] = T.float32(0)
+                    compute_2[cse_var_2 + 28] = T.float32(0)
+                    compute_2[cse_var_2 + 29] = T.float32(0)
+                    compute_2[cse_var_2 + 30] = T.float32(0)
+                    compute_2[cse_var_2 + 31] = T.float32(0)
+                    for elem_idx in range(placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
                         placeholder_5 = T.Buffer((33,), "int32", data=placeholder_3.data)
-                        cse_var_2: T.int32 = i0_outer_i1_outer_fused % 32
-                        if T.likely(elem_idx < placeholder_5[cse_var_2 + 1] - placeholder_5[cse_var_2]):
-                            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)
-                            cse_var_3: T.int32 = i_outer_inner * 128 + i_inner * 16 + j
-                            compute_2[cse_var_3] = compute_2[cse_var_3] + placeholder_6[placeholder_5[cse_var_2] * 16 + elem_idx * 16 + j] * T.max(placeholder_7[i0_outer_i1_outer_fused // 32 * 16384 + i_outer_inner * 2048 + i_inner * 256 + placeholder_8[placeholder_5[cse_var_2] + elem_idx]], T.float32(0))
-                for i0_inner in range(64):
-                    cse_var_4: T.int32 = i0_outer_i1_outer_fused // 32 * 32768 + i0_inner * 512 + i0_outer_i1_outer_fused % 32 * 16
+                        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)
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            compute_2[cse_var_2] = compute_2[cse_var_2] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_3: T.int32 = cse_var_2 + 1
+                            compute_2[cse_var_3] = compute_2[cse_var_3] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 1] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_4: T.int32 = cse_var_2 + 2
+                            compute_2[cse_var_4] = compute_2[cse_var_4] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 2] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_5: T.int32 = cse_var_2 + 3
+                            compute_2[cse_var_5] = compute_2[cse_var_5] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 3] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_6: T.int32 = cse_var_2 + 4
+                            compute_2[cse_var_6] = compute_2[cse_var_6] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 4] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_7: T.int32 = cse_var_2 + 5
+                            compute_2[cse_var_7] = compute_2[cse_var_7] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 5] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_8: T.int32 = cse_var_2 + 6
+                            compute_2[cse_var_8] = compute_2[cse_var_8] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 6] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_9: T.int32 = cse_var_2 + 7
+                            compute_2[cse_var_9] = compute_2[cse_var_9] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 7] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_10: T.int32 = cse_var_2 + 8
+                            compute_2[cse_var_10] = compute_2[cse_var_10] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_11: T.int32 = cse_var_2 + 9
+                            compute_2[cse_var_11] = compute_2[cse_var_11] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 1] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_12: T.int32 = cse_var_2 + 10
+                            compute_2[cse_var_12] = compute_2[cse_var_12] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 2] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_13: T.int32 = cse_var_2 + 11
+                            compute_2[cse_var_13] = compute_2[cse_var_13] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 3] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_14: T.int32 = cse_var_2 + 12
+                            compute_2[cse_var_14] = compute_2[cse_var_14] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 4] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_15: T.int32 = cse_var_2 + 13
+                            compute_2[cse_var_15] = compute_2[cse_var_15] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 5] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_16: T.int32 = cse_var_2 + 14
+                            compute_2[cse_var_16] = compute_2[cse_var_16] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 6] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_17: T.int32 = cse_var_2 + 15
+                            compute_2[cse_var_17] = compute_2[cse_var_17] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 7] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_18: T.int32 = cse_var_2 + 16
+                            compute_2[cse_var_18] = compute_2[cse_var_18] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_19: T.int32 = cse_var_2 + 17
+                            compute_2[cse_var_19] = compute_2[cse_var_19] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 1] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_20: T.int32 = cse_var_2 + 18
+                            compute_2[cse_var_20] = compute_2[cse_var_20] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 2] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_21: T.int32 = cse_var_2 + 19
+                            compute_2[cse_var_21] = compute_2[cse_var_21] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 3] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_22: T.int32 = cse_var_2 + 20
+                            compute_2[cse_var_22] = compute_2[cse_var_22] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 4] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_23: T.int32 = cse_var_2 + 21
+                            compute_2[cse_var_23] = compute_2[cse_var_23] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 5] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_24: T.int32 = cse_var_2 + 22
+                            compute_2[cse_var_24] = compute_2[cse_var_24] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 6] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_25: T.int32 = cse_var_2 + 23
+                            compute_2[cse_var_25] = compute_2[cse_var_25] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 7] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_26: T.int32 = cse_var_2 + 24
+                            compute_2[cse_var_26] = compute_2[cse_var_26] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_27: T.int32 = cse_var_2 + 25
+                            compute_2[cse_var_27] = compute_2[cse_var_27] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 1] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_28: T.int32 = cse_var_2 + 26
+                            compute_2[cse_var_28] = compute_2[cse_var_28] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 2] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_29: T.int32 = cse_var_2 + 27
+                            compute_2[cse_var_29] = compute_2[cse_var_29] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 3] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_30: T.int32 = cse_var_2 + 28
+                            compute_2[cse_var_30] = compute_2[cse_var_30] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 4] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_31: T.int32 = cse_var_2 + 29
+                            compute_2[cse_var_31] = compute_2[cse_var_31] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 5] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_32: T.int32 = cse_var_2 + 30
+                            compute_2[cse_var_32] = compute_2[cse_var_32] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 6] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                        if T.likely(elem_idx < placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                            cse_var_33: T.int32 = cse_var_2 + 31
+                            compute_2[cse_var_33] = compute_2[cse_var_33] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 7] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                for i0_inner in range(16):
+                    cse_var_34: T.int32 = i0_outer_i1_outer_fused // 64 * 8192 + i0_inner * 512 + i0_outer_i1_outer_fused % 64 * 8
                     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 + 16] = T.max(compute_2[i0_inner * 16:i0_inner * 16 + 16] + placeholder_5[cse_var_4:cse_var_4 + 16], T.Broadcast(T.float32(0), 16))
+                    compute_3[cse_var_34:cse_var_34 + 8] = T.max(compute_2[i0_inner * 8:i0_inner * 8 + 8] + placeholder_5[cse_var_34:cse_var_34 + 8], T.Broadcast(T.float32(0), 8))
 
 
 
@@ -459,7 +581,7 @@ We build the binary and check its correctness and performance.
 
  .. code-block:: none
 
-    Execution time of this operator: 1.546 ms
+    Execution time of this operator: 1.918 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 5458b42f9c..eb9dc68393 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
 =================
-**00:43.881** total execution time for **how_to_tune_with_autotvm** files:
+**00:36.058** 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``)           | 00:43.847 | 0.0 MB |
+| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_conv2d_cuda.py` (``tune_conv2d_cuda.py``)           | 00:36.024 | 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_x86.py` (``tune_relay_x86.py``)               | 00:00.020 | 0.0 MB |
 +--------------------------------------------------------------------------------------------------+-----------+--------+
 | :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_cuda.py` (``tune_relay_cuda.py``)             | 00:00.005 | 0.0 MB |
 +--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_mobile_gpu.py` (``tune_relay_mobile_gpu.py``) | 00:00.005 | 0.0 MB |
-+--------------------------------------------------------------------------------------------------+-----------+--------+
 | :ref:`sphx_glr_how_to_tune_with_autotvm_tune_relay_arm.py` (``tune_relay_arm.py``)               | 00:00.004 | 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 98a3fc1e4e..43ca16d201 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
@@ -268,7 +268,8 @@ for this template
     waiting for device...
     device available
     Get devices for measurement successfully!
-    No: 1   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+    No: 1   GFLOPS: 142.96/142.96   result: MeasureResult(costs=(0.0016193390869565216,), error_no=MeasureErrorNo.NO_ERROR, all_cost=7.002098798751831, timestamp=1679387620.600002)        [('tile_f', [-1, 2, 8, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10364940
+    No: 2   GFLOPS: 0.00/142.96     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
@@ -390,8 +391,10 @@ for this template
       File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
       File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
         raise InstantiationError("Skipped because of invalid gpu kernel")
-    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 4, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 16, 16]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,3233414
-    No: 2   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 2, 1, 256]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 8, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9177077
+    No: 3   GFLOPS: 0.98/142.96     result: MeasureResult(costs=(0.23536286925,), error_no=MeasureErrorNo.NO_ERROR, all_cost=7.865373849868774, timestamp=1679387625.744464)        [('tile_f', [-1, 1, 1, 256]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7074976
+    No: 4   GFLOPS: 1278.20/1278.20 result: MeasureResult(costs=(0.00018111530568181819,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.7274179458618164, timestamp=1679387626.8660135)     [('tile_f', [-1, 2, 8, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8563748
+    No: 5   GFLOPS: 0.00/1278.20    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
@@ -513,8 +516,8 @@ 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, 128]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 2, 128]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1727650
-    No: 3   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 128, 2, 2]), ('tile_y', [-1, 1, 7, 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', 512), ('unroll_explicit', 0)],None,2414351
+    No: 6   GFLOPS: 0.00/1278.20    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
@@ -636,8 +639,8 @@ for this template
       File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
       File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
         raise InstantiationError("Skipped because of invalid gpu kernel")
-    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 8, 8, 2]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 2, 128]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10436882
-    No: 4   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 1, 128]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 256]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9093912
+    No: 7   GFLOPS: 0.00/1278.20    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
@@ -759,8 +762,8 @@ for this template
       File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
       File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
         raise InstantiationError("Skipped because of invalid gpu kernel")
-    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 16, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 8]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6875967
-    No: 5   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 16, 2, 2]), ('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, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,9075948
+    No: 8   GFLOPS: 0.00/1278.20    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
@@ -882,9 +885,8 @@ 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, 16]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 32]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8665085
-    No: 6   GFLOPS: 25.12/25.12     result: MeasureResult(costs=(0.009215251235294116,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.549553871154785, timestamp=1679386605.377113) [('tile_f', [-1, 2, 1, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 8, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7986716
-    No: 7   GFLOPS: 0.00/25.12      result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 128, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 16]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6518431
+    No: 9   GFLOPS: 0.00/1278.20    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
@@ -1006,8 +1008,8 @@ 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, 4, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 128]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8698036
-    No: 8   GFLOPS: 0.00/25.12      result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 32, 16, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 64]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,361059
+    No: 10  GFLOPS: 0.00/1278.20    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
@@ -1129,28 +1131,8 @@ 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, 512, 1, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 32, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1055789
-    No: 9   GFLOPS: 3.66/25.12      result: MeasureResult(costs=(0.06323308175,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.009866237640381, timestamp=1679386618.4774506)       [('tile_f', [-1, 32, 2, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 2, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,779475
-    No: 10  GFLOPS: 0.00/25.12      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, 16, 4, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4687253
-    No: 11  GFLOPS: 195.97/195.97   result: MeasureResult(costs=(0.0011813070823529412,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7828633785247803, timestamp=1679386619.329636)       [('tile_f', [-1, 1, 64, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 32, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8730354
-    No: 12  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 8, 1, 32]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 32]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,3826868
+    No: 11  GFLOPS: 0.00/1278.20    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
@@ -1272,8 +1254,8 @@ 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, 32]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 64]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7516274
-    No: 13  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 32, 1]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 16, 32]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6546582
+    No: 12  GFLOPS: 0.00/1278.20    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
@@ -1395,8 +1377,9 @@ for this template
       File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
       File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
         raise InstantiationError("Skipped because of invalid gpu kernel")
-    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 1, 4, 128]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 128]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5787975
-    No: 14  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 2, 16, 4]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 1, 128]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,10242887
+    No: 13  GFLOPS: 523.22/1278.20  result: MeasureResult(costs=(0.00044245605217391306,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.484649658203125, timestamp=1679387632.097078)       [('tile_f', [-1, 1, 8, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,819087
+    No: 14  GFLOPS: 0.00/1278.20    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
@@ -1518,8 +1501,9 @@ for this template
       File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
       File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
         raise InstantiationError("Skipped because of invalid gpu kernel")
-    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 2, 2]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 64, 4]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2024286
-    No: 15  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 2, 16, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 64, 2]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4316527
+    No: 15  GFLOPS: 35.46/1278.20   result: MeasureResult(costs=(0.006527616,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.4571244716644287, timestamp=1679387632.9939733)        [('tile_f', [-1, 2, 2, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 64, 1]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,216325
+    No: 16  GFLOPS: 0.00/1278.20    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
@@ -1641,8 +1625,8 @@ 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, 128, 1]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,7654070
-    No: 16  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 32, 8, 2]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 128, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,3384124
+    No: 17  GFLOPS: 0.00/1278.20    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
@@ -1764,8 +1748,8 @@ 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, 128, 1, 4]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 32, 16]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,4399447
-    No: 17  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 32, 1, 8]), ('tile_y', [-1, 7, 1, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 32, 16]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1108281
+    No: 18  GFLOPS: 0.00/1278.20    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
@@ -1887,10 +1871,253 @@ for this template
       File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
       File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
         raise InstantiationError("Skipped because of invalid gpu kernel")
-    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 32, 4]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 4, 8]), ('tile_ry', [-1, 3, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,5523452
-    No: 18  GFLOPS: 109.98/195.97   result: MeasureResult(costs=(0.0021049593125000003,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.143249034881592, timestamp=1679386623.6207368)       [('tile_f', [-1, 1, 8, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 128, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 0)],None,3511667
-    No: 19  GFLOPS: 116.11/195.97   result: MeasureResult(costs=(0.001993747470588235,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6696715354919434, timestamp=1679386624.47567) [('tile_f', [-1, 16, 8, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 4]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2783031
-    No: 20  GFLOPS: 129.42/195.97   result: MeasureResult(costs=(0.0017887947857142855,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.4931731224060059, timestamp=1679386625.3033702)      [('tile_f', [-1, 1, 16, 8]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 2, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 1)],None,6816638
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 2, 4, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 128, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 3, 1]), ('auto_unroll_max_step', 512), ('unroll_explicit', 0)],None,2350333
+    No: 19  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
+      File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
+        func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
+      File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
+        func = build(s, args, target=target, runtime=runtime)
+      File "/workspace/python/tvm/driver/build_module.py", line 227, in build
+        input_mod = lower(inputs, args, name=name, binds=binds)
+      File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
+        return ffi.lower_schedule(inp, args, name, binds, simple_mode)
+      File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
+      File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
+      File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
+    tvm._ffi.base.TVMError: Traceback (most recent call last):
+      24: TVMFuncCall
+            at ../src/runtime/c_runtime_api.cc:477
+      23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+            at ../include/tvm/runtime/packed_func.h:1217
+      22: Call
+            at ../include/tvm/runtime/packed_func.h:1213
+      21: operator()
+            at ../include/tvm/runtime/packed_func.h:1734
+      20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
+            at ../include/tvm/runtime/packed_func.h:1674
+      19: run<>
+            at ../include/tvm/runtime/packed_func.h:1634
+      18: run<tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1649
+      13: operator()
+            at ../src/driver/driver_api.cc:402
+      12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
+            at ../src/driver/driver_api.cc:388
+      11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
+            at ../src/driver/driver_api.cc:283
+      10: tvm::transform::Pass::operator()(tvm::IRModule) const
+            at ../src/ir/transform.cc:258
+      9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:274
+      8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:451
+      7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:274
+      6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/tir/ir/transform.cc:100
+      5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
+            at ../include/tvm/runtime/packed_func.h:1753
+      4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
+            at ../include/tvm/runtime/packed_func.h:1697
+      3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
+            at ../include/tvm/runtime/packed_func.h:1621
+      2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+            at ../include/tvm/runtime/packed_func.h:1217
+      1: Call
+            at ../include/tvm/runtime/packed_func.h:1213
+      0: operator()
+            at ../src/runtime/c_runtime_api.cc:534
+      File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
+      File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
+        raise InstantiationError("Skipped because of invalid gpu kernel")
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
+
+    Traceback (most recent call last):
+      24: TVMFuncCall
+            at ../src/runtime/c_runtime_api.cc:477
+      23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+            at ../include/tvm/runtime/packed_func.h:1217
+      22: Call
+            at ../include/tvm/runtime/packed_func.h:1213
+      21: operator()
+            at ../include/tvm/runtime/packed_func.h:1734
+      20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
+            at ../include/tvm/runtime/packed_func.h:1674
+      19: run<>
+            at ../include/tvm/runtime/packed_func.h:1634
+      18: run<tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1649
+      13: operator()
+            at ../src/driver/driver_api.cc:402
+      12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
+            at ../src/driver/driver_api.cc:388
+      11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
+            at ../src/driver/driver_api.cc:283
+      10: tvm::transform::Pass::operator()(tvm::IRModule) const
+            at ../src/ir/transform.cc:258
+      9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:274
+      8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:451
+      7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:274
+      6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/tir/ir/transform.cc:100
+      5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
+            at ../include/tvm/runtime/packed_func.h:1753
+      4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
+            at ../include/tvm/runtime/packed_func.h:1697
+      3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
+            at ../include/tvm/runtime/packed_func.h:1621
+      2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+            at ../include/tvm/runtime/packed_func.h:1217
+      1: Call
+            at ../include/tvm/runtime/packed_func.h:1213
+      0: operator()
+            at ../src/runtime/c_runtime_api.cc:534
+      File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
+      File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
+        raise InstantiationError("Skipped because of invalid gpu kernel")
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 2, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 7, 1]), ('tile_rc', [-1, 8, 32]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 0), ('unroll_explicit', 0)],None,1702372
+    No: 20  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
+      File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 592, in __call__
+        func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
+      File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 544, in _build_func_common
+        func = build(s, args, target=target, runtime=runtime)
+      File "/workspace/python/tvm/driver/build_module.py", line 227, in build
+        input_mod = lower(inputs, args, name=name, binds=binds)
+      File "/workspace/python/tvm/driver/build_module.py", line 134, in lower
+        return ffi.lower_schedule(inp, args, name, binds, simple_mode)
+      File "tvm/_ffi/_cython/./packed_func.pxi", line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
+      File "tvm/_ffi/_cython/./packed_func.pxi", line 276, in tvm._ffi._cy3.core.FuncCall
+      File "tvm/_ffi/_cython/./base.pxi", line 181, in tvm._ffi._cy3.core.CHECK_CALL
+    tvm._ffi.base.TVMError: Traceback (most recent call last):
+      24: TVMFuncCall
+            at ../src/runtime/c_runtime_api.cc:477
+      23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+            at ../include/tvm/runtime/packed_func.h:1217
+      22: Call
+            at ../include/tvm/runtime/packed_func.h:1213
+      21: operator()
+            at ../include/tvm/runtime/packed_func.h:1734
+      20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
+            at ../include/tvm/runtime/packed_func.h:1674
+      19: run<>
+            at ../include/tvm/runtime/packed_func.h:1634
+      18: run<tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1649
+      13: operator()
+            at ../src/driver/driver_api.cc:402
+      12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
+            at ../src/driver/driver_api.cc:388
+      11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
+            at ../src/driver/driver_api.cc:283
+      10: tvm::transform::Pass::operator()(tvm::IRModule) const
+            at ../src/ir/transform.cc:258
+      9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:274
+      8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:451
+      7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:274
+      6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/tir/ir/transform.cc:100
+      5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
+            at ../include/tvm/runtime/packed_func.h:1753
+      4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
+            at ../include/tvm/runtime/packed_func.h:1697
+      3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
+            at ../include/tvm/runtime/packed_func.h:1621
+      2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+            at ../include/tvm/runtime/packed_func.h:1217
+      1: Call
+            at ../include/tvm/runtime/packed_func.h:1213
+      0: operator()
+            at ../src/runtime/c_runtime_api.cc:534
+      File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
+      File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
+        raise InstantiationError("Skipped because of invalid gpu kernel")
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
+
+    Traceback (most recent call last):
+      24: TVMFuncCall
+            at ../src/runtime/c_runtime_api.cc:477
+      23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+            at ../include/tvm/runtime/packed_func.h:1217
+      22: Call
+            at ../include/tvm/runtime/packed_func.h:1213
+      21: operator()
+            at ../include/tvm/runtime/packed_func.h:1734
+      20: unpack_call<tvm::IRModule, 5, tvm::<lambda(tvm::te::Schedule, const tvm::runtime::Array<tvm::runtime::ObjectRef>&, const tvm::runtime::String&, const tvm::runtime::Map<tvm::te::Tensor, tvm::tir::Buffer>&, bool)> >
+            at ../include/tvm/runtime/packed_func.h:1674
+      19: run<>
+            at ../include/tvm/runtime/packed_func.h:1634
+      18: run<tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      17: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      16: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      15: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1634
+      14: run<tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_>
+            at ../include/tvm/runtime/packed_func.h:1649
+      13: operator()
+            at ../src/driver/driver_api.cc:402
+      12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array<tvm::runtime::ObjectRef, void> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::unordered_map<tvm::te::Tensor, tvm::tir::Buffer, std::hash<tvm::te::Tensor>, std::equal_to<tvm::te::Tensor>, std::allocator<std::pair<tvm::te::Tensor const, tvm::tir::Buffer> > > const&, tvm::GlobalVarSupply, bool)
+            at ../src/driver/driver_api.cc:388
+      11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array<tvm::transform::Pass, void>)
+            at ../src/driver/driver_api.cc:283
+      10: tvm::transform::Pass::operator()(tvm::IRModule) const
+            at ../src/ir/transform.cc:258
+      9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:274
+      8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:451
+      7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/ir/transform.cc:274
+      6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&) const
+            at ../src/tir/ir/transform.cc:100
+      5: tvm::runtime::TypedPackedFunc<tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)>::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
+            at ../include/tvm/runtime/packed_func.h:1753
+      4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher<tvm::tir::PrimFunc>::run<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::runtime::PackedFunc const&, tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&)
+            at ../include/tvm/runtime/packed_func.h:1697
+      3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()<tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext>(tvm::tir::PrimFunc&&, tvm::IRModule&&, tvm::transform::PassContext&&) const
+            at ../include/tvm/runtime/packed_func.h:1621
+      2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+            at ../include/tvm/runtime/packed_func.h:1217
+      1: Call
+            at ../include/tvm/runtime/packed_func.h:1213
+      0: operator()
+            at ../src/runtime/c_runtime_api.cc:534
+      File "tvm/_ffi/_cython/./packed_func.pxi", line 56, in tvm._ffi._cy3.core.tvm_callback
+      File "/workspace/python/tvm/autotvm/measure/measure_methods.py", line 875, in verify_pass
+        raise InstantiationError("Skipped because of invalid gpu kernel")
+    tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [('tile_f', [-1, 4, 32, 4]), ('tile_y', [-1, 1, 1, 1]), ('tile_x', [-1, 7, 1, 1]), ('tile_rc', [-1, 1, 32]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8853812
 
 
 
@@ -1945,9 +2172,9 @@ and measure running time.
     Finish loading 20 records
 
     Best config:
-    [('tile_f', [-1, 1, 64, 2]), ('tile_y', [-1, 1, 1, 7]), ('tile_x', [-1, 1, 1, 1]), ('tile_rc', [-1, 32, 1]), ('tile_ry', [-1, 1, 1]), ('tile_rx', [-1, 1, 1]), ('auto_unroll_max_step', 1500), ('unroll_explicit', 1)],None,8730354
+    [('tile_f', [-1, 2, 8, 1]), ('tile_y', [-1, 1, 7, 1]), ('tile_x', [-1, 1, 1, 7]), ('tile_rc', [-1, 4, 2]), ('tile_ry', [-1, 1, 3]), ('tile_rx', [-1, 1, 3]), ('auto_unroll_max_step', 512), ('unroll_explicit', 1)],None,8563748
     Finish loading 20 records
-    Time cost of this operator: 0.001574
+    Time cost of this operator: 0.000404
 
 
 
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 a64fa0da46..85002e1785 100644
--- a/docs/_sources/how_to/work_with_microtvm/micro_autotune.rst.txt
+++ b/docs/_sources/how_to/work_with_microtvm/micro_autotune.rst.txt
@@ -360,10 +360,10 @@ Timing the untuned program
     ########## Build without Autotuning ##########
     Node Name                                     Ops                                           Time(us)  Time(%)  Shape              Inputs  Outputs  Measurements(us)  
     ---------                                     ---                                           --------  -------  -----              ------  -------  ----------------  
-    tvmgen_default_fused_nn_contrib_conv2d_NCHWc  tvmgen_default_fused_nn_contrib_conv2d_NCHWc  315.7     98.728   (1, 2, 10, 10, 3)  2       1        [315.7]           
-    tvmgen_default_fused_layout_transform_1       tvmgen_default_fused_layout_transform_1       3.101     0.97     (1, 6, 10, 10)     1       1        [3.101]           
-    tvmgen_default_fused_layout_transform         tvmgen_default_fused_layout_transform         0.966     0.302    (1, 1, 10, 10, 3)  1       1        [0.966]           
-    Total_time                                    -                                             319.766   -        -                  -       -        -                 
+    tvmgen_default_fused_nn_contrib_conv2d_NCHWc  tvmgen_default_fused_nn_contrib_conv2d_NCHWc  314.5     98.74    (1, 2, 10, 10, 3)  2       1        [314.5]           
+    tvmgen_default_fused_layout_transform_1       tvmgen_default_fused_layout_transform_1       3.048     0.957    (1, 6, 10, 10)     1       1        [3.048]           
+    tvmgen_default_fused_layout_transform         tvmgen_default_fused_layout_transform         0.966     0.303    (1, 1, 10, 10, 3)  1       1        [0.966]           
+    Total_time                                    -                                             318.514   -        -                  -       -        -                 
 
 
 
@@ -428,10 +428,10 @@ Timing the tuned program
     ########## Build with Autotuning ##########
     Node Name                                     Ops                                           Time(us)  Time(%)  Shape              Inputs  Outputs  Measurements(us)  
     ---------                                     ---                                           --------  -------  -----              ------  -------  ----------------  
-    tvmgen_default_fused_nn_contrib_conv2d_NCHWc  tvmgen_default_fused_nn_contrib_conv2d_NCHWc  103.5     97.496   (1, 6, 10, 10, 1)  2       1        [103.5]           
-    tvmgen_default_fused_layout_transform_1       tvmgen_default_fused_layout_transform_1       1.798     1.693    (1, 6, 10, 10)     1       1        [1.798]           
-    tvmgen_default_fused_layout_transform         tvmgen_default_fused_layout_transform         0.86      0.81     (1, 3, 10, 10, 1)  1       1        [0.86]            
-    Total_time                                    -                                             106.158   -        -                  -       -        -                 
+    tvmgen_default_fused_nn_contrib_conv2d_NCHWc  tvmgen_default_fused_nn_contrib_conv2d_NCHWc  105.0     97.564   (1, 6, 10, 10, 1)  2       1        [105.0]           
+    tvmgen_default_fused_layout_transform_1       tvmgen_default_fused_layout_transform_1       1.78      1.654    (1, 6, 10, 10)     1       1        [1.78]            
+    tvmgen_default_fused_layout_transform         tvmgen_default_fused_layout_transform         0.842     0.782    (1, 3, 10, 10, 1)  1       1        [0.842]           
+    Total_time                                    -                                             107.622   -        -                  -       -        -                 
 
 
 
@@ -439,7 +439,7 @@ Timing the tuned program
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 1 minutes  21.994 seconds)
+   **Total running time of the script:** ( 1 minutes  22.179 seconds)
 
 
 .. _sphx_glr_download_how_to_work_with_microtvm_micro_autotune.py:
diff --git a/docs/_sources/how_to/work_with_microtvm/micro_pytorch.rst.txt b/docs/_sources/how_to/work_with_microtvm/micro_pytorch.rst.txt
index 412b58948d..3487d1147d 100644
--- a/docs/_sources/how_to/work_with_microtvm/micro_pytorch.rst.txt
+++ b/docs/_sources/how_to/work_with_microtvm/micro_pytorch.rst.txt
@@ -118,7 +118,7 @@ download a cat image and preprocess it to use as the model input.
     /venv/apache-tvm-py3.7/lib/python3.7/site-packages/torch/ao/quantization/utils.py:281: UserWarning: must run observer before calling calculate_qparams. Returning default values.
       "must run observer before calling calculate_qparams. " +
     Downloading: "https://download.pytorch.org/models/quantized/mobilenet_v2_qnnpack_37f702c5.pth" to /workspace/.cache/torch/hub/checkpoints/mobilenet_v2_qnnpack_37f702c5.pth
-
      0%|          | 0.00/3.42M [00:00<?, ?B/s]
    100%|##########| 3.42M/3.42M [00:00<00:00, 53.9MB/s]
+
      0%|          | 0.00/3.42M [00:00<?, ?B/s]
     61%|######    | 2.09M/3.42M [00:00<00:00, 11.9MB/s]
    100%|##########| 3.42M/3.42M [00:00<00:00, 18.6MB/s]
     /workspace/python/tvm/relay/frontend/pytorch_utils.py:47: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
       return LooseVersion(torch_ver) > ver
     /venv/apache-tvm-py3.7/lib/python3.7/site-packages/setuptools/_distutils/version.py:346: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
@@ -324,7 +324,7 @@ Look up prediction top 1 index in 1000 class synset.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 1 minutes  18.461 seconds)
+   **Total running time of the script:** ( 1 minutes  19.009 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 6d5d762158..33faa58a29 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/tmp_rfyuq_w/images/random'
+    '/tmp/tmpn9bjfr1u/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], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0], [1.0, 0.0]
+   :alt: [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0]
    :srcset: /how_to/work_with_microtvm/images/sphx_glr_micro_train_001.png
    :class: sphx-glr-single-img
 
@@ -318,8 +318,8 @@ objects to other stuff? We can display some examples from our datasets using ``m
 
  .. code-block:: none
 
-    /tmp/tmp_rfyuq_w/images/target contains 8144 images
-    /tmp/tmp_rfyuq_w/images/random contains 5000 images
+    /tmp/tmpn9bjfr1u/images/target contains 8144 images
+    /tmp/tmpn9bjfr1u/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: 0.2213 - accuracy: 0.9245 - val_loss: 0.1149 - val_accuracy: 0.9588 - 47s/epoch - 145ms/step
+    328/328 - 47s - loss: 0.2282 - accuracy: 0.9221 - val_loss: 0.1165 - val_accuracy: 0.9592 - 47s/epoch - 144ms/step
     Epoch 2/3
-    328/328 - 43s - loss: 0.1003 - accuracy: 0.9627 - val_loss: 0.0947 - val_accuracy: 0.9649 - 43s/epoch - 132ms/step
+    328/328 - 44s - loss: 0.0989 - accuracy: 0.9645 - val_loss: 0.1235 - val_accuracy: 0.9600 - 44s/epoch - 133ms/step
     Epoch 3/3
-    328/328 - 43s - loss: 0.0658 - accuracy: 0.9743 - val_loss: 0.0970 - val_accuracy: 0.9656 - 43s/epoch - 131ms/step
+    328/328 - 43s - loss: 0.0618 - accuracy: 0.9753 - val_loss: 0.1142 - val_accuracy: 0.9683 - 43s/epoch - 132ms/step
 
-    <keras.callbacks.History object at 0x7fa816b52a10>
+    <keras.callbacks.History object at 0x7f68b6ad09d0>
 
 
 
@@ -861,7 +861,7 @@ Arduino tutorial for how to do that `on GitHub <https://github.com/guberti/tvm-a
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 4 minutes  33.636 seconds)
+   **Total running time of the script:** ( 4 minutes  34.489 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 c3ed60f436..74be003afb 100644
--- a/docs/_sources/how_to/work_with_microtvm/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/work_with_microtvm/sg_execution_times.rst.txt
@@ -5,20 +5,20 @@
 
 Computation times
 =================
-**07:39.956** total execution time for **how_to_work_with_microtvm** files:
+**07:41.083** total execution time for **how_to_work_with_microtvm** files:
 
 +-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``)           | 04:33.636 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_train.py` (``micro_train.py``)           | 04:34.489 | 0.0 MB |
 +-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``)     | 01:21.994 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_autotune.py` (``micro_autotune.py``)     | 01:22.179 | 0.0 MB |
 +-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_pytorch.py` (``micro_pytorch.py``)       | 01:18.461 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_pytorch.py` (``micro_pytorch.py``)       | 01:19.009 | 0.0 MB |
 +-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``)               | 00:10.341 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_aot.py` (``micro_aot.py``)               | 00:10.186 | 0.0 MB |
 +-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_custom_ide.py` (``micro_custom_ide.py``) | 00:08.017 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_custom_ide.py` (``micro_custom_ide.py``) | 00:07.951 | 0.0 MB |
 +-----------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``)         | 00:07.507 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_microtvm_micro_tflite.py` (``micro_tflite.py``)         | 00:07.270 | 0.0 MB |
 +-----------------------------------------------------------------------------------------+-----------+--------+
 | :ref:`sphx_glr_how_to_work_with_microtvm_micro_ethosu.py` (``micro_ethosu.py``)         | 00:00.000 | 0.0 MB |
 +-----------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt b/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt
index 29622ad0b1..3ed742711b 100644
--- a/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/work_with_relay/sg_execution_times.rst.txt
@@ -5,14 +5,14 @@
 
 Computation times
 =================
-**00:46.763** total execution time for **how_to_work_with_relay** files:
+**00:46.557** total execution time for **how_to_work_with_relay** files:
 
 +----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_using_pipeline_executor.py` (``using_pipeline_executor.py``) | 00:34.578 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_pipeline_executor.py` (``using_pipeline_executor.py``) | 00:34.426 | 0.0 MB |
 +----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``)           | 00:10.588 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_using_external_lib.py` (``using_external_lib.py``)           | 00:10.330 | 0.0 MB |
 +----------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``)                             | 00:01.591 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_relay_build_gcn.py` (``build_gcn.py``)                             | 00:01.795 | 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 d1e00389ec..1e97f93f34 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 0x7fa6bc15f290>
+    <function my_cuda_math_rule at 0x7f67574853b0>
 
 
 
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 18be2b7431..c8b17e6fd0 100644
--- a/docs/_sources/how_to/work_with_schedules/sg_execution_times.rst.txt
+++ b/docs/_sources/how_to/work_with_schedules/sg_execution_times.rst.txt
@@ -5,20 +5,20 @@
 
 Computation times
 =================
-**00:06.608** total execution time for **how_to_work_with_schedules** files:
+**00:08.138** total execution time for **how_to_work_with_schedules** files:
 
 +------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``)                 | 00:04.026 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_intrin_math.py` (``intrin_math.py``)                 | 00:05.598 | 0.0 MB |
 +------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``)                     | 00:01.177 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tensorize.py` (``tensorize.py``)                     | 00:01.130 | 0.0 MB |
 +------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``)                     | 00:00.582 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_reduction.py` (``reduction.py``)                     | 00:00.588 | 0.0 MB |
 +------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``)                               | 00:00.565 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_scan.py` (``scan.py``)                               | 00:00.566 | 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``)                     | 00:00.118 | 0.0 MB |
 +------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_how_to_work_with_schedules_tedd.py` (``tedd.py``)                               | 00:00.060 | 0.0 MB |
+| :ref:`sphx_glr_how_to_work_with_schedules_tedd.py` (``tedd.py``)                               | 00:00.059 | 0.0 MB |
 +------------------------------------------------------------------------------------------------+-----------+--------+
 | :ref:`sphx_glr_how_to_work_with_schedules_schedule_primitives.py` (``schedule_primitives.py``) | 00:00.052 | 0.0 MB |
 +------------------------------------------------------------------------------------------------+-----------+--------+
diff --git a/docs/_sources/topic/vta/tutorials/autotvm/sg_execution_times.rst.txt b/docs/_sources/topic/vta/tutorials/autotvm/sg_execution_times.rst.txt
index f958f39780..d8e490a6e2 100644
--- a/docs/_sources/topic/vta/tutorials/autotvm/sg_execution_times.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/autotvm/sg_execution_times.rst.txt
@@ -5,10 +5,10 @@
 
 Computation times
 =================
-**00:31.679** total execution time for **topic_vta_tutorials_autotvm** files:
+**00:31.586** total execution time for **topic_vta_tutorials_autotvm** files:
 
 +---------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:31.672 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_autotvm_tune_relay_vta.py` (``tune_relay_vta.py``) | 00:31.580 | 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 418c3e901b..1b06957336 100644
--- a/docs/_sources/topic/vta/tutorials/frontend/deploy_classification.rst.txt
+++ b/docs/_sources/topic/vta/tutorials/frontend/deploy_classification.rst.txt
@@ -293,7 +293,7 @@ The compilation steps are:
       DeprecationWarning,
     /workspace/vta/tutorials/frontend/deploy_classification.py:213: DeprecationWarning: legacy graph executor behavior of producing json / lib / params will be removed in the next release. Please see documents of tvm.contrib.graph_executor.GraphModule for the  new recommended usage.
       relay_prog, target=tvm.target.Target(target, host=env.target_host), params=params
-    resnet18_v1 inference graph built in 33.80s!
+    resnet18_v1 inference graph built in 34.03s!
 
 
 
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 c78ea3cca1..8c15b27e3a 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 23.18s!
+    yolov3-tiny inference graph built in 23.13s!
 
 
 
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 930b682903..1ca376136b 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
 =================
-**01:40.787** total execution time for **topic_vta_tutorials_frontend** files:
+**01:40.827** total execution time for **topic_vta_tutorials_frontend** files:
 
 +------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:50.648 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_classification.py` (``deploy_classification.py``) | 00:50.842 | 0.0 MB |
 +------------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``)           | 00:50.139 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_frontend_deploy_detection.py` (``deploy_detection.py``)           | 00:49.986 | 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 0202534ed7..f238933869 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
 =================
-**00:03.178** total execution time for **topic_vta_tutorials_optimize** files:
+**00:03.162** total execution time for **topic_vta_tutorials_optimize** files:
 
 +--------------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_optimize_convolution_opt.py` (``convolution_opt.py``)         | 00:02.707 | 0.0 MB |
+| :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_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.470 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_optimize_matrix_multiply_opt.py` (``matrix_multiply_opt.py``) | 00:00.471 | 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 138fffc4f2..c426de2563 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
 =================
-**00:00.797** total execution time for **topic_vta_tutorials** files:
+**00:00.794** total execution time for **topic_vta_tutorials** files:
 
 +---------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.414 | 0.0 MB |
+| :ref:`sphx_glr_topic_vta_tutorials_matrix_multiply.py` (``matrix_multiply.py``) | 00:00.411 | 0.0 MB |
 +---------------------------------------------------------------------------------+-----------+--------+
 | :ref:`sphx_glr_topic_vta_tutorials_vta_get_started.py` (``vta_get_started.py``) | 00:00.383 | 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 de40edf870..b55f64bbc2 100644
--- a/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
+++ b/docs/_sources/tutorial/auto_scheduler_matmul_x86.rst.txt
@@ -207,13 +207,6 @@ trials, we can load the best schedule from the log file and apply it.
 
 
 
-.. rst-class:: sphx-glr-script-out
-
- .. code-block:: none
-
-    .T
-
-
 
 
 
@@ -325,7 +318,7 @@ We build the binary and check its correctness and performance.
 
  .. code-block:: none
 
-    Execution time of this operator: 95.259 ms
+    Execution time of this operator: 97.437 ms
 
 
 
@@ -423,7 +416,7 @@ resume the status and do more 5 trials.
  .. code-block:: none
 
     Resume search:
-    *E
+
 
 
 
@@ -441,7 +434,7 @@ operations.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 1 minutes  47.961 seconds)
+   **Total running time of the script:** ( 1 minutes  33.638 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 62bd6363ee..019946c96f 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: 1.88/1.88       result: MeasureResult(costs=(0.14288205199999998,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.5259623527526855, timestamp=1679384947.1457846)        [('tile_y', [-1, 1]), ('tile_x', [-1, 4])],None,20
-    No: 2   GFLOPS: 3.32/3.32       result: MeasureResult(costs=(0.0807752306,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.5558521747589111, timestamp=1679384949.9663463)       [('tile_y', [-1, 32]), ('tile_x', [-1, 8])],None,35
-    No: 3   GFLOPS: 13.11/13.11     result: MeasureResult(costs=(0.020481232600000003,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.643653392791748, timestamp=1679384950.5591009)        [('tile_y', [-1, 256]), ('tile_x', [-1, 128])],None,78
-    No: 4   GFLOPS: 14.39/14.39     result: MeasureResult(costs=(0.0186509502,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5796802043914795, timestamp=1679384952.3880312)       [('tile_y', [-1, 32]), ('tile_x', [-1, 64])],None,65
-    No: 5   GFLOPS: 3.93/14.39      result: MeasureResult(costs=(0.0683804154,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.3462045192718506, timestamp=1679384953.8633971)       [('tile_y', [-1, 16]), ('tile_x', [-1, 8])],None,34
-    No: 6   GFLOPS: 2.11/14.39      result: MeasureResult(costs=(0.12726343639999999,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.294757843017578, timestamp=1679384957.439452)  [('tile_y', [-1, 128]), ('tile_x', [-1, 4])],None,27
-    No: 7   GFLOPS: 11.22/14.39     result: MeasureResult(costs=(0.0239204612,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6556861400604248, timestamp=1679384958.087621)        [('tile_y', [-1, 128]), ('tile_x', [-1, 256])],None,87
-    No: 8   GFLOPS: 2.09/14.39      result: MeasureResult(costs=(0.1285214744,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.2971298694610596, timestamp=1679384960.3970604)       [('tile_y', [-1, 256]), ('tile_x', [-1, 4])],None,28
-    No: 9   GFLOPS: 1.36/14.39      result: MeasureResult(costs=(0.1976289706,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.3668715953826904, timestamp=1679384963.8987393)       [('tile_y', [-1, 2]), ('tile_x', [-1, 1])],None,1
-    No: 10  GFLOPS: 11.42/14.39     result: MeasureResult(costs=(0.0235062912,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.598099946975708, timestamp=1679384964.5302613)        [('tile_y', [-1, 128]), ('tile_x', [-1, 32])],None,57
+    No: 1   GFLOPS: 4.42/4.42       result: MeasureResult(costs=(0.060736017399999995,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.2457153797149658, timestamp=1679385982.5549886)       [('tile_y', [-1, 16]), ('tile_x', [-1, 16])],None,44
+    No: 2   GFLOPS: 12.93/12.93     result: MeasureResult(costs=(0.0207570296,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.5916378498077393, timestamp=1679385983.1463542)       [('tile_y', [-1, 256]), ('tile_x', [-1, 64])],None,68
+    No: 3   GFLOPS: 11.08/12.93     result: MeasureResult(costs=(0.024226987600000003,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.633718729019165, timestamp=1679385985.0716083)        [('tile_y', [-1, 64]), ('tile_x', [-1, 256])],None,86
+    No: 4   GFLOPS: 10.48/12.93     result: MeasureResult(costs=(0.0256093362,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.68430495262146, timestamp=1679385987.000635)  [('tile_y', [-1, 512]), ('tile_x', [-1, 64])],None,69
+    No: 5   GFLOPS: 11.85/12.93     result: MeasureResult(costs=(0.022652811399999997,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6896207332611084, timestamp=1679385989.0740974)       [('tile_y', [-1, 64]), ('tile_x', [-1, 32])],None,56
+    No: 6   GFLOPS: 10.32/12.93     result: MeasureResult(costs=(0.026005339600000003,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6879167556762695, timestamp=1679385989.7467787)       [('tile_y', [-1, 8]), ('tile_x', [-1, 32])],None,53
+    No: 7   GFLOPS: 11.43/12.93     result: MeasureResult(costs=(0.0234760276,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6295592784881592, timestamp=1679385990.382936)        [('tile_y', [-1, 4]), ('tile_x', [-1, 512])],None,92
+    No: 8   GFLOPS: 11.52/12.93     result: MeasureResult(costs=(0.023292248,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6384110450744629, timestamp=1679385991.0148263)        [('tile_y', [-1, 8]), ('tile_x', [-1, 256])],None,83
+    No: 9   GFLOPS: 10.18/12.93     result: MeasureResult(costs=(0.026363145600000006,), error_no=MeasureErrorNo.NO_ERROR, all_cost=0.6564569473266602, timestamp=1679385991.7865014)       [('tile_y', [-1, 4]), ('tile_x', [-1, 32])],None,52
+    No: 10  GFLOPS: 2.07/12.93      result: MeasureResult(costs=(0.1297377124,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.291069746017456, timestamp=1679385994.1183388)        [('tile_y', [-1, 1]), ('tile_x', [-1, 4])],None,20
 
 
 
diff --git a/docs/_sources/tutorial/autotvm_relay_x86.rst.txt b/docs/_sources/tutorial/autotvm_relay_x86.rst.txt
index 2424ea2cca..a1157ce1d0 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': 517.4032972500004, 'median': 518.1137195500014, 'std': 1.692627985874011}
+    {'mean': 514.8689825700013, 'median': 514.807255200003, 'std': 1.1113574405842}
 
 
 
@@ -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:    6.27/  23.15 GFLOPS | Progress: (4/20) | 9.74 s
    [Task  1/25]  Current/Best:    9.91/  23.15 GFLOPS | Progress: (8/20) | 15.24 s
    [Task  1/25]  Current/Best:   17.57/  23.15 GFLOPS | Progress: (12/20) | 18.07 s
    [Task  1/25]  Current/Best:    9.21/  23.15 GFLOPS | Progress: (16/20) | 20.70 s
    [Task  1/25]  Current/Best:   21.93/  23.15 GFLOPS | Progress: (20/20) | 23.40 s Done.
-
    [Task  2/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  2/25]  Current/Best:   17.87/  17.87 GFLOPS | Progress: (4/20) | 4.71 s
    [Task  2/25]  Current/Best:   12.10/  17.87 GFLOPS | Progress: (8/20) | 7.08 s
    [Task  2/25]  Current/Best:    9.24/  17.87 GFLOPS | Progress: (12/20) | 9.14 s
    [Task  2/25]  Current/Best:   10.71/  22.06 GFLOPS | Progress: (16/20) | 11.06 s
    [Task  2/25]  Current/Best:   13.15/  22.06 GFLOPS | Progress: (20/20) | 13.29 s Done.
-
    [Task  3/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  3/25]  Current/Best:   17.75/  17.75 GFLOPS | Progress: (4/20) | 5.78 s
    [Task  3/25]  Current/Best:   21.41/  21.41 GFLOPS | Progress: (8/20) | 7.88 s
    [Task  3/25]  Current/Best:    6.83/  21.41 GFLOPS | Progress: (12/20) | 10.05 s
    [Task  3/25]  Current/Best:    6.97/  23.46 GFLOPS | Progress: (16/20) | 13.14 s
    [Task  3/25]  Current/Best:   15.84/  23.46 GFLOPS | Progress: (20/20) | 15.17 s Done.
-
    [Task  4/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  4/25]  Current/Best:   18.26/  18.26 GFLOPS | Progress: (4/20) | 11.05 s
    [Task  4/25]  Current/Best:   13.97/  18.26 GFLOPS | Progress: (8/20) | 15.57 s
    [Task  4/25]  Current/Best:   11.97/  18.26 GFLOPS | Progress: (12/20) | 18.78 s
    [Task  4/25]  Current/Best:   21.60/  21.60 GFLOPS | Progress: (16/20) | 20.96 s
    [Task  4/25]  Current/Best:   17.00/  21.60 GFLOPS | Progress: (20/20) | 23.46 s Done.
-
    [Task  5/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  5/25]  Current/Best:    1.60/  15.46 GFLOPS | Progress: (4/20) | 5.61 s
    [Task  5/25]  Current/Best:    9.68/  22.81 GFLOPS | Progress: (8/20) | 7.43 s
    [Task  5/25]  Current/Best:   11.33/  22.81 GFLOPS | Progress: (12/20) | 9.61 s
    [Task  5/25]  Current/Best:   12.35/  22.81 GFLOPS | Progress: (16/20) | 12.18 s
    [Task  5/25]  Current/Best:   10.93/  22.81 GFLOPS | Progress: (20/20) | 14.55 s Done.
-
    [Task  6/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  6/25]  Current/Best:    5.40/  15.21 GFLOPS | Progress: (4/20) | 5.81 s
    [Task  6/25]  Current/Best:    8.13/  15.21 GFLOPS | Progress: (8/20) | 9.12 s
    [Task  6/25]  Current/Best:   10.99/  15.21 GFLOPS | Progress: (12/20) | 12.19 s
    [Task  6/25]  Current/Best:   15.58/  15.58 GFLOPS | Progress: (16/20) | 15.83 s
    [Task  6/25]  Current/Best:   18.55/  20.48 GFLOPS | Progress: (20/20) | 17.69 s Done.
-
    [Task  7/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  7/25]  Current/Best:    8.57/  12.33 GFLOPS | Progress: (4/20) | 5.58 s
    [Task  7/25]  Current/Best:    7.37/  19.47 GFLOPS | Progress: (8/20) | 8.18 s
    [Task  7/25]  Current/Best:   11.72/  19.47 GFLOPS | Progress: (12/20) | 12.76 s
    [Task  7/25]  Current/Best:   11.36/  19.67 GFLOPS | Progress: (16/20) | 14.75 s
    [Task  7/25]  Current/Best:   19.03/  19.67 GFLOPS | Progress: (20/20) | 17.33 s Done.
-
    [Task  8/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  8/25]  Current/Best:   12.29/  12.34 GFLOPS | Progress: (4/20) | 7.42 s
    [Task  8/25]  Current/Best:    3.75/  16.17 GFLOPS | Progress: (8/20) | 10.07 s
    [Task  8/25]  Current/Best:   10.11/  16.17 GFLOPS | Progress: (12/20) | 18.57 s
    [Task  8/25]  Current/Best:    9.37/  16.17 GFLOPS | Progress: (16/20) | 26.49 s
    [Task  8/25]  Current/Best:   17.72/  17.72 GFLOPS | Progress: (20/20) | 34.24 s Done.
-
    [Task  9/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  9/25]  Current/Best:   20.83/  22.67 GFLOPS | Progress: (4/20) | 13.69 s
    [Task  9/25]  Current/Best:   13.85/  22.67 GFLOPS | Progress: (8/20) | 16.10 s
    [Task  9/25]  Current/Best:   19.10/  22.67 GFLOPS | Progress: (12/20) | 17.73 s
    [Task  9/25]  Current/Best:    7.67/  22.67 GFLOPS | Progress: (16/20) | 21.18 s
    [Task  9/25]  Current/Best:   14.25/  22.67 GFLOPS | Progress: (20/20) | 26.82 s Done.
-
    [Task 10/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 10/25]  Current/Best:    9.99/  18.14 GFLOPS | Progress: (4/20) | 4.66 s
    [Task 10/25]  Current/Best:   16.37/  18.61 GFLOPS | Progress: (8/20) | 6.59 s
    [Task 10/25]  Current/Best:   15.11/  18.61 GFLOPS | Progress: (12/20) | 8.16 s
    [Task 10/25]  Current/Best:   20.75/  20.75 GFLOPS | Progress: (16/20) | 10.31 s
    [Task 10/25]  Current/Best:   15.88/  20.75 GFLOPS | Progress: (20/20) | 14.46 s Done.
-
    [Task 11/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 11/25]  Current/Best:   20.45/  20.45 GFLOPS | Progress: (4/20) | 5.92 s
    [Task 11/25]  Current/Best:   11.01/  20.45 GFLOPS | Progress: (8/20) | 10.20 s
    [Task 11/25]  Current/Best:   19.16/  20.82 GFLOPS | Progress: (12/20) | 12.82 s
    [Task 11/25]  Current/Best:    5.68/  20.82 GFLOPS | Progress: (16/20) | 15.68 s
    [Task 11/25]  Current/Best:   10.23/  20.82 GFLOPS | Progress: (20/20) | 18.16 s Done.
-
    [Task 12/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 12/25]  Current/Best:   15.31/  18.21 GFLOPS | Progress: (4/20) | 5.73 s
    [Task 12/25]  Current/Best:   20.56/  20.56 GFLOPS | Progress: (8/20) | 14.21 s
    [Task 12/25]  Current/Best:    9.32/  21.36 GFLOPS | Progress: (12/20) | 19.85 s
    [Task 12/25]  Current/Best:    9.82/  21.36 GFLOPS | Progress: (16/20) | 23.25 s
    [Task 12/25]  Current/Best:    8.50/  21.36 GFLOPS | Progress: (20/20) | 27.49 s Done.
-
    [Task 13/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 13/25]  Current/Best:   12.25/  16.80 GFLOPS | Progress: (4/20) | 4.97 s
    [Task 13/25]  Current/Best:   20.35/  22.35 GFLOPS | Progress: (8/20) | 7.71 s
    [Task 13/25]  Current/Best:   13.17/  22.35 GFLOPS | Progress: (12/20) | 10.99 s
    [Task 13/25]  Current/Best:   14.14/  22.35 GFLOPS | Progress: (16/20) | 14.41 s
    [Task 13/25]  Current/Best:   10.05/  22.35 GFLOPS | Progress: (20/20) | 17.53 s Done.
-
    [Task 14/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 14/25]  Current/Best:   18.24/  18.24 GFLOPS | Progress: (4/20) | 5.00 s
    [Task 14/25]  Current/Best:   18.58/  18.58 GFLOPS | Progress: (8/20) | 9.49 s
    [Task 14/25]  Current/Best:   10.98/  19.14 GFLOPS | Progress: (12/20) | 14.42 s
    [Task 14/25]  Current/Best:   18.94/  19.14 GFLOPS | Progress: (16/20) | 18.58 s
    [Task 14/25]  Current/Best:   12.40/  19.14 GFLOPS | Progress: (20/20) | 28.09 s
    [Task 15/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s Done.
-
    [Task 15/25]  Current/Best:   15.83/  17.23 GFLOPS | Progress: (4/20) | 6.57 s
    [Task 15/25]  Current/Best:   18.13/  18.13 GFLOPS | Progress: (8/20) | 8.59 s
    [Task 15/25]  Current/Best:    9.25/  18.13 GFLOPS | Progress: (12/20) | 12.32 s
    [Task 15/25]  Current/Best:    8.32/  18.13 GFLOPS | Progress: (16/20) | 15.74 s
    [Task 15/25]  Current/Best:   13.33/  18.13 GFLOPS | Progress: (20/20) | 17.73 s Done.
-
    [Task 16/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 16/25]  Current/Best:   13.56/  20.81 GFLOPS | Progress: (4/20) | 4.33 s
    [Task 16/25]  Current/Best:   11.63/  20.81 GFLOPS | Progress: (8/20) | 6.53 s
    [Task 16/25]  Current/Best:   14.23/  20.81 GFLOPS | Progress: (12/20) | 8.45 s
    [Task 16/25]  Current/Best:    6.07/  20.81 GFLOPS | Progress: (16/20) | 10.32 s
    [Task 16/25]  Current/Best:   14.99/  20.81 GFLOPS | Progress: (20/20) | 12.14 s Done.
-
    [Task 17/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 17/25]  Current/Best:   20.24/  20.24 GFLOPS | Progress: (4/20) | 5.36 s
    [Task 17/25]  Current/Best:    6.20/  22.75 GFLOPS | Progress: (8/20) | 7.65 s
    [Task 17/25]  Current/Best:   11.71/  22.75 GFLOPS | Progress: (12/20) | 11.19 s
    [Task 17/25]  Current/Best:   11.62/  22.75 GFLOPS | Progress: (16/20) | 13.87 s
    [Task 17/25]  Current/Best:   11.68/  22.75 GFLOPS | Progress: (20/20) | 16.63 s Done.
-
    [Task 18/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 18/25]  Current/Best:   13.92/  13.92 GFLOPS | Progress: (4/20) | 5.67 s
    [Task 18/25]  Current/Best:   12.07/  17.52 GFLOPS | Progress: (8/20) | 8.01 s
    [Task 18/25]  Current/Best:    7.86/  17.52 GFLOPS | Progress: (12/20) | 11.91 s
    [Task 18/25]  Current/Best:   12.22/  17.52 GFLOPS | Progress: (16/20) | 15.32 s
    [Task 18/25]  Current/Best:   10.23/  17.52 GFLOPS | Progress: (20/20) | 19.64 s Done.
-
    [Task 19/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 19/25]  Current/Best:   12.01/  20.07 GFLOPS | Progress: (4/20) | 7.34 s
    [Task 19/25]  Current/Best:   12.05/  20.07 GFLOPS | Progress: (8/20) | 10.76 s
    [Task 19/25]  Current/Best:   11.63/  20.07 GFLOPS | Progress: (12/20) | 13.93 s
    [Task 19/25]  Current/Best:    8.37/  20.72 GFLOPS | Progress: (16/20) | 17.42 s
    [Task 19/25]  Current/Best:    9.60/  20.87 GFLOPS | Progress: (20/20) | 20.56 s Done.
-
    [Task 20/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 20/25]  Current/Best:   13.40/  13.40 GFLOPS | Progress: (4/20) | 6.10 s
    [Task 20/25]  Current/Best:   19.00/  19.00 GFLOPS | Progress: (8/20) | 9.81 s
    [Task 20/25]  Current/Best:    8.27/  19.00 GFLOPS | Progress: (12/20) | 11.96 s
    [Task 20/25]  Current/Best:    3.10/  19.00 GFLOPS | Progress: (16/20) | 16.02 s
    [Task 20/25]  Current/Best:   16.60/  19.00 GFLOPS | Progress: (20/20) | 19.33 s
    [Task 21/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 21/25]  Current/Best:   21.15/  21.15 GFLOPS | Progress: (4/20) | 5.74 s Done.
-
    [Task 21/25]  Current/Best:    1.62/  21.15 GFLOPS | Progress: (8/20) | 8.42 s
    [Task 21/25]  Current/Best:   20.61/  21.15 GFLOPS | Progress: (12/20) | 11.69 s
    [Task 21/25]  Current/Best:    5.39/  21.15 GFLOPS | Progress: (16/20) | 14.80 s
    [Task 21/25]  Current/Best:   13.18/  21.15 GFLOPS | Progress: (20/20) | 22.30 s
    [Task 22/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 22/25]  Current/Best:    9.54/   9.54 GFLOPS | Progress: (4/20) | 5.98 s
    [Task 22/25]  Current/Best:   19.37/  19.37 GFLOPS | Progress: (8/20) | 9.24 s
    [Task 22/25]  Current/Best:   20.67/  20.67 GFLOPS | Progress: (12/20) | 11.61 s
    [Task 22/25]  Current/Best:   17.91/  20.67 GFLOPS | Progress: (16/20) | 13.51 s
    [Task 22/25]  Current/Best:    5.58/  20.67 GFLOPS | Progress: (20/20) | 17.52 s Done.
-
    [Task 23/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 23/25]  Current/Best:   18.89/  21.54 GFLOPS | Progress: (4/20) | 5.13 s
    [Task 23/25]  Current/Best:   17.53/  21.54 GFLOPS | Progress: (8/20) | 10.70 s
    [Task 23/25]  Current/Best:    9.17/  21.54 GFLOPS | Progress: (12/20) | 13.08 s
    [Task 23/25]  Current/Best:   17.63/  21.54 GFLOPS | Progress: (16/20) | 17.21 s
    [Task 23/25]  Current/Best:   10.35/  21.54 GFLOPS | Progress: (20/20) | 22.02 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/   5.25 GFLOPS | Progress: (4/20) | 5.86 s
    [Task 24/25]  Current/Best:    6.95/   9.41 GFLOPS | Progress: (8/20) | 7.32 s
    [Task 24/25]  Current/Best:    5.37/   9.41 GFLOPS | Progress: (12/20) | 9.26 s
    [Task 24/25]  Current/Best:    0.61/   9.41 GFLOPS | Progress: (16/20) | 19.99 s
    [Task 24/25]  Current/Best:    3.43/   9.41 GFLOPS | Progress: (20/20) | 31.02 s
    [Task 25/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 25/25]  Current/Best:    8.66/   9.12 GFLOPS | Progress: (4/20) | 5.44 s Done.
-
    [Task 25/25]  Current/Best:    3.37/   9.68 GFLOPS | Progress: (8/20) | 16.11 s
    [Task 25/25]  Current/Best:    5.75/   9.68 GFLOPS | Progress: (12/20) | 27.05 s
    [Task 25/25]  Current/Best:    3.00/   9.68 GFLOPS | Progress: (16/20) | 29.79 s
    [Task 25/25]  Current/Best:    8.13/   9.68 GFLOPS | Progress: (20/20) | 40.75 s
+
    [Task  1/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  1/25]  Current/Best:    7.83/   9.61 GFLOPS | Progress: (4/20) | 12.69 s
    [Task  1/25]  Current/Best:    7.69/  19.26 GFLOPS | Progress: (8/20) | 16.77 s
    [Task  1/25]  Current/Best:   12.11/  19.26 GFLOPS | Progress: (12/20) | 19.37 s
    [Task  1/25]  Current/Best:    6.86/  19.26 GFLOPS | Progress: (16/20) | 25.68 s
    [Task  1/25]  Current/Best:   16.79/  19.26 GFLOPS | Progress: (20/20) | 28.04 s Done.
+
    [Task  2/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  2/25]  Current/Best:    8.54/  18.51 GFLOPS | Progress: (4/20) | 5.25 s
    [Task  2/25]  Current/Best:   10.82/  18.51 GFLOPS | Progress: (8/20) | 7.36 s
    [Task  2/25]  Current/Best:   15.86/  18.51 GFLOPS | Progress: (12/20) | 9.28 s
    [Task  2/25]  Current/Best:   10.33/  18.51 GFLOPS | Progress: (16/20) | 12.23 s
    [Task  2/25]  Current/Best:    8.96/  18.51 GFLOPS | Progress: (20/20) | 13.99 s Done.
+
    [Task  3/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  3/25]  Current/Best:   15.43/  20.37 GFLOPS | Progress: (4/20) | 5.53 s
    [Task  3/25]  Current/Best:   11.50/  20.37 GFLOPS | Progress: (8/20) | 9.51 s
    [Task  3/25]  Current/Best:   13.17/  20.37 GFLOPS | Progress: (12/20) | 11.79 s
    [Task  3/25]  Current/Best:   12.67/  20.37 GFLOPS | Progress: (16/20) | 15.60 s
    [Task  3/25]  Current/Best:    6.21/  20.37 GFLOPS | Progress: (20/20) | 18.09 s Done.
+
    [Task  4/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  4/25]  Current/Best:    8.66/  16.65 GFLOPS | Progress: (4/20) | 5.00 s
    [Task  4/25]  Current/Best:   16.77/  16.77 GFLOPS | Progress: (8/20) | 6.91 s
    [Task  4/25]  Current/Best:   13.57/  20.92 GFLOPS | Progress: (12/20) | 9.02 s
    [Task  4/25]  Current/Best:   12.18/  22.43 GFLOPS | Progress: (16/20) | 11.13 s
    [Task  4/25]  Current/Best:    6.39/  22.43 GFLOPS | Progress: (20/20) | 13.96 s Done.
+
    [Task  5/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  5/25]  Current/Best:   11.02/  17.13 GFLOPS | Progress: (4/20) | 6.08 s
    [Task  5/25]  Current/Best:   10.81/  17.13 GFLOPS | Progress: (8/20) | 9.60 s
    [Task  5/25]  Current/Best:   18.25/  18.25 GFLOPS | Progress: (12/20) | 11.25 s
    [Task  5/25]  Current/Best:   14.65/  18.25 GFLOPS | Progress: (16/20) | 13.23 s
    [Task  5/25]  Current/Best:    2.82/  18.25 GFLOPS | Progress: (20/20) | 15.77 s Done.
+
    [Task  6/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  6/25]  Current/Best:    8.38/  17.39 GFLOPS | Progress: (4/20) | 6.29 s
    [Task  6/25]  Current/Best:   10.30/  17.39 GFLOPS | Progress: (8/20) | 9.39 s
    [Task  6/25]  Current/Best:   13.65/  23.06 GFLOPS | Progress: (12/20) | 11.73 s
    [Task  6/25]  Current/Best:   16.05/  23.06 GFLOPS | Progress: (16/20) | 14.43 s
    [Task  6/25]  Current/Best:   13.16/  23.06 GFLOPS | Progress: (20/20) | 21.18 s Done.
+
    [Task  7/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  7/25]  Current/Best:    9.80/  12.23 GFLOPS | Progress: (4/20) | 6.78 s
    [Task  7/25]  Current/Best:   11.57/  18.67 GFLOPS | Progress: (8/20) | 9.63 s
    [Task  7/25]  Current/Best:   12.59/  18.67 GFLOPS | Progress: (12/20) | 11.98 s
    [Task  7/25]  Current/Best:   22.36/  22.36 GFLOPS | Progress: (16/20) | 14.55 s
    [Task  7/25]  Current/Best:   11.55/  22.36 GFLOPS | Progress: (20/20) | 17.26 s Done.
+
    [Task  8/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  8/25]  Current/Best:   11.96/  11.96 GFLOPS | Progress: (4/20) | 5.73 s
    [Task  8/25]  Current/Best:   10.39/  11.96 GFLOPS | Progress: (8/20) | 11.71 s
    [Task  8/25]  Current/Best:   14.67/  19.07 GFLOPS | Progress: (12/20) | 15.12 s
    [Task  8/25]  Current/Best:   17.89/  19.07 GFLOPS | Progress: (16/20) | 26.80 s
    [Task  8/25]  Current/Best:    7.99/  19.07 GFLOPS | Progress: (20/20) | 29.23 s Done.
+
    [Task  9/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task  9/25]  Current/Best:    6.79/  16.33 GFLOPS | Progress: (4/20) | 7.77 s
    [Task  9/25]  Current/Best:    9.10/  19.50 GFLOPS | Progress: (8/20) | 11.33 s
    [Task  9/25]  Current/Best:    6.13/  19.50 GFLOPS | Progress: (12/20) | 15.93 s
    [Task  9/25]  Current/Best:   11.88/  19.50 GFLOPS | Progress: (16/20) | 19.77 s
    [Task  9/25]  Current/Best:   16.24/  19.50 GFLOPS | Progress: (20/20) | 21.63 s Done.
+
    [Task 10/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 10/25]  Current/Best:   10.06/  16.69 GFLOPS | Progress: (4/20) | 5.50 s
    [Task 10/25]  Current/Best:   15.95/  17.46 GFLOPS | Progress: (8/20) | 7.91 s
    [Task 10/25]  Current/Best:   15.30/  17.79 GFLOPS | Progress: (12/20) | 9.70 s
    [Task 10/25]  Current/Best:    3.88/  22.17 GFLOPS | Progress: (16/20) | 11.69 s
    [Task 10/25]  Current/Best:   15.16/  22.17 GFLOPS | Progress: (20/20) | 13.80 s Done.
+
    [Task 11/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 11/25]  Current/Best:   10.08/  18.49 GFLOPS | Progress: (4/20) | 5.46 s
    [Task 11/25]  Current/Best:    9.68/  18.49 GFLOPS | Progress: (8/20) | 8.61 s
    [Task 11/25]  Current/Best:   12.39/  18.49 GFLOPS | Progress: (12/20) | 11.24 s
    [Task 11/25]  Current/Best:   16.59/  18.49 GFLOPS | Progress: (16/20) | 13.39 s
    [Task 11/25]  Current/Best:   12.33/  23.62 GFLOPS | Progress: (20/20) | 15.73 s Done.
+
    [Task 12/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 12/25]  Current/Best:   12.04/  18.26 GFLOPS | Progress: (4/20) | 6.61 s
    [Task 12/25]  Current/Best:   13.01/  18.26 GFLOPS | Progress: (8/20) | 9.11 s
    [Task 12/25]  Current/Best:   14.32/  18.26 GFLOPS | Progress: (12/20) | 11.59 s
    [Task 12/25]  Current/Best:    8.82/  18.64 GFLOPS | Progress: (16/20) | 14.74 s
    [Task 12/25]  Current/Best:   12.87/  18.64 GFLOPS | Progress: (20/20) | 17.31 s Done.
+
    [Task 13/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 13/25]  Current/Best:   19.81/  19.81 GFLOPS | Progress: (4/20) | 4.86 s
    [Task 13/25]  Current/Best:   15.86/  19.81 GFLOPS | Progress: (8/20) | 8.98 s
    [Task 13/25]  Current/Best:   18.43/  21.04 GFLOPS | Progress: (12/20) | 12.34 s
    [Task 13/25]  Current/Best:    6.88/  21.04 GFLOPS | Progress: (16/20) | 15.84 s
    [Task 13/25]  Current/Best:    6.74/  21.04 GFLOPS | Progress: (20/20) | 20.38 s Done.
+
    [Task 14/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 14/25]  Current/Best:   13.99/  13.99 GFLOPS | Progress: (4/20) | 5.26 s
    [Task 14/25]  Current/Best:    8.24/  14.00 GFLOPS | Progress: (8/20) | 13.64 s
    [Task 14/25]  Current/Best:   10.22/  14.00 GFLOPS | Progress: (12/20) | 17.61 s
    [Task 14/25]  Current/Best:   15.84/  15.84 GFLOPS | Progress: (16/20) | 20.07 s
    [Task 14/25]  Current/Best:    9.38/  15.84 GFLOPS | Progress: (20/20) | 22.44 s
    [Task 15/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 15/25]  Current/Best:   15.49/  15.49 GFLOPS | Progress: (4/20) | 5.53 s
    [Task 15/25]  Current/Best:   11.17/  16.09 GFLOPS | Progress: (8/20) | 8.41 s
    [Task 15/25]  Current/Best:   18.25/  18.25 GFLOPS | Progress: (12/20) | 10.60 s
    [Task 15/25]  Current/Best:    6.62/  19.79 GFLOPS | Progress: (16/20) | 14.09 s
    [Task 15/25]  Current/Best:   11.60/  19.79 GFLOPS | Progress: (20/2
 0) | 15.92 s
    [Task 16/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 16/25]  Current/Best:   20.48/  20.48 GFLOPS | Progress: (4/20) | 5.66 s
    [Task 16/25]  Current/Best:    7.56/  20.48 GFLOPS | Progress: (8/20) | 8.84 s
    [Task 16/25]  Current/Best:    6.36/  22.91 GFLOPS | Progress: (12/20) | 10.41 s
    [Task 16/25]  Current/Best:   21.12/  22.91 GFLOPS | Progress: (16/20) | 12.22 s
    [Task 16/25]  Current/Best:    4.87/  22.91 GFLOPS | Progress: (20/20) | 14.66 s Done.
+
    [Task 17/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 17/25]  Current/Best:   20.13/  20.13 GFLOPS | Progress: (4/20) | 5.55 s
    [Task 17/25]  Current/Best:    6.43/  20.13 GFLOPS | Progress: (8/20) | 8.52 s
    [Task 17/25]  Current/Best:    9.05/  22.48 GFLOPS | Progress: (12/20) | 12.41 s
    [Task 17/25]  Current/Best:   11.95/  22.48 GFLOPS | Progress: (16/20) | 15.55 s
    [Task 17/25]  Current/Best:   19.24/  22.48 GFLOPS | Progress: (20/20) | 18.06 s Done.
+
    [Task 18/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 18/25]  Current/Best:    9.13/  19.05 GFLOPS | Progress: (4/20) | 8.97 s
    [Task 18/25]  Current/Best:   14.22/  19.05 GFLOPS | Progress: (8/20) | 11.46 s
    [Task 18/25]  Current/Best:   14.07/  19.07 GFLOPS | Progress: (12/20) | 14.46 s
    [Task 18/25]  Current/Best:    6.57/  19.07 GFLOPS | Progress: (16/20) | 18.44 s
    [Task 18/25]  Current/Best:    9.28/  19.07 GFLOPS | Progress: (20/20) | 24.99 s Done.
+
    [Task 19/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 19/25]  Current/Best:   10.81/  19.76 GFLOPS | Progress: (4/20) | 6.28 s
    [Task 19/25]  Current/Best:   14.55/  19.76 GFLOPS | Progress: (8/20) | 9.64 s
    [Task 19/25]  Current/Best:    9.37/  19.76 GFLOPS | Progress: (12/20) | 13.43 s
    [Task 19/25]  Current/Best:   16.81/  19.76 GFLOPS | Progress: (16/20) | 15.72 s
    [Task 19/25]  Current/Best:   22.16/  22.16 GFLOPS | Progress: (20/20) | 18.40 s Done.
+
    [Task 20/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 20/25]  Current/Best:   11.11/  11.11 GFLOPS | Progress: (4/20) | 5.03 s
    [Task 20/25]  Current/Best:   15.74/  17.15 GFLOPS | Progress: (8/20) | 8.48 s Done.
+     Done.
+
    [Task 20/25]  Current/Best:    9.24/  17.15 GFLOPS | Progress: (12/20) | 10.90 s
    [Task 20/25]  Current/Best:    8.84/  18.64 GFLOPS | Progress: (16/20) | 14.36 s
    [Task 20/25]  Current/Best:   13.09/  18.64 GFLOPS | Progress: (20/20) | 17.13 s
    [Task 21/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 21/25]  Current/Best:    7.21/  15.67 GFLOPS | Progress: (4/20) | 5.48 s
    [Task 21/25]  Current/Best:   11.23/  16.99 GFLOPS | Progress: (8/20) | 8.35 s
    [Task 21/25]  Current/Best:    5.30/  16.99 GFLOPS | Progress: (12/20) | 11.96 s
    [Task 21/25]  Current/Best:    2.72/  19.44 GFLOPS | Progress: (16/20) | 15.73 s
    [Task 21/25]  Current/Best:   16.62/  19.44 GFLOPS | Progress: (20/20) | 19.36 s
    [Task 22/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 22/25]  Current/Best:   21.69/  21.69 GFLOPS | Progress: (4/20) | 6.73 s
    [Task 22/25]  Current/Best:   17.94/  21.69 GFLOPS | Progress: (8/2
 0) | 8.55 s
    [Task 22/25]  Current/Best:   10.52/  21.69 GFLOPS | Progress: (12/20) | 10.38 s
    [Task 22/25]  Current/Best:   11.95/  21.69 GFLOPS | Progress: (16/20) | 12.36 s
    [Task 22/25]  Current/Best:    4.73/  21.69 GFLOPS | Progress: (20/20) | 14.39 s Done.
+
    [Task 23/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 23/25]  Current/Best:   18.94/  18.94 GFLOPS | Progress: (4/20) | 5.38 s
    [Task 23/25]  Current/Best:   22.74/  22.74 GFLOPS | Progress: (8/20) | 9.53 s
    [Task 23/25]  Current/Best:    1.55/  22.74 GFLOPS | Progress: (12/20) | 14.81 s
    [Task 23/25]  Current/Best:   11.97/  22.74 GFLOPS | Progress: (16/20) | 20.15 s
    [Task 23/25]  Current/Best:   19.56/  22.74 GFLOPS | Progress: (20/20) | 24.45 s Done.
+
    [Task 24/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 24/25]  Current/Best:    5.60/   5.60 GFLOPS | Progress: (4/20) | 13.82 s
    [Task 24/25]  Current/Best:    1.89/   6.78 GFLOPS | Progress: (8/20) | 24.78 s Done.
+     Done.
+
    [Task 24/25]  Current/Best:    9.90/   9.90 GFLOPS | Progress: (12/20) | 34.78 s
    [Task 24/25]  Current/Best:    5.58/   9.90 GFLOPS | Progress: (16/20) | 39.00 s
    [Task 24/25]  Current/Best:    1.60/   9.90 GFLOPS | Progress: (20/20) | 46.04 s
    [Task 25/25]  Current/Best:    0.00/   0.00 GFLOPS | Progress: (0/20) | 0.00 s
    [Task 25/25]  Current/Best:    5.51/   7.64 GFLOPS | Progress: (4/20) | 5.97 s
    [Task 25/25]  Current/Best:    1.54/   7.64 GFLOPS | Progress: (8/20) | 16.97 s
    [Task 25/25]  Current/Best:    5.70/   8.81 GFLOPS | Progress: (12/20) | 19.62 s
    [Task 25/25]  Current/Best:    1.53/   8.81 GFLOPS | Progress: (16/20) | 24.24 s
    [Task 25/25]  Current/Best:    2.92/   8.81 GFLOPS | Progress: (20/20) | 27.07 s
 
 
 
@@ -722,8 +722,8 @@ improvement in comparing the optimized model to the unoptimized model.
 
  .. code-block:: none
 
-    optimized: {'mean': 414.89879594000286, 'median': 413.83022729999084, 'std': 3.08408415957333}
-    unoptimized: {'mean': 517.4032972500004, 'median': 518.1137195500014, 'std': 1.692627985874011}
+    optimized: {'mean': 408.57807502999776, 'median': 407.6914375499996, 'std': 2.4356376885080357}
+    unoptimized: {'mean': 514.8689825700013, 'median': 514.807255200003, 'std': 1.1113574405842}
 
 
 
@@ -746,7 +746,7 @@ profiling/benchmarking.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 12 minutes  44.977 seconds)
+   **Total running time of the script:** ( 12 minutes  22.302 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 4c20798589..d8ee460faf 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
 
-    1.552e-07 secs/op
+    1.341e-07 secs/op
 
 
 
diff --git a/docs/_sources/tutorial/intro_topi.rst.txt b/docs/_sources/tutorial/intro_topi.rst.txt
index 38a8eb315d..089ce77c99 100644
--- a/docs/_sources/tutorial/intro_topi.rst.txt
+++ b/docs/_sources/tutorial/intro_topi.rst.txt
@@ -270,7 +270,7 @@ As you can see, scheduled stages of computation have been accumulated and we can
 
  .. code-block:: none
 
-    [stage(a, placeholder(a, 0xfe66c70)), stage(b, placeholder(b, 0x23c15e40)), stage(T_add, compute(T_add, body=[a[ax0, ax1, ax2] + b[ax1, ax2]], axis=[T.iter_var(ax0, T.Range(0, 100), "DataPar", ""), T.iter_var(ax1, T.Range(0, 10), "DataPar", ""), T.iter_var(ax2, T.Range(0, 10), "DataPar", "")], reduce_axis=[], tag=broadcast, attrs={})), stage(T_multiply, compute(T_multiply, body=[a[ax0, ax1, ax2] * b[ax1, ax2]], axis=[T.iter_var(ax0, T.Range(0, 100), "DataPar", ""), T.iter_var(ax1, T. [...]
+    [stage(a, placeholder(a, 0x2e385120)), stage(b, placeholder(b, 0x70ed8f0)), 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 1b7f9b9be1..b0c1ee3a64 100644
--- a/docs/_sources/tutorial/sg_execution_times.rst.txt
+++ b/docs/_sources/tutorial/sg_execution_times.rst.txt
@@ -5,32 +5,32 @@
 
 Computation times
 =================
-**16:39.130** total execution time for **tutorial** files:
+**15:55.606** total execution time for **tutorial** files:
 
 +------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``)                 | 12:44.977 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_relay_x86.py` (``autotvm_relay_x86.py``)                 | 12:22.302 | 0.0 MB |
 +------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 01:47.961 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_auto_scheduler_matmul_x86.py` (``auto_scheduler_matmul_x86.py``) | 01:33.638 | 0.0 MB |
 +------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``)     | 01:00.225 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_expr_get_started.py` (``tensor_expr_get_started.py``)     | 01:00.851 | 0.0 MB |
 +------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``)                 | 00:36.923 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_relay_quick_start.py` (``relay_quick_start.py``)                 | 00:36.970 | 0.0 MB |
 +------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``)               | 00:26.719 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_autotvm_matmul_x86.py` (``autotvm_matmul_x86.py``)               | 00:19.385 | 0.0 MB |
 +------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``)       | 00:01.275 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_tensor_ir_blitz_course.py` (``tensor_ir_blitz_course.py``)       | 00:01.407 | 0.0 MB |
 +------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``)                               | 00:00.866 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_intro_topi.py` (``intro_topi.py``)                               | 00:00.858 | 0.0 MB |
 +------------------------------------------------------------------------------------------+-----------+--------+
-| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.182 | 0.0 MB |
+| :ref:`sphx_glr_tutorial_cross_compilation_and_rpc.py` (``cross_compilation_and_rpc.py``) | 00:00.194 | 0.0 MB |
 +------------------------------------------------------------------------------------------+-----------+--------+
 | :ref:`sphx_glr_tutorial_uma.py` (``uma.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_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_introduction.py` (``introduction.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 df1b01fb59..89bd2ae119 100644
--- a/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
+++ b/docs/_sources/tutorial/tensor_expr_get_started.rst.txt
@@ -285,7 +285,7 @@ helper function to run a profile of the TVM generated code.
 
  .. code-block:: none
 
-    Numpy running time: 0.000007
+    Numpy running time: 0.000008
     naive: 0.000007
 
 
@@ -389,7 +389,7 @@ compile and run this new schedule with the parallel operation applied:
 
  .. code-block:: none
 
-    parallel: 0.000012
+    parallel: 0.000007
 
 
 
@@ -498,10 +498,10 @@ We can now compare the different schedules
  .. code-block:: none
 
                 Operator                  Timing             Performance
-                   numpy    7.281870000497292e-06                    1.0
-                   naive              6.6897e-06      0.9186788557806096
-                parallel             1.15118e-05      1.5808851296732622
-                  vector    2.4609400000000002e-05      3.37954399052982
+                   numpy    7.727219999651425e-06                    1.0
+                   naive    6.691099999999999e-06     0.8659129674451919
+                parallel    6.984899999999999e-06     0.9039344033578813
+                  vector             2.46369e-05       3.188326461665563
 
 
 
@@ -922,7 +922,7 @@ matrix multiplication.
 
  .. code-block:: none
 
-    Numpy running time: 0.018844
+    Numpy running time: 0.019485
 
 
 
@@ -980,7 +980,7 @@ optimizations.
 
  .. code-block:: none
 
-    none: 3.315918
+    none: 3.366190
 
 
 
@@ -1080,7 +1080,7 @@ schedule.
 
  .. code-block:: none
 
-    blocking: 0.317580
+    blocking: 0.313734
 
 
 
@@ -1164,7 +1164,7 @@ already cache friendly from our previous optimizations.
 
  .. code-block:: none
 
-    vectorization: 0.348403
+    vectorization: 0.344156
     # from tvm.script import ir as I
     # from tvm.script import tir as T
 
@@ -1230,7 +1230,7 @@ more cache friendly.
 
  .. code-block:: none
 
-    loop permutation: 0.115508
+    loop permutation: 0.119324
     # from tvm.script import ir as I
     # from tvm.script import tir as T
 
@@ -1321,7 +1321,7 @@ optimized schedule.
 
  .. code-block:: none
 
-    array packing: 0.109137
+    array packing: 0.108431
     # from tvm.script import ir as I
     # from tvm.script import tir as T
 
@@ -1404,7 +1404,7 @@ to `C` when all the block results are ready.
 
  .. code-block:: none
 
-    block caching: 0.110537
+    block caching: 0.110456
     # from tvm.script import ir as I
     # from tvm.script import tir as T
 
@@ -1478,7 +1478,7 @@ of thread-level parallelization.
 
  .. code-block:: none
 
-    parallelization: 0.146267
+    parallelization: 0.146389
     # from tvm.script import ir as I
     # from tvm.script import tir as T
 
@@ -1548,13 +1548,13 @@ working, we can compare the results.
  .. code-block:: none
 
                 Operator                  Timing             Performance
-                    none      3.3159181699999998                     1.0
-                blocking            0.3175798164      0.0957743225611626
-           vectorization     0.34840252569999997     0.10506969950347116
-        loop permutation     0.11550810550000001    0.034834425814555016
-           array packing     0.10913673809999999     0.03291297689050029
-           block caching            0.1105371131    0.033335295816422396
-         parallelization            0.1462669291    0.044110536388779466
+                    none            3.3661901395                     1.0
+                blocking            0.3137343787     0.09320162132808124
+           vectorization            0.3441560657     0.10223904516312306
+        loop permutation            0.1193241219     0.03544782586693808
+           array packing     0.10843102769999999    0.032211795295706586
+           block caching     0.11045565930000001     0.03281325615088595
+         parallelization            0.1463894652     0.04348817480100637
 
 
 
@@ -1596,7 +1596,7 @@ the computation for specific platforms.
 
 .. rst-class:: sphx-glr-timing
 
-   **Total running time of the script:** ( 1 minutes  0.225 seconds)
+   **Total running time of the script:** ( 1 minutes  0.851 seconds)
 
 
 .. _sphx_glr_download_tutorial_tensor_expr_get_started.py:
diff --git a/docs/commit_hash b/docs/commit_hash
index edda12e136..a758b8bfd4 100644
--- a/docs/commit_hash
+++ b/docs/commit_hash
@@ -1 +1 @@
-7f6da090528b683bb285fcc657a46e45501ca48c
+4819300803fd2477180d3b38972e24eb080f1c08
diff --git a/docs/how_to/compile_models/from_darknet.html b/docs/how_to/compile_models/from_darknet.html
index db65238fb4..6be328f0d0 100644
--- a/docs/how_to/compile_models/from_darknet.html
+++ b/docs/how_to/compile_models/from_darknet.html
@@ -590,7 +590,7 @@ class:[&#39;truck 0.9266&#39;] left:471 top:83 right:689 bottom:169
 class:[&#39;bicycle 0.9984&#39;] 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  17.627 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  20.524 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 945f1259bc..1a921f1017 100644
--- a/docs/how_to/compile_models/from_keras.html
+++ b/docs/how_to/compile_models/from_keras.html
@@ -511,7 +511,7 @@ Tensorflow is also required since it’s used as the default backend of keras.</
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Relay top-1 id: 285, class name: Egyptian cat
 
 1/1 [==============================] - ETA: 0s
-1/1 [==============================] - 1s 946ms/step
+1/1 [==============================] - 1s 953ms/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 7ed7f5c5a2..556c86cd82 100644
--- a/docs/how_to/compile_models/from_mxnet.html
+++ b/docs/how_to/compile_models/from_mxnet.html
@@ -444,7 +444,7 @@
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;x&quot;</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.zipff49ba47-aac5-4505-a39d-7b52e8e5d528 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.zip7edf74da-5053-4323-a546-fa38db9222a0 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 9c76b7f1d7..189a4b8387 100644
--- a/docs/how_to/compile_models/from_oneflow.html
+++ b/docs/how_to/compile_models/from_oneflow.html
@@ -454,11 +454,12 @@ Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdo
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading: &quot;https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/ResNet/resnet18.zip&quot; to /workspace/.oneflow/flowvision_cache/resnet18.zip
 
   0%|          | 0.00/41.5M [00:00&lt;?, ?B/s]
- 19%|#9        | 7.99M/41.5M [00:00&lt;00:00, 80.8MB/s]
- 39%|###8      | 16.1M/41.5M [00:00&lt;00:00, 71.3MB/s]
- 55%|#####5    | 23.0M/41.5M [00:00&lt;00:00, 66.1MB/s]
- 77%|#######7  | 32.1M/41.5M [00:00&lt;00:00, 73.8MB/s]
-100%|##########| 41.5M/41.5M [00:00&lt;00:00, 82.3MB/s]
+ 22%|##1       | 9.04M/41.5M [00:00&lt;00:00, 94.8MB/s]
+ 44%|####3     | 18.1M/41.5M [00:00&lt;00:00, 78.3MB/s]
+ 62%|######1   | 25.7M/41.5M [00:00&lt;00:00, 77.1MB/s]
+ 80%|#######9  | 33.2M/41.5M [00:00&lt;00:00, 73.6MB/s]
+ 97%|#########6| 40.2M/41.5M [00:00&lt;00:00, 62.9MB/s]
+100%|##########| 41.5M/41.5M [00:00&lt;00:00, 70.8MB/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 b41a89ee13..0f04d068fd 100644
--- a/docs/how_to/compile_models/from_pytorch.html
+++ b/docs/how_to/compile_models/from_pytorch.html
@@ -437,10 +437,10 @@ be unstable.</p>
 Downloading: &quot;https://download.pytorch.org/models/resnet18-f37072fd.pth&quot; to /workspace/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth
 
   0%|          | 0.00/44.7M [00:00&lt;?, ?B/s]
- 28%|##7       | 12.3M/44.7M [00:00&lt;00:00, 128MB/s]
- 55%|#####4    | 24.5M/44.7M [00:00&lt;00:00, 110MB/s]
- 79%|#######8  | 35.1M/44.7M [00:00&lt;00:00, 106MB/s]
-100%|##########| 44.7M/44.7M [00:00&lt;00:00, 106MB/s]
+ 26%|##6       | 11.6M/44.7M [00:00&lt;00:00, 122MB/s]
+ 54%|#####3    | 24.0M/44.7M [00:00&lt;00:00, 107MB/s]
+ 77%|#######7  | 34.5M/44.7M [00:00&lt;00:00, 109MB/s]
+100%|##########| 44.7M/44.7M [00:00&lt;00:00, 71.2MB/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 2948c3df3b..cc2bca83c3 100644
--- a/docs/how_to/compile_models/from_tensorflow.html
+++ b/docs/how_to/compile_models/from_tensorflow.html
@@ -654,7 +654,7 @@ banana (score = 0.00022)
 desk (score = 0.00019)
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  23.887 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  23.477 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 7cd013c559..6bcab147fa 100644
--- a/docs/how_to/compile_models/sg_execution_times.html
+++ b/docs/how_to/compile_models/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-compile-models-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>06:37.466</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
+<p><strong>06:39.807</strong> total execution time for <strong>how_to_compile_models</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 81%" />
@@ -354,43 +354,43 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="from_tensorflow.html#sphx-glr-how-to-compile-models-from-tensorflow-py"><span class="std std-ref">Compile Tensorflow Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_tensorflow.py</span></code>)</p></td>
-<td><p>01:23.887</p></td>
+<td><p>01:23.477</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>01:17.627</p></td>
+<td><p>01:20.524</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="from_paddle.html#sphx-glr-how-to-compile-models-from-paddle-py"><span class="std std-ref">Compile PaddlePaddle Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_paddle.py</span></code>)</p></td>
-<td><p>00:56.604</p></td>
+<td><p>00:56.418</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="from_oneflow.html#sphx-glr-how-to-compile-models-from-oneflow-py"><span class="std std-ref">Compile OneFlow Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_oneflow.py</span></code>)</p></td>
-<td><p>00:37.764</p></td>
+<td><p>00:37.570</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="from_coreml.html#sphx-glr-how-to-compile-models-from-coreml-py"><span class="std std-ref">Compile CoreML Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_coreml.py</span></code>)</p></td>
-<td><p>00:32.531</p></td>
+<td><p>00:32.478</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="from_mxnet.html#sphx-glr-how-to-compile-models-from-mxnet-py"><span class="std std-ref">Compile MXNet Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_mxnet.py</span></code>)</p></td>
-<td><p>00:31.432</p></td>
+<td><p>00:31.420</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>00:26.843</p></td>
+<td><p>00:26.647</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="from_pytorch.html#sphx-glr-how-to-compile-models-from-pytorch-py"><span class="std std-ref">Compile PyTorch Models</span></a> (<code class="docutils literal notranslate"><span class="pre">from_pytorch.py</span></code>)</p></td>
-<td><p>00:26.083</p></td>
+<td><p>00:26.337</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>00:21.991</p></td>
+<td><p>00:22.175</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>00:02.704</p></td>
+<td><p>00:02.761</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 d63162b345..ec4947582d 100644
--- a/docs/how_to/deploy_models/deploy_model_on_adreno.html
+++ b/docs/how_to/deploy_models/deploy_model_on_adreno.html
@@ -799,7 +799,7 @@ Top5 predictions:
 Evaluate inference time cost...
 Execution time summary:
  mean (ms)   median (ms)    max (ms)     min (ms)     std (ms)
- 2535.3280    2534.9226    2538.4680    2533.7433      1.4229
+ 2538.1084    2537.6719    2541.9637    2536.2990      1.6237
 </pre></div>
 </div>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-deploy-models-deploy-model-on-adreno-py">
diff --git a/docs/how_to/deploy_models/deploy_model_on_adreno_tvmc.html b/docs/how_to/deploy_models/deploy_model_on_adreno_tvmc.html
index 838850e8c2..519713b417 100644
--- a/docs/how_to/deploy_models/deploy_model_on_adreno_tvmc.html
+++ b/docs/how_to/deploy_models/deploy_model_on_adreno_tvmc.html
@@ -443,25 +443,21 @@ to run this tutorial with a real device over rpc.</p>
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels.h5
 
      8192/102967424 [..............................] - ETA: 0s
-  6995968/102967424 [=&gt;............................] - ETA: 0s
- 12099584/102967424 [==&gt;...........................] - ETA: 0s
+  8380416/102967424 [=&gt;............................] - ETA: 1s
  16769024/102967424 [===&gt;..........................] - ETA: 0s
- 22142976/102967424 [=====&gt;........................] - ETA: 0s
- 27222016/102967424 [======&gt;.......................] - ETA: 0s
- 32186368/102967424 [========&gt;.....................] - ETA: 0s
- 37208064/102967424 [=========&gt;....................] - ETA: 0s
- 41934848/102967424 [===========&gt;..................] - ETA: 0s
- 48603136/102967424 [=============&gt;................] - ETA: 0s
- 53714944/102967424 [==============&gt;...............] - ETA: 0s
+ 26320896/102967424 [======&gt;.......................] - ETA: 0s
+ 33546240/102967424 [========&gt;.....................] - ETA: 0s
+ 39624704/102967424 [==========&gt;...................] - ETA: 0s
+ 42065920/102967424 [===========&gt;..................] - ETA: 0s
+ 53239808/102967424 [==============&gt;...............] - ETA: 0s
  58712064/102967424 [================&gt;.............] - ETA: 0s
- 64036864/102967424 [=================&gt;............] - ETA: 0s
- 69025792/102967424 [===================&gt;..........] - ETA: 0s
- 74006528/102967424 [====================&gt;.........] - ETA: 0s
- 79011840/102967424 [======================&gt;.......] - ETA: 0s
- 84803584/102967424 [=======================&gt;......] - ETA: 0s
- 89776128/102967424 [=========================&gt;....] - ETA: 0s
- 94330880/102967424 [==========================&gt;...] - ETA: 0s
-100114432/102967424 [============================&gt;.] - ETA: 0s
+ 65732608/102967424 [==================&gt;...........] - ETA: 0s
+ 75169792/102967424 [====================&gt;.........] - ETA: 0s
+ 75628544/102967424 [=====================&gt;........] - ETA: 0s
+ 77684736/102967424 [=====================&gt;........] - ETA: 0s
+ 83877888/102967424 [=======================&gt;......] - ETA: 0s
+ 92405760/102967424 [=========================&gt;....] - ETA: 0s
+100646912/102967424 [============================&gt;.] - ETA: 0s
 102967424/102967424 [==============================] - 1s 0us/step
 </pre></div>
 </div>
diff --git a/docs/how_to/deploy_models/deploy_model_on_android.html b/docs/how_to/deploy_models/deploy_model_on_android.html
index c0d87fc2ac..17d3905f61 100644
--- a/docs/how_to/deploy_models/deploy_model_on_android.html
+++ b/docs/how_to/deploy_models/deploy_model_on_android.html
@@ -667,7 +667,7 @@ to the remote android device.</p>
 Evaluate inference time cost...
 Execution time summary:
  mean (ms)   median (ms)    max (ms)     min (ms)     std (ms)
-  16.0745      16.0362      16.3460      15.9037       0.1235
+  16.1313      16.0852      16.4169      15.9376       0.1567
 </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 6e4930445d..55babbea15 100644
--- a/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
+++ b/docs/how_to/deploy_models/deploy_object_detection_pytorch.html
@@ -459,22 +459,30 @@ be unstable.</p>
 Downloading: &quot;https://download.pytorch.org/models/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth&quot; to /workspace/.cache/torch/hub/checkpoints/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth
 
   0%|          | 0.00/170M [00:00&lt;?, ?B/s]
-  5%|4         | 7.99M/170M [00:00&lt;00:02, 69.9MB/s]
- 10%|9         | 16.6M/170M [00:00&lt;00:01, 80.9MB/s]
- 15%|#5        | 25.6M/170M [00:00&lt;00:01, 86.6MB/s]
- 20%|#9        | 33.9M/170M [00:00&lt;00:01, 80.9MB/s]
- 26%|##6       | 44.5M/170M [00:00&lt;00:01, 91.3MB/s]
- 31%|###1      | 53.4M/170M [00:00&lt;00:01, 79.6MB/s]
- 38%|###7      | 64.0M/170M [00:00&lt;00:01, 88.2MB/s]
- 43%|####2     | 72.7M/170M [00:00&lt;00:01, 80.4MB/s]
- 52%|#####1    | 88.0M/170M [00:01&lt;00:00, 94.9MB/s]
- 57%|#####7    | 97.2M/170M [00:01&lt;00:00, 91.1MB/s]
- 66%|######5   | 112M/170M [00:01&lt;00:00, 106MB/s]
- 72%|#######1  | 122M/170M [00:01&lt;00:00, 96.3MB/s]
- 80%|########  | 136M/170M [00:01&lt;00:00, 98.9MB/s]
- 89%|########9 | 152M/170M [00:01&lt;00:00, 103MB/s]
- 95%|#########5| 162M/170M [00:02&lt;00:00, 70.0MB/s]
-100%|##########| 170M/170M [00:02&lt;00:00, 86.6MB/s]
+  4%|3         | 6.30M/170M [00:00&lt;00:06, 28.0MB/s]
+  5%|5         | 8.98M/170M [00:00&lt;00:06, 26.6MB/s]
+  9%|9         | 16.0M/170M [00:00&lt;00:04, 32.5MB/s]
+ 13%|#3        | 22.3M/170M [00:00&lt;00:03, 41.5MB/s]
+ 18%|#7        | 30.3M/170M [00:00&lt;00:02, 50.3MB/s]
+ 23%|##2       | 38.5M/170M [00:00&lt;00:02, 59.9MB/s]
+ 26%|##6       | 44.6M/170M [00:01&lt;00:02, 54.8MB/s]
+ 31%|###       | 52.2M/170M [00:01&lt;00:02, 61.2MB/s]
+ 37%|###6      | 62.3M/170M [00:01&lt;00:01, 69.3MB/s]
+ 41%|####      | 69.1M/170M [00:01&lt;00:01, 63.3MB/s]
+ 44%|####4     | 75.4M/170M [00:01&lt;00:01, 63.3MB/s]
+ 48%|####8     | 81.6M/170M [00:01&lt;00:01, 61.1MB/s]
+ 56%|#####5    | 94.6M/170M [00:01&lt;00:00, 81.1MB/s]
+ 60%|######    | 103M/170M [00:01&lt;00:00, 74.6MB/s]
+ 65%|######4   | 110M/170M [00:01&lt;00:00, 72.0MB/s]
+ 69%|######8   | 117M/170M [00:02&lt;00:00, 66.5MB/s]
+ 73%|#######2  | 124M/170M [00:02&lt;00:00, 64.2MB/s]
+ 76%|#######6  | 130M/170M [00:02&lt;00:00, 56.6MB/s]
+ 80%|########  | 136M/170M [00:02&lt;00:00, 54.2MB/s]
+ 85%|########4 | 144M/170M [00:02&lt;00:00, 60.0MB/s]
+ 89%|########9 | 152M/170M [00:02&lt;00:00, 60.7MB/s]
+ 94%|#########4| 160M/170M [00:02&lt;00:00, 59.0MB/s]
+ 99%|#########8| 168M/170M [00:02&lt;00:00, 64.3MB/s]
+100%|##########| 170M/170M [00:03&lt;00:00, 58.9MB/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 &#39;trunc&#39; function NOT &#39;floor&#39;). This results in incorrect rounding for negative values. To keep the current behavior, use torch.div(a, b, rounding_mode=&#39;trunc&#39;), or for actual floor division, use torch.div(a, b, rounding_mode=& [...]
@@ -572,7 +580,7 @@ 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  39.868 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes  38.121 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 9cbade063e..6db6846662 100644
--- a/docs/how_to/deploy_models/deploy_prequantized.html
+++ b/docs/how_to/deploy_models/deploy_prequantized.html
@@ -500,8 +500,9 @@ training. Other models require a full post training calibration.</p>
 Downloading: &quot;https://download.pytorch.org/models/mobilenet_v2-b0353104.pth&quot; to /workspace/.cache/torch/hub/checkpoints/mobilenet_v2-b0353104.pth
 
   0%|          | 0.00/13.6M [00:00&lt;?, ?B/s]
- 59%|#####8    | 7.99M/13.6M [00:00&lt;00:00, 63.1MB/s]
-100%|##########| 13.6M/13.6M [00:00&lt;00:00, 78.1MB/s]
+ 59%|#####8    | 7.99M/13.6M [00:00&lt;00:00, 53.1MB/s]
+ 96%|#########6| 13.1M/13.6M [00:00&lt;00:00, 51.2MB/s]
+100%|##########| 13.6M/13.6M [00:00&lt;00:00, 52.9MB/s]
 </pre></div>
 </div>
 </div>
@@ -592,7 +593,7 @@ 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)
-  90.4523      90.4264      91.4663      90.0434       0.2777
+  90.4230      90.2571      96.1396      90.0245       0.7508
 </pre></div>
 </div>
 <div class="admonition note">
@@ -631,7 +632,7 @@ 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  18.206 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  17.766 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 1014fe3689..a6f15bea3e 100644
--- a/docs/how_to/deploy_models/deploy_prequantized_tflite.html
+++ b/docs/how_to/deploy_models/deploy_prequantized_tflite.html
@@ -585,7 +585,7 @@ TFLite Top-5 labels: [387 102 386 341 349]
 </div>
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time summary:
  mean (ms)   median (ms)    max (ms)     min (ms)     std (ms)
-  119.5424     119.5118     120.5062     118.7352      0.3464
+  118.1251     118.0449     124.9242     117.3214      0.7493
 </pre></div>
 </div>
 <div class="admonition note">
@@ -613,7 +613,7 @@ network for ARM CPU</span></a>.</p></li>
 </ul>
 </div></blockquote>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 2 minutes  30.194 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 2 minutes  26.632 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 3acec84925..1a4ffbe1ae 100644
--- a/docs/how_to/deploy_models/deploy_quantized.html
+++ b/docs/how_to/deploy_models/deploy_quantized.html
@@ -526,7 +526,7 @@ for calibration. But the accuracy might be impacted.</p>
   DeprecationWarning,
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  37.937 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  39.156 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 04d48e1bf3..e3feab7eaf 100644
--- a/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
+++ b/docs/how_to/deploy_models/deploy_ssd_gluoncv.html
@@ -468,24 +468,24 @@ to your device.</p>
 Downloading /workspace/.mxnet/models/ssd_512_resnet50_v1_voc-9c8b225a.zip from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/ssd_512_resnet50_v1_voc-9c8b225a.zip...
 
   0%|          | 0/132723 [00:00&lt;?, ?KB/s]
-  1%|          | 1081/132723 [00:00&lt;00:12, 10793.34KB/s]
-  3%|2         | 3398/132723 [00:00&lt;00:07, 18067.79KB/s]
-  6%|6         | 8046/132723 [00:00&lt;00:04, 30979.96KB/s]
- 12%|#2        | 16404/132723 [00:00&lt;00:02, 51716.28KB/s]
- 19%|#8        | 25204/132723 [00:00&lt;00:01, 64787.28KB/s]
- 26%|##5       | 34017/132723 [00:00&lt;00:01, 72715.63KB/s]
- 32%|###2      | 42881/132723 [00:00&lt;00:01, 77917.45KB/s]
- 39%|###8      | 51683/132723 [00:00&lt;00:00, 81131.35KB/s]
- 46%|####5     | 60503/132723 [00:00&lt;00:00, 83339.27KB/s]
- 52%|#####2    | 69308/132723 [00:01&lt;00:00, 84792.17KB/s]
- 59%|#####8    | 78180/132723 [00:01&lt;00:00, 85991.28KB/s]
- 66%|######5   | 87068/132723 [00:01&lt;00:00, 86867.72KB/s]
- 72%|#######2  | 96009/132723 [00:01&lt;00:00, 87634.97KB/s]
- 79%|#######9  | 104959/132723 [00:01&lt;00:00, 88195.92KB/s]
- 86%|########5 | 113779/132723 [00:01&lt;00:00, 87914.72KB/s]
- 92%|#########2| 122571/132723 [00:01&lt;00:00, 87870.97KB/s]
- 99%|#########9| 131485/132723 [00:01&lt;00:00, 88248.87KB/s]
-100%|##########| 132723/132723 [00:01&lt;00:00, 77254.29KB/s]
+  1%|          | 1233/132723 [00:00&lt;00:10, 12289.05KB/s]
+  3%|2         | 3719/132723 [00:00&lt;00:06, 19669.36KB/s]
+  7%|7         | 9384/132723 [00:00&lt;00:03, 36539.49KB/s]
+ 13%|#3        | 17902/132723 [00:00&lt;00:02, 55729.40KB/s]
+ 20%|#9        | 26501/132723 [00:00&lt;00:01, 66634.48KB/s]
+ 26%|##6       | 35066/132723 [00:00&lt;00:01, 73095.74KB/s]
+ 33%|###2      | 43717/132723 [00:00&lt;00:01, 77476.55KB/s]
+ 39%|###9      | 52365/132723 [00:00&lt;00:01, 80338.75KB/s]
+ 46%|####5     | 61016/132723 [00:00&lt;00:00, 82266.36KB/s]
+ 52%|#####2    | 69667/132723 [00:01&lt;00:00, 83573.07KB/s]
+ 59%|#####9    | 78313/132723 [00:01&lt;00:00, 84453.78KB/s]
+ 66%|######5   | 86977/132723 [00:01&lt;00:00, 85114.80KB/s]
+ 72%|#######2  | 95662/132723 [00:01&lt;00:00, 85638.38KB/s]
+ 79%|#######8  | 104311/132723 [00:01&lt;00:00, 85888.13KB/s]
+ 85%|########5 | 112983/132723 [00:01&lt;00:00, 86136.53KB/s]
+ 92%|#########1| 121599/132723 [00:01&lt;00:00, 86140.40KB/s]
+ 98%|#########8| 130254/132723 [00:01&lt;00:00, 86262.16KB/s]
+100%|##########| 132723/132723 [00:01&lt;00:00, 76651.91KB/s]
 </pre></div>
 </div>
 <p>Create TVM runtime and do inference
@@ -524,7 +524,7 @@ Downloading /workspace/.mxnet/models/ssd_512_resnet50_v1_voc-9c8b225a.zip from h
 <span class="n">plt</span><span class="o">.</span><span class="n">show</span><span class="p">()</span>
 </pre></div>
 </div>
-<img src="../../_images/sphx_glr_deploy_ssd_gluoncv_001.png" srcset="../../_images/sphx_glr_deploy_ssd_gluoncv_001.png" alt="deploy ssd gluoncv" class = "sphx-glr-single-img"/><p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 3 minutes  48.374 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  48.677 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 89afdad936..2b78e5db37 100644
--- a/docs/how_to/deploy_models/sg_execution_times.html
+++ b/docs/how_to/deploy_models/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-deploy-models-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>16:20.971</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
+<p><strong>16:16.916</strong> total execution time for <strong>how_to_deploy_models</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 86%" />
@@ -354,43 +354,43 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="deploy_ssd_gluoncv.html#sphx-glr-how-to-deploy-models-deploy-ssd-gluoncv-py"><span class="std std-ref">Deploy Single Shot Multibox Detector(SSD) model</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_ssd_gluoncv.py</span></code>)</p></td>
-<td><p>03:48.374</p></td>
+<td><p>03:48.677</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>03:39.868</p></td>
+<td><p>03:38.121</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>02:30.194</p></td>
+<td><p>02:26.632</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>01:37.937</p></td>
+<td><p>01:39.156</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>01:18.206</p></td>
+<td><p>01:17.766</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 Adrenoâ„¢</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_adreno.py</span></code>)</p></td>
-<td><p>00:55.133</p></td>
+<td><p>00:55.145</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="deploy_model_on_adreno_tvmc.html#sphx-glr-how-to-deploy-models-deploy-model-on-adreno-tvmc-py"><span class="std std-ref">Deploy the Pretrained Model on Adrenoâ„¢ with tvmc Interface</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_adreno_tvmc.py</span></code>)</p></td>
-<td><p>00:51.213</p></td>
+<td><p>00:51.542</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="deploy_model_on_android.html#sphx-glr-how-to-deploy-models-deploy-model-on-android-py"><span class="std std-ref">Deploy the Pretrained Model on Android</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_android.py</span></code>)</p></td>
-<td><p>00:43.500</p></td>
+<td><p>00:43.172</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="deploy_model_on_nano.html#sphx-glr-how-to-deploy-models-deploy-model-on-nano-py"><span class="std std-ref">Deploy the Pretrained Model on Jetson Nano</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_nano.py</span></code>)</p></td>
-<td><p>00:28.498</p></td>
+<td><p>00:28.682</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="deploy_model_on_rasp.html#sphx-glr-how-to-deploy-models-deploy-model-on-rasp-py"><span class="std std-ref">Deploy the Pretrained Model on Raspberry Pi</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_model_on_rasp.py</span></code>)</p></td>
-<td><p>00:28.043</p></td>
+<td><p>00:28.017</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="deploy_sparse.html#sphx-glr-how-to-deploy-models-deploy-sparse-py"><span class="std std-ref">Deploy a Hugging Face Pruned Model on CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">deploy_sparse.py</span></code>)</p></td>
diff --git a/docs/how_to/extend_tvm/bring_your_own_datatypes.html b/docs/how_to/extend_tvm/bring_your_own_datatypes.html
index e4cb3df4a0..de7fd44e26 100644
--- a/docs/how_to/extend_tvm/bring_your_own_datatypes.html
+++ b/docs/how_to/extend_tvm/bring_your_own_datatypes.html
@@ -624,7 +624,7 @@ In this alpha state of the Bring Your Own Datatypes framework, we have not imple
 <span class="n">module</span><span class="p">,</span> <a href="https://docs.python.org/3/library/stdtypes.html#dict" title="builtins.dict" class="sphx-glr-backref-module-builtins sphx-glr-backref-type-py-class sphx-glr-backref-instance"><span class="n">params</span></a> <span class="o">=</span> <span class="n">get_mobilenet</span><span class="p">()</span>
 </pre></div>
 </div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Downloading /workspace/.mxnet/models/mobilenet0.25-9f83e440.zip76c23e2b-be0e-45c2-b357-b52ba13382ed 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.zipd1bde805-15e1-4a94-8ac8-775af63ae770 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 b76590b7cf..7b9c08b85d 100644
--- a/docs/how_to/extend_tvm/sg_execution_times.html
+++ b/docs/how_to/extend_tvm/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-extend-tvm-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:55.422</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
+<p><strong>00:55.248</strong> total execution time for <strong>how_to_extend_tvm</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 84%" />
@@ -354,15 +354,15 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="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>00:51.500</p></td>
+<td><p>00:51.342</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>00:02.813</p></td>
+<td><p>00:02.802</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>00:01.102</p></td>
+<td><p>00:01.097</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 5d897db973..d91682678c 100644
--- a/docs/how_to/extend_tvm/use_pass_instrument.html
+++ b/docs/how_to/extend_tvm/use_pass_instrument.html
@@ -531,10 +531,10 @@ profile the execution time of each passes.</p>
 </pre></div>
 </div>
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Printing results of timing profile...
-InferType: 22343us [22343us] (47.86%; 47.86%)
-FoldScaleAxis: 24339us [8us] (52.14%; 52.14%)
-        FoldConstant: 24330us [1883us] (52.12%; 99.97%)
-                InferType: 22447us [22447us] (48.09%; 92.26%)
+InferType: 22041us [22041us] (48.58%; 48.58%)
+FoldScaleAxis: 23332us [7us] (51.42%; 51.42%)
+        FoldConstant: 23325us [1701us] (51.41%; 99.97%)
+                InferType: 21625us [21625us] (47.66%; 92.71%)
 </pre></div>
 </div>
 </div>
@@ -556,10 +556,10 @@ Refer to following sections and <a class="reference internal" href="../../refere
 </pre></div>
 </div>
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Printing results of timing profile...
-InferType: 22191us [22191us] (48.42%; 48.42%)
-FoldScaleAxis: 23642us [7us] (51.58%; 51.58%)
-        FoldConstant: 23634us [1796us] (51.57%; 99.97%)
-                InferType: 21839us [21839us] (47.65%; 92.40%)
+InferType: 21885us [21885us] (48.26%; 48.26%)
+FoldScaleAxis: 23467us [6us] (51.74%; 51.74%)
+        FoldConstant: 23461us [1748us] (51.73%; 99.97%)
+                InferType: 21713us [21713us] (47.88%; 92.55%)
 </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 a099016ee4..a5769e15f6 100644
--- a/docs/how_to/optimize_operators/opt_conv_cuda.html
+++ b/docs/how_to/optimize_operators/opt_conv_cuda.html
@@ -580,7 +580,7 @@ latency of convolution.</p>
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Convolution: </span><span class="si">%f</span><span class="s2"> ms&quot;</span> <span class="o">%</span> <span class="p">(</span><span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">w</span><span class="p">,</span> <span class="n">b</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span> <span class="o">*</span> <span cl [...]
 </pre></div>
 </div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Convolution: 54.226974 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Convolution: 40.327167 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 80ddd9278f..9e98777816 100644
--- a/docs/how_to/optimize_operators/opt_conv_tensorcore.html
+++ b/docs/how_to/optimize_operators/opt_conv_tensorcore.html
@@ -862,7 +862,7 @@ be able to run on our build server</p>
     <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;conv2d with tensor core: </span><span class="si">%f</span><span class="s2"> ms&quot;</span> <span class="o">%</span> <span class="p">(</span><span class="n">evaluator</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">w</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span> <span class="o">* [...]
 </pre></div>
 </div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>conv2d with tensor core: 13.375407 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>conv2d with tensor core: 13.350711 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 153a228cff..f90f210f38 100644
--- a/docs/how_to/optimize_operators/opt_gemm.html
+++ b/docs/how_to/optimize_operators/opt_gemm.html
@@ -477,8 +477,8 @@ Then we write a baseline implementation, the simplest way to write a matrix mult
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Baseline: </span><span class="si">%f</span><span class="s2">&quot;</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: 0.018962
-Baseline: 3.307704
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Numpy running time: 0.018989
+Baseline: 3.314976
 </pre></div>
 </div>
 <p>In TVM, we can always inspect lower level IR to debug or optimize our schedule.
@@ -537,7 +537,7 @@ fill 32 * 32 * sizeof(float) which is 4KB in the cache whose total size is 32KB
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Opt1: </span><span class="si">%f</span><span class="s2">&quot;</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: 0.303799
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt1: 0.309745
 </pre></div>
 </div>
 <p>Here is the generated IR after blocking.</p>
@@ -594,7 +594,7 @@ vastly.</p>
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Opt2: </span><span class="si">%f</span><span class="s2">&quot;</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: 0.339524
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt2: 0.341284
 </pre></div>
 </div>
 <p>Here is the generated IR after vectorization.</p>
@@ -649,7 +649,7 @@ the access pattern for A matrix is more cache friendly.</p>
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Opt3: </span><span class="si">%f</span><span class="s2">&quot;</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: 0.117044
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt3: 0.116433
 </pre></div>
 </div>
 <p>Here is the generated IR after loop permutation.</p>
@@ -726,7 +726,7 @@ flattening.</p>
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Opt4: </span><span class="si">%f</span><span class="s2">&quot;</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: 0.114789
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt4: 0.109233
 </pre></div>
 </div>
 <p>Here is the generated IR after array packing.</p>
@@ -804,7 +804,7 @@ write to C when all the block results are ready.</p>
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Opt5: </span><span class="si">%f</span><span class="s2">&quot;</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: 0.110691
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt5: 0.110824
 </pre></div>
 </div>
 <p>Here is the generated IR after blocking.</p>
@@ -884,7 +884,7 @@ class Module:
 <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Opt6: </span><span class="si">%f</span><span class="s2">&quot;</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: 0.146671
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Opt6: 0.147740
 </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 7802faa9f7..b54e8223e1 100644
--- a/docs/how_to/optimize_operators/sg_execution_times.html
+++ b/docs/how_to/optimize_operators/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-optimize-operators-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:35.049</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
+<p><strong>00:35.022</strong> total execution time for <strong>how_to_optimize_operators</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 83%" />
@@ -354,15 +354,15 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="opt_gemm.html#sphx-glr-how-to-optimize-operators-opt-gemm-py"><span class="std std-ref">How to optimize GEMM on CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">opt_gemm.py</span></code>)</p></td>
-<td><p>00:32.194</p></td>
+<td><p>00:32.313</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>00:01.707</p></td>
+<td><p>00:01.627</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>00:01.147</p></td>
+<td><p>00:01.083</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 af4c349a16..63552516ef 100644
--- a/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
+++ b/docs/how_to/tune_with_autoscheduler/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-tune-with-autoscheduler-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>10:02.671</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
+<p><strong>10:06.943</strong> total execution time for <strong>how_to_tune_with_autoscheduler</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 85%" />
@@ -354,27 +354,27 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="tune_conv2d_layer_cuda.html#sphx-glr-how-to-tune-with-autoscheduler-tune-conv2d-layer-cuda-py"><span class="std std-ref">Auto-scheduling a Convolution Layer for GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_conv2d_layer_cuda.py</span></code>)</p></td>
-<td><p>06:10.917</p></td>
+<td><p>06:12.035</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>01:42.867</p></td>
+<td><p>01:42.823</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="tune_network_cuda.html#sphx-glr-how-to-tune-with-autoscheduler-tune-network-cuda-py"><span class="std std-ref">Auto-scheduling a Neural Network for NVIDIA GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_network_cuda.py</span></code>)</p></td>
-<td><p>01:07.853</p></td>
+<td><p>01:08.262</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="tune_sparse_x86.html#sphx-glr-how-to-tune-with-autoscheduler-tune-sparse-x86-py"><span class="std std-ref">Auto-scheduling Sparse Matrix Multiplication on CPU with Custom Sketch Rule</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_sparse_x86.py</span></code>)</p></td>
-<td><p>00:32.813</p></td>
+<td><p>00:35.774</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>00:14.364</p></td>
+<td><p>00:14.335</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>00:13.856</p></td>
+<td><p>00:13.715</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 551771c0f4..f2c88864c1 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
@@ -510,769 +510,479 @@ class Module:
     @T.prim_func
     def main(data: T.Buffer((1, 512, 7, 7), &quot;float32&quot;), kernel: T.Buffer((512, 512, 3, 3), &quot;float32&quot;), bias: T.Buffer((1, 512, 1, 1), &quot;float32&quot;), compute: T.Buffer((1, 512, 7, 7), &quot;float32&quot;)):
         T.func_attr({&quot;from_legacy_te_schedule&quot;: True, &quot;global_symbol&quot;: &quot;main&quot;, &quot;tir.noalias&quot;: True})
-        blockIdx_x = T.launch_thread(&quot;blockIdx.x&quot;, 8)
-        conv2d_nchw = T.allocate([28], &quot;float32&quot;, &quot;local&quot;)
-        pad_temp_shared = T.allocate([504], &quot;float32&quot;, &quot;shared&quot;)
-        kernel_shared = T.allocate([1536], &quot;float32&quot;, &quot;shared&quot;)
-        threadIdx_x = T.launch_thread(&quot;threadIdx.x&quot;, 112)
-        conv2d_nchw_1 = T.Buffer((196,), data=conv2d_nchw, scope=&quot;local&quot;, align=32)
+        blockIdx_x = T.launch_thread(&quot;blockIdx.x&quot;, 28)
+        conv2d_nchw = T.allocate([14], &quot;float32&quot;, &quot;local&quot;)
+        pad_temp_shared = T.allocate([72], &quot;float32&quot;, &quot;shared&quot;)
+        kernel_shared = T.allocate([3072], &quot;float32&quot;, &quot;shared&quot;)
+        threadIdx_x = T.launch_thread(&quot;threadIdx.x&quot;, 64)
+        conv2d_nchw_1 = T.Buffer((14,), data=conv2d_nchw, scope=&quot;local&quot;, align=32)
         conv2d_nchw_1[0] = T.float32(0)
-        conv2d_nchw_1[14] = T.float32(0)
         conv2d_nchw_1[1] = T.float32(0)
-        conv2d_nchw_1[15] = T.float32(0)
         conv2d_nchw_1[2] = T.float32(0)
-        conv2d_nchw_1[16] = T.float32(0)
         conv2d_nchw_1[3] = T.float32(0)
-        conv2d_nchw_1[17] = T.float32(0)
         conv2d_nchw_1[4] = T.float32(0)
-        conv2d_nchw_1[18] = T.float32(0)
         conv2d_nchw_1[5] = T.float32(0)
-        conv2d_nchw_1[19] = T.float32(0)
         conv2d_nchw_1[6] = T.float32(0)
-        conv2d_nchw_1[20] = T.float32(0)
         conv2d_nchw_1[7] = T.float32(0)
-        conv2d_nchw_1[21] = T.float32(0)
         conv2d_nchw_1[8] = T.float32(0)
-        conv2d_nchw_1[22] = T.float32(0)
         conv2d_nchw_1[9] = T.float32(0)
-        conv2d_nchw_1[23] = T.float32(0)
         conv2d_nchw_1[10] = T.float32(0)
-        conv2d_nchw_1[24] = T.float32(0)
         conv2d_nchw_1[11] = T.float32(0)
-        conv2d_nchw_1[25] = T.float32(0)
         conv2d_nchw_1[12] = T.float32(0)
-        conv2d_nchw_1[26] = T.float32(0)
         conv2d_nchw_1[13] = T.float32(0)
-        conv2d_nchw_1[27] = T.float32(0)
         for rc_outer_outer, ry_outer_outer in T.grid(64, 3):
-            cse_var_4: T.int32 = rc_outer_outer * 392
-            cse_var_3: T.int32 = ry_outer_outer * 7
             cse_var_2: T.int32 = rc_outer_outer * 72
             cse_var_1: T.int32 = ry_outer_outer * 3
+            pad_temp_shared_1 = T.Buffer((72,), data=pad_temp_shared, scope=&quot;shared&quot;)
+            with T.launch_thread(&quot;threadIdx.x&quot;, 64) as threadIdx_x_1:
+                data_1 = T.Buffer((25088,), data=data.data)
+                if T.likely(threadIdx_x_1 &lt; 18):
+                    pad_temp_shared_1[threadIdx_x_1 * 4] = T.if_then_else(1 &lt;= ry_outer_outer + blockIdx_x % 7 and ry_outer_outer + blockIdx_x % 7 &lt; 8 and 1 &lt;= threadIdx_x_1 * 4 % 9 and threadIdx_x_1 * 4 % 9 &lt; 8, data_1[rc_outer_outer * 392 + threadIdx_x_1 * 4 // 9 * 49 + ry_outer_outer * 7 + blockIdx_x % 7 * 7 + threadIdx_x_1 * 4 % 9 - 8], T.float32(0))
+                if T.likely(threadIdx_x_1 &lt; 18):
+                    pad_temp_shared_1[threadIdx_x_1 * 4 + 1] = T.if_then_else(1 &lt;= ry_outer_outer + blockIdx_x % 7 and ry_outer_outer + blockIdx_x % 7 &lt; 8 and 1 &lt;= (threadIdx_x_1 * 4 + 1) % 9 and (threadIdx_x_1 * 4 + 1) % 9 &lt; 8, data_1[rc_outer_outer * 392 + (threadIdx_x_1 * 4 + 1) // 9 * 49 + ry_outer_outer * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 * 4 + 1) % 9 - 8], T.float32(0))
+                if T.likely(threadIdx_x_1 &lt; 18):
+                    pad_temp_shared_1[threadIdx_x_1 * 4 + 2] = T.if_then_else(1 &lt;= ry_outer_outer + blockIdx_x % 7 and ry_outer_outer + blockIdx_x % 7 &lt; 8 and 1 &lt;= (threadIdx_x_1 * 4 + 2) % 9 and (threadIdx_x_1 * 4 + 2) % 9 &lt; 8, data_1[rc_outer_outer * 392 + (threadIdx_x_1 * 4 + 2) // 9 * 49 + ry_outer_outer * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 * 4 + 2) % 9 - 8], T.float32(0))
+                if T.likely(threadIdx_x_1 &lt; 18):
+                    pad_temp_shared_1[threadIdx_x_1 * 4 + 3] = T.if_then_else(1 &lt;= ry_outer_outer + blockIdx_x % 7 and ry_outer_outer + blockIdx_x % 7 &lt; 8 and 1 &lt;= (threadIdx_x_1 * 4 + 3) % 9 and (threadIdx_x_1 * 4 + 3) % 9 &lt; 8, data_1[rc_outer_outer * 392 + (threadIdx_x_1 * 4 + 3) // 9 * 49 + ry_outer_outer * 7 + blockIdx_x % 7 * 7 + (threadIdx_x_1 * 4 + 3) % 9 - 8], T.float32(0))
             threadIdx_x_1 = T.env_thread(&quot;threadIdx.x&quot;)
-            pad_temp_shared_1 = T.Buffer((504,), data=pad_temp_shared, scope=&quot;shared&quot;)
-            data_1 = T.Buffer((25088,), data=data.data)
-            with T.launch_thread(threadIdx_x_1, 112):
-                pad_temp_shared_1[threadIdx_x_1] = T.if_then_else(1 &lt;= threadIdx_x_1 % 63 // 9 + ry_outer_outer and threadIdx_x_1 % 63 // 9 + ry_outer_outer &lt; 8 and 1 &lt;= threadIdx_x_1 % 9 and threadIdx_x_1 % 9 &lt; 8, data_1[cse_var_4 + threadIdx_x_1 // 9 * 7 + cse_var_3 + threadIdx_x_1 % 9 - 8], T.float32(0))
-            with T.launch_thread(threadIdx_x_1, 112):
-                pad_temp_shared_1[threadIdx_x_1 + 112] = T.if_then_else(1 &lt;= (threadIdx_x_1 + 49) % 63 // 9 + ry_outer_outer and (threadIdx_x_1 + 49) % 63 // 9 + ry_outer_outer &lt; 8 and 1 &lt;= (threadIdx_x_1 + 4) % 9 and (threadIdx_x_1 + 4) % 9 &lt; 8, data_1[cse_var_4 + (threadIdx_x_1 + 112) // 9 * 7 + cse_var_3 + (threadIdx_x_1 + 4) % 9 - 8], T.float32(0))
-            with T.launch_thread(threadIdx_x_1, 112):
-                pad_temp_shared_1[threadIdx_x_1 + 224] = T.if_then_else(1 &lt;= (threadIdx_x_1 + 35) % 63 // 9 + ry_outer_outer and (threadIdx_x_1 + 35) % 63 // 9 + ry_outer_outer &lt; 8 and 1 &lt;= (threadIdx_x_1 + 8) % 9 and (threadIdx_x_1 + 8) % 9 &lt; 8, data_1[cse_var_4 + (threadIdx_x_1 + 224) // 9 * 7 + cse_var_3 + (threadIdx_x_1 + 8) % 9 - 8], T.float32(0))
-            with T.launch_thread(threadIdx_x_1, 112):
-                pad_temp_shared_1[threadIdx_x_1 + 336] = T.if_then_else(1 &lt;= (threadIdx_x_1 + 21) % 63 // 9 + ry_outer_outer and (threadIdx_x_1 + 21) % 63 // 9 + ry_outer_outer &lt; 8 and 1 &lt;= (threadIdx_x_1 + 3) % 9 and (threadIdx_x_1 + 3) % 9 &lt; 8, data_1[cse_var_4 + (threadIdx_x_1 + 336) // 9 * 7 + cse_var_3 + (threadIdx_x_1 + 3) % 9 - 8], T.float32(0))
-            with T.launch_thread(threadIdx_x_1, 112):
-                if T.likely(threadIdx_x_1 &lt; 56):
-                    pad_temp_shared_1[threadIdx_x_1 + 448] = T.if_then_else(1 &lt;= (threadIdx_x_1 + 7) // 9 + ry_outer_outer and (threadIdx_x_1 + 7) // 9 + ry_outer_outer &lt; 8 and 1 &lt;= (threadIdx_x_1 + 7) % 9 and (threadIdx_x_1 + 7) % 9 &lt; 8, data_1[cse_var_4 + (threadIdx_x_1 + 448) // 9 * 7 + cse_var_3 + (threadIdx_x_1 + 7) % 9 - 8], T.float32(0))
-            threadIdx_x_2 = T.env_thread(&quot;threadIdx.x&quot;)
-            kernel_shared_1 = T.Buffer((1536,), data=kernel_shared, scope=&quot;shared&quot;)
+            kernel_shared_1 = T.Buffer((3072,), data=kernel_shared, scope=&quot;shared&quot;)
             kernel_1 = T.Buffer((2359296,), data=kernel.data)
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[threadIdx_x_2] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[(threadIdx_x_2 + 112) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 112) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[(threadIdx_x_2 + 224) // 24 * 24 + (threadIdx_x_2 + 8) % 24 // 3 * 3 + (threadIdx_x_2 + 2) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 224) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 2) % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[threadIdx_x_2 + 336] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3 + 64512]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[(threadIdx_x_2 + 448) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 448) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[(threadIdx_x_2 + 560) // 24 * 24 + (threadIdx_x_2 + 8) % 24 // 3 * 3 + (threadIdx_x_2 + 2) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 560) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 2) % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[threadIdx_x_2 + 672] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3 + 129024]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[(threadIdx_x_2 + 784) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 784) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[(threadIdx_x_2 + 896) // 24 * 24 + (threadIdx_x_2 + 8) % 24 // 3 * 3 + (threadIdx_x_2 + 2) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 896) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 2) % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[threadIdx_x_2 + 1008] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3 + 193536]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[(threadIdx_x_2 + 1120) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 1120) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[(threadIdx_x_2 + 1232) // 24 * 24 + (threadIdx_x_2 + 8) % 24 // 3 * 3 + (threadIdx_x_2 + 2) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 1232) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 2) % 3]
-            with T.launch_thread(threadIdx_x_2, 112):
-                kernel_shared_1[threadIdx_x_2 + 1344] = kernel_1[blockIdx_x * 294912 + threadIdx_x_2 // 24 * 4608 + cse_var_2 + threadIdx_x_2 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_2 % 3 + 258048]
-            with T.launch_thread(threadIdx_x_2, 112):
-                if T.likely(threadIdx_x_2 &lt; 80):
-                    kernel_shared_1[(threadIdx_x_2 + 1456) // 24 * 24 + (threadIdx_x_2 + 16) % 24 // 3 * 3 + (threadIdx_x_2 + 1) % 3] = kernel_1[blockIdx_x * 294912 + (threadIdx_x_2 + 1456) // 24 * 4608 + cse_var_2 + (threadIdx_x_2 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_2 + 1) % 3]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9] * kernel_shared_1[threadIdx_x // 7 * 48]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 63] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 63] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 126] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 126] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 189] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 189] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 768]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 3]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 771]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 6]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 774]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 9]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 777]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 63] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 63] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 126] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 126] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 189] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 189] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 24]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 792]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 27]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 795]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 30]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 798]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 33]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 801]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 1]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 769]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 4]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 772]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 7]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 775]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 10]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 778]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 1] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 64] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 127] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 190] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 25]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 793]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 28]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 796]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 31]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 799]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 34]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 802]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 8] * kernel_shared_1[threadIdx_x // 7 * 48 + 2]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 8] * kernel_shared_1[threadIdx_x // 7 * 48 + 770]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 71] * kernel_shared_1[threadIdx_x // 7 * 48 + 5]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 71] * kernel_shared_1[threadIdx_x // 7 * 48 + 773]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 134] * kernel_shared_1[threadIdx_x // 7 * 48 + 8]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 134] * kernel_shared_1[threadIdx_x // 7 * 48 + 776]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 197] * kernel_shared_1[threadIdx_x // 7 * 48 + 11]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 197] * kernel_shared_1[threadIdx_x // 7 * 48 + 779]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 2] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 65] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 128] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 191] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 3] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 66] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 129] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 192] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 4] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 67] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 130] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 193] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 5] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 68] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 131] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 194] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 6] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 69] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 132] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 195] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 7] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 70] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 133] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 196] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 8] * kernel_shared_1[threadIdx_x // 7 * 48 + 26]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 8] * kernel_shared_1[threadIdx_x // 7 * 48 + 794]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 71] * kernel_shared_1[threadIdx_x // 7 * 48 + 29]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 71] * kernel_shared_1[threadIdx_x // 7 * 48 + 797]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 134] * kernel_shared_1[threadIdx_x // 7 * 48 + 32]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 134] * kernel_shared_1[threadIdx_x // 7 * 48 + 800]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 197] * kernel_shared_1[threadIdx_x // 7 * 48 + 35]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 197] * kernel_shared_1[threadIdx_x // 7 * 48 + 803]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 252] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 252] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 315] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 315] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 378] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 378] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 441] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 441] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 12]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 780]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 15]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 783]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 18]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 786]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 21]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 789]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 252] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 252] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 315] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 315] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 378] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 378] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 441] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 441] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 36]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 804]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 39]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 807]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 42]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 810]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 45]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 813]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 13]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 781]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 16]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 784]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 19]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 787]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 22]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 790]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 253] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 316] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 379] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 442] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 37]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 805]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 40]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 808]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 43]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 811]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 46]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 814]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-            conv2d_nchw_1[14] = conv2d_nchw_1[14] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-            conv2d_nchw_1[15] = conv2d_nchw_1[15] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-            conv2d_nchw_1[16] = conv2d_nchw_1[16] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-            conv2d_nchw_1[17] = conv2d_nchw_1[17] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-            conv2d_nchw_1[18] = conv2d_nchw_1[18] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-            conv2d_nchw_1[19] = conv2d_nchw_1[19] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 260] * kernel_shared_1[threadIdx_x // 7 * 48 + 14]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 260] * kernel_shared_1[threadIdx_x // 7 * 48 + 782]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 323] * kernel_shared_1[threadIdx_x // 7 * 48 + 17]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 323] * kernel_shared_1[threadIdx_x // 7 * 48 + 785]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 386] * kernel_shared_1[threadIdx_x // 7 * 48 + 20]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 386] * kernel_shared_1[threadIdx_x // 7 * 48 + 788]
-            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 449] * kernel_shared_1[threadIdx_x // 7 * 48 + 23]
-            conv2d_nchw_1[20] = conv2d_nchw_1[20] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 449] * kernel_shared_1[threadIdx_x // 7 * 48 + 791]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 254] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 317] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 380] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-            conv2d_nchw_1[21] = conv2d_nchw_1[21] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 443] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 255] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 318] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 381] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-            conv2d_nchw_1[22] = conv2d_nchw_1[22] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 444] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 256] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 319] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 382] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-            conv2d_nchw_1[23] = conv2d_nchw_1[23] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 445] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 257] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 320] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 383] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-            conv2d_nchw_1[24] = conv2d_nchw_1[24] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 446] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 258] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 321] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 384] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-            conv2d_nchw_1[25] = conv2d_nchw_1[25] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 447] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 259] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 322] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 385] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-            conv2d_nchw_1[26] = conv2d_nchw_1[26] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 448] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 260] * kernel_shared_1[threadIdx_x // 7 * 48 + 38]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 260] * kernel_shared_1[threadIdx_x // 7 * 48 + 806]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 323] * kernel_shared_1[threadIdx_x // 7 * 48 + 41]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 323] * kernel_shared_1[threadIdx_x // 7 * 48 + 809]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 386] * kernel_shared_1[threadIdx_x // 7 * 48 + 44]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 386] * kernel_shared_1[threadIdx_x // 7 * 48 + 812]
-            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 449] * kernel_shared_1[threadIdx_x // 7 * 48 + 47]
-            conv2d_nchw_1[27] = conv2d_nchw_1[27] + pad_temp_shared_1[threadIdx_x % 7 * 9 + 449] * kernel_shared_1[threadIdx_x // 7 * 48 + 815]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 64) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 64) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 128) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 128) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 192] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 36864]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 256) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 256) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 320) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 320) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 384] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 73728]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 448) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 448) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 512) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 512) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 576] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 110592]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 640) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 640) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 704) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 704) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 768] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 147456]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 832) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 832) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 896) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 896) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 960] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 184320]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1024) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1024) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1088) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1088) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 1152] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 221184]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1216) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1216) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1280) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1280) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 1344] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 258048]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1408) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1408) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1472) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1472) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 1536] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 294912]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1600) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1600) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1664) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1664) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 1728] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 331776]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1792) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1792) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1856) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1856) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 1920] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 368640]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 1984) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 1984) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2048) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2048) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 2112] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 405504]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2176) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2176) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2240) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2240) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 2304] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 442368]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2368) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2368) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2432) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2432) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 2496] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 479232]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2560) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2560) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2624) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2624) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 2688] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 516096]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2752) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2752) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2816) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2816) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[threadIdx_x_1 + 2880] = kernel_1[blockIdx_x // 7 * 589824 + threadIdx_x_1 // 24 * 4608 + cse_var_2 + threadIdx_x_1 % 24 // 3 * 9 + cse_var_1 + threadIdx_x_1 % 3 + 552960]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 2944) // 24 * 24 + (threadIdx_x_1 + 16) % 24 // 3 * 3 + (threadIdx_x_1 + 1) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 2944) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 16) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 1) % 3]
+            with T.launch_thread(threadIdx_x_1, 64):
+                kernel_shared_1[(threadIdx_x_1 + 3008) // 24 * 24 + (threadIdx_x_1 + 8) % 24 // 3 * 3 + (threadIdx_x_1 + 2) % 3] = kernel_1[blockIdx_x // 7 * 589824 + (threadIdx_x_1 + 3008) // 24 * 4608 + cse_var_2 + (threadIdx_x_1 + 8) % 24 // 3 * 9 + cse_var_1 + (threadIdx_x_1 + 2) % 3]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[0] * kernel_shared_1[threadIdx_x * 48]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[9] * kernel_shared_1[threadIdx_x * 48 + 3]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 48]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 48 + 3]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 3]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 3]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 3]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 3]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 3]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[0] * kernel_shared_1[threadIdx_x * 48 + 24]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[9] * kernel_shared_1[threadIdx_x * 48 + 27]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 48 + 24]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 48 + 27]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 24]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 27]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 24]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 27]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 24]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 27]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 24]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 27]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 24]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 27]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 48 + 1]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 48 + 4]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 1]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 4]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 1]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 4]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 1]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 4]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 1]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 4]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 1]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 4]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 48 + 1]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 48 + 4]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[1] * kernel_shared_1[threadIdx_x * 48 + 25]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[10] * kernel_shared_1[threadIdx_x * 48 + 28]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 25]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 28]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 25]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 28]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 25]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 28]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 25]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 28]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 25]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 28]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 48 + 25]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 48 + 28]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 2]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 5]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 2]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 5]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 2]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 5]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 2]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 5]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 2]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 5]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 48 + 2]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 48 + 5]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[8] * kernel_shared_1[threadIdx_x * 48 + 2]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[17] * kernel_shared_1[threadIdx_x * 48 + 5]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[2] * kernel_shared_1[threadIdx_x * 48 + 26]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[11] * kernel_shared_1[threadIdx_x * 48 + 29]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[3] * kernel_shared_1[threadIdx_x * 48 + 26]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[12] * kernel_shared_1[threadIdx_x * 48 + 29]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[4] * kernel_shared_1[threadIdx_x * 48 + 26]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[13] * kernel_shared_1[threadIdx_x * 48 + 29]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[5] * kernel_shared_1[threadIdx_x * 48 + 26]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[14] * kernel_shared_1[threadIdx_x * 48 + 29]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[6] * kernel_shared_1[threadIdx_x * 48 + 26]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[15] * kernel_shared_1[threadIdx_x * 48 + 29]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[7] * kernel_shared_1[threadIdx_x * 48 + 26]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[16] * kernel_shared_1[threadIdx_x * 48 + 29]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[8] * kernel_shared_1[threadIdx_x * 48 + 26]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[17] * kernel_shared_1[threadIdx_x * 48 + 29]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[18] * kernel_shared_1[threadIdx_x * 48 + 6]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[27] * kernel_shared_1[threadIdx_x * 48 + 9]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 48 + 6]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 48 + 9]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 6]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 9]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 6]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 9]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 6]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 9]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 6]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 9]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 6]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 9]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[18] * kernel_shared_1[threadIdx_x * 48 + 30]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[27] * kernel_shared_1[threadIdx_x * 48 + 33]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 48 + 30]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 48 + 33]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 30]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 33]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 30]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 33]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 30]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 33]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 30]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 33]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 30]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 33]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 48 + 7]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 48 + 10]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 7]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 10]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 7]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 10]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 7]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 10]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 7]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 10]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 7]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 10]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 48 + 7]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 48 + 10]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[19] * kernel_shared_1[threadIdx_x * 48 + 31]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[28] * kernel_shared_1[threadIdx_x * 48 + 34]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 31]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 34]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 31]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 34]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 31]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 34]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 31]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 34]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 31]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 34]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 48 + 31]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 48 + 34]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 8]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 11]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 8]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 11]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 8]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 11]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 8]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 11]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 8]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 11]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 48 + 8]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 48 + 11]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[26] * kernel_shared_1[threadIdx_x * 48 + 8]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[35] * kernel_shared_1[threadIdx_x * 48 + 11]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[20] * kernel_shared_1[threadIdx_x * 48 + 32]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[29] * kernel_shared_1[threadIdx_x * 48 + 35]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[21] * kernel_shared_1[threadIdx_x * 48 + 32]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[30] * kernel_shared_1[threadIdx_x * 48 + 35]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[22] * kernel_shared_1[threadIdx_x * 48 + 32]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[31] * kernel_shared_1[threadIdx_x * 48 + 35]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[23] * kernel_shared_1[threadIdx_x * 48 + 32]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[32] * kernel_shared_1[threadIdx_x * 48 + 35]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[24] * kernel_shared_1[threadIdx_x * 48 + 32]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[33] * kernel_shared_1[threadIdx_x * 48 + 35]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[25] * kernel_shared_1[threadIdx_x * 48 + 32]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[34] * kernel_shared_1[threadIdx_x * 48 + 35]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[26] * kernel_shared_1[threadIdx_x * 48 + 32]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[35] * kernel_shared_1[threadIdx_x * 48 + 35]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[36] * kernel_shared_1[threadIdx_x * 48 + 12]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[45] * kernel_shared_1[threadIdx_x * 48 + 15]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 48 + 12]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 48 + 15]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 12]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 15]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 12]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 15]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 12]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 15]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 12]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 15]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 12]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 15]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[36] * kernel_shared_1[threadIdx_x * 48 + 36]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[45] * kernel_shared_1[threadIdx_x * 48 + 39]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 48 + 36]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 48 + 39]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 36]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 39]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 36]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 39]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 36]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 39]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 36]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 39]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 36]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 39]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 48 + 13]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 48 + 16]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 13]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 16]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 13]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 16]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 13]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 16]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 13]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 16]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 13]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 16]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 48 + 13]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 48 + 16]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[37] * kernel_shared_1[threadIdx_x * 48 + 37]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[46] * kernel_shared_1[threadIdx_x * 48 + 40]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 37]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 40]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 37]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 40]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 37]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 40]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 37]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 40]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 37]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 40]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 48 + 37]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 48 + 40]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 14]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 17]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 14]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 17]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 14]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 17]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 14]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 17]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 14]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 17]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 48 + 14]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 48 + 17]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[44] * kernel_shared_1[threadIdx_x * 48 + 14]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[53] * kernel_shared_1[threadIdx_x * 48 + 17]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[38] * kernel_shared_1[threadIdx_x * 48 + 38]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[47] * kernel_shared_1[threadIdx_x * 48 + 41]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[39] * kernel_shared_1[threadIdx_x * 48 + 38]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[48] * kernel_shared_1[threadIdx_x * 48 + 41]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[40] * kernel_shared_1[threadIdx_x * 48 + 38]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[49] * kernel_shared_1[threadIdx_x * 48 + 41]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[41] * kernel_shared_1[threadIdx_x * 48 + 38]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[50] * kernel_shared_1[threadIdx_x * 48 + 41]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[42] * kernel_shared_1[threadIdx_x * 48 + 38]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[51] * kernel_shared_1[threadIdx_x * 48 + 41]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[43] * kernel_shared_1[threadIdx_x * 48 + 38]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[52] * kernel_shared_1[threadIdx_x * 48 + 41]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[44] * kernel_shared_1[threadIdx_x * 48 + 38]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[53] * kernel_shared_1[threadIdx_x * 48 + 41]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[54] * kernel_shared_1[threadIdx_x * 48 + 18]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[63] * kernel_shared_1[threadIdx_x * 48 + 21]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 48 + 18]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 48 + 21]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 18]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 21]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 18]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 21]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 18]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 21]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 18]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 21]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 18]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 21]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[54] * kernel_shared_1[threadIdx_x * 48 + 42]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[63] * kernel_shared_1[threadIdx_x * 48 + 45]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 48 + 42]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 48 + 45]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 42]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 45]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 42]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 45]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 42]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 45]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 42]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 45]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 42]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 45]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 48 + 19]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 48 + 22]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 19]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 22]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 19]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 22]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 19]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 22]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 19]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 22]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 19]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 22]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 48 + 19]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 48 + 22]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[55] * kernel_shared_1[threadIdx_x * 48 + 43]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[64] * kernel_shared_1[threadIdx_x * 48 + 46]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 43]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 46]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 43]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 46]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 43]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 46]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 43]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 46]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 43]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 46]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 48 + 43]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 48 + 46]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 20]
+            conv2d_nchw_1[0] = conv2d_nchw_1[0] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 23]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 20]
+            conv2d_nchw_1[1] = conv2d_nchw_1[1] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 23]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 20]
+            conv2d_nchw_1[2] = conv2d_nchw_1[2] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 23]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 20]
+            conv2d_nchw_1[3] = conv2d_nchw_1[3] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 23]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 20]
+            conv2d_nchw_1[4] = conv2d_nchw_1[4] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 23]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 48 + 20]
+            conv2d_nchw_1[5] = conv2d_nchw_1[5] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 48 + 23]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[62] * kernel_shared_1[threadIdx_x * 48 + 20]
+            conv2d_nchw_1[6] = conv2d_nchw_1[6] + pad_temp_shared_1[71] * kernel_shared_1[threadIdx_x * 48 + 23]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[56] * kernel_shared_1[threadIdx_x * 48 + 44]
+            conv2d_nchw_1[7] = conv2d_nchw_1[7] + pad_temp_shared_1[65] * kernel_shared_1[threadIdx_x * 48 + 47]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[57] * kernel_shared_1[threadIdx_x * 48 + 44]
+            conv2d_nchw_1[8] = conv2d_nchw_1[8] + pad_temp_shared_1[66] * kernel_shared_1[threadIdx_x * 48 + 47]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[58] * kernel_shared_1[threadIdx_x * 48 + 44]
+            conv2d_nchw_1[9] = conv2d_nchw_1[9] + pad_temp_shared_1[67] * kernel_shared_1[threadIdx_x * 48 + 47]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[59] * kernel_shared_1[threadIdx_x * 48 + 44]
+            conv2d_nchw_1[10] = conv2d_nchw_1[10] + pad_temp_shared_1[68] * kernel_shared_1[threadIdx_x * 48 + 47]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[60] * kernel_shared_1[threadIdx_x * 48 + 44]
+            conv2d_nchw_1[11] = conv2d_nchw_1[11] + pad_temp_shared_1[69] * kernel_shared_1[threadIdx_x * 48 + 47]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[61] * kernel_shared_1[threadIdx_x * 48 + 44]
+            conv2d_nchw_1[12] = conv2d_nchw_1[12] + pad_temp_shared_1[70] * kernel_shared_1[threadIdx_x * 48 + 47]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[62] * kernel_shared_1[threadIdx_x * 48 + 44]
+            conv2d_nchw_1[13] = conv2d_nchw_1[13] + pad_temp_shared_1[71] * kernel_shared_1[threadIdx_x * 48 + 47]
         for i1_inner, i3_inner in T.grid(2, 7):
-            cse_var_5: T.int32 = i1_inner * 7 + i3_inner
             compute_1 = T.Buffer((25088,), data=compute.data)
             bias_1 = T.Buffer((512,), data=bias.data)
-            compute_1[blockIdx_x * 3136 + threadIdx_x // 7 * 98 + i1_inner * 49 + threadIdx_x % 7 * 7 + i3_inner] = T.max(conv2d_nchw_1[cse_var_5] + bias_1[blockIdx_x * 64 + threadIdx_x // 7 * 2 + i1_inner], T.float32(0))
-            compute_1[blockIdx_x * 3136 + threadIdx_x // 7 * 98 + i1_inner * 49 + threadIdx_x % 7 * 7 + i3_inner + 1568] = T.max(conv2d_nchw_1[cse_var_5 + 14] + bias_1[blockIdx_x * 64 + threadIdx_x // 7 * 2 + i1_inner + 32], T.float32(0))
+            compute_1[blockIdx_x // 7 * 6272 + threadIdx_x * 98 + i1_inner * 49 + blockIdx_x % 7 * 7 + i3_inner] = T.max(conv2d_nchw_1[i1_inner * 7 + i3_inner] + bias_1[blockIdx_x // 7 * 128 + threadIdx_x * 2 + i1_inner], T.float32(0))
 </pre></div>
 </div>
 </div>
@@ -1306,7 +1016,7 @@ 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: 0.381 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 0.362 ms
 </pre></div>
 </div>
 </div>
@@ -1337,18 +1047,18 @@ 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=2)
-conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=16)
-conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=2)
+conv2d_nchw_ff_o_o_o_i, conv2d_nchw_ff_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_i, factor=64)
+conv2d_nchw_ff_o_o_o_o, conv2d_nchw_ff_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_ff_o_o_o_i, factor=1)
 conv2d_nchw_yy_o_i, conv2d_nchw_yy_i = s[conv2d_nchw].split(conv2d_nchw_yy, factor=1)
 conv2d_nchw_yy_o_o_i, conv2d_nchw_yy_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_i, factor=1)
-conv2d_nchw_yy_o_o_o_i, conv2d_nchw_yy_o_o_i = s[conv2d_nchw].split(conv2d_nchw_yy_o_o_i, factor=7)
+conv2d_nchw_yy_o_o_o_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_o_i, conv2d_nchw_xx_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_i, factor=1)
 conv2d_nchw_xx_o_o_o_o, conv2d_nchw_xx_o_o_o_i = s[conv2d_nchw].split(conv2d_nchw_xx_o_o_o_i, factor=1)
-conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=4)
-conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=2)
+conv2d_nchw_rc_o_i, conv2d_nchw_rc_i = s[conv2d_nchw].split(conv2d_nchw_rc, factor=2)
+conv2d_nchw_rc_o_o, conv2d_nchw_rc_o_i = s[conv2d_nchw].split(conv2d_nchw_rc_o_i, factor=4)
 conv2d_nchw_ry_o_i, conv2d_nchw_ry_i = s[conv2d_nchw].split(conv2d_nchw_ry, factor=1)
 conv2d_nchw_ry_o_o, conv2d_nchw_ry_o_i = s[conv2d_nchw].split(conv2d_nchw_ry_o_i, factor=1)
 conv2d_nchw_rx_o_i, conv2d_nchw_rx_i = s[conv2d_nchw].split(conv2d_nchw_rx, factor=1)
@@ -1358,10 +1068,10 @@ compute_i0_o_i, compute_i0_i = s[compute].split(compute_i0, factor=1)
 compute_i0_o_o_i, compute_i0_o_i = s[compute].split(compute_i0_o_i, factor=1)
 compute_i0_o_o_o, compute_i0_o_o_i = s[compute].split(compute_i0_o_o_i, factor=1)
 compute_i1_o_i, compute_i1_i = s[compute].split(compute_i1, factor=2)
-compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=16)
-compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=2)
+compute_i1_o_o_i, compute_i1_o_i = s[compute].split(compute_i1_o_i, factor=64)
+compute_i1_o_o_o, compute_i1_o_o_i = s[compute].split(compute_i1_o_o_i, factor=1)
 compute_i2_o_i, compute_i2_i = s[compute].split(compute_i2, factor=1)
-compute_i2_o_o_i, compute_i2_o_i = s[compute].split(compute_i2_o_i, factor=7)
+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_o_i, compute_i3_o_i = s[compute].split(compute_i3_o_i, factor=1)
@@ -1384,14 +1094,14 @@ 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=112)
+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=64)
 s[kernel_shared].bind(kernel_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis(&quot;threadIdx.x&quot;))
 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)
+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=4)
 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=112)
+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=64)
 s[pad_temp_shared].bind(pad_temp_shared_ax0_ax1_fused_ax2_fused_ax3_fused_o_i, te.thread_axis(&quot;threadIdx.x&quot;))
-s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, &quot;auto_unroll_max_step&quot;, 1024)
+s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, &quot;auto_unroll_max_step&quot;, 512)
 s[conv2d_nchw].pragma(conv2d_nchw_nn_o_o_o_o, &quot;unroll_explicit&quot;, True)
 
 CUDA source code:
@@ -1409,743 +1119,429 @@ CUDA source code:
   #define int64_t long long
   #define uint64_t unsigned long long
 #endif
-extern &quot;C&quot; __global__ void __launch_bounds__(112) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
-  float conv2d_nchw[28];
-  __shared__ float pad_temp_shared[504];
-  __shared__ float kernel_shared[1536];
+extern &quot;C&quot; __global__ void __launch_bounds__(64) default_function_kernel0(float* __restrict__ data, float* __restrict__ kernel, float* __restrict__ compute, float* __restrict__ bias) {
+  float conv2d_nchw[14];
+  __shared__ float pad_temp_shared[72];
+  __shared__ float kernel_shared[3072];
   conv2d_nchw[0] = 0.000000e+00f;
-  conv2d_nchw[14] = 0.000000e+00f;
   conv2d_nchw[1] = 0.000000e+00f;
-  conv2d_nchw[15] = 0.000000e+00f;
   conv2d_nchw[2] = 0.000000e+00f;
-  conv2d_nchw[16] = 0.000000e+00f;
   conv2d_nchw[3] = 0.000000e+00f;
-  conv2d_nchw[17] = 0.000000e+00f;
   conv2d_nchw[4] = 0.000000e+00f;
-  conv2d_nchw[18] = 0.000000e+00f;
   conv2d_nchw[5] = 0.000000e+00f;
-  conv2d_nchw[19] = 0.000000e+00f;
   conv2d_nchw[6] = 0.000000e+00f;
-  conv2d_nchw[20] = 0.000000e+00f;
   conv2d_nchw[7] = 0.000000e+00f;
-  conv2d_nchw[21] = 0.000000e+00f;
   conv2d_nchw[8] = 0.000000e+00f;
-  conv2d_nchw[22] = 0.000000e+00f;
   conv2d_nchw[9] = 0.000000e+00f;
-  conv2d_nchw[23] = 0.000000e+00f;
   conv2d_nchw[10] = 0.000000e+00f;
-  conv2d_nchw[24] = 0.000000e+00f;
   conv2d_nchw[11] = 0.000000e+00f;
-  conv2d_nchw[25] = 0.000000e+00f;
   conv2d_nchw[12] = 0.000000e+00f;
-  conv2d_nchw[26] = 0.000000e+00f;
   conv2d_nchw[13] = 0.000000e+00f;
-  conv2d_nchw[27] = 0.000000e+00f;
   for (int rc_outer_outer = 0; rc_outer_outer &lt; 64; ++rc_outer_outer) {
     for (int ry_outer_outer = 0; ry_outer_outer &lt; 3; ++ry_outer_outer) {
       __syncthreads();
-      pad_temp_shared[((int)threadIdx.x)] = (((((1 &lt;= (((((int)threadIdx.x) % 63) / 9) + ry_outer_outer)) &amp;&amp; ((((((int)threadIdx.x) % 63) / 9) + ry_outer_outer) &lt; 8)) &amp;&amp; (1 &lt;= (((int)threadIdx.x) % 9))) &amp;&amp; ((((int)threadIdx.x) % 9) &lt; 8)) ? data[(((((rc_outer_outer * 392) + ((((int)threadIdx.x) / 9) * 7)) + (ry_outer_outer * 7)) + (((int)threadIdx.x) % 9)) - 8)] : 0.000000e+00f);
-      pad_temp_shared[(((int)threadIdx.x) + 112)] = (((((1 &lt;= ((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer)) &amp;&amp; (((((((int)threadIdx.x) + 49) % 63) / 9) + ry_outer_outer) &lt; 8)) &amp;&amp; (1 &lt;= ((((int)threadIdx.x) + 4) % 9))) &amp;&amp; (((((int)threadIdx.x) + 4) % 9) &lt; 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 112) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 4) % 9)) - 8)] : 0.000000e+00f);
-      pad_temp_shared[(((int)threadIdx.x) + 224)] = (((((1 &lt;= ((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer)) &amp;&amp; (((((((int)threadIdx.x) + 35) % 63) / 9) + ry_outer_outer) &lt; 8)) &amp;&amp; (1 &lt;= ((((int)threadIdx.x) + 8) % 9))) &amp;&amp; (((((int)threadIdx.x) + 8) % 9) &lt; 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 224) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 8) % 9)) - 8)] : 0.000000e+00f);
-      pad_temp_shared[(((int)threadIdx.x) + 336)] = (((((1 &lt;= ((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer)) &amp;&amp; (((((((int)threadIdx.x) + 21) % 63) / 9) + ry_outer_outer) &lt; 8)) &amp;&amp; (1 &lt;= ((((int)threadIdx.x) + 3) % 9))) &amp;&amp; (((((int)threadIdx.x) + 3) % 9) &lt; 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 336) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 3) % 9)) - 8)] : 0.000000e+00f);
-      if (((int)threadIdx.x) &lt; 56) {
-        pad_temp_shared[(((int)threadIdx.x) + 448)] = (((((1 &lt;= (((((int)threadIdx.x) + 7) / 9) + ry_outer_outer)) &amp;&amp; ((((((int)threadIdx.x) + 7) / 9) + ry_outer_outer) &lt; 8)) &amp;&amp; (1 &lt;= ((((int)threadIdx.x) + 7) % 9))) &amp;&amp; (((((int)threadIdx.x) + 7) % 9) &lt; 8)) ? data[(((((rc_outer_outer * 392) + (((((int)threadIdx.x) + 448) / 9) * 7)) + (ry_outer_outer * 7)) + ((((int)threadIdx.x) + 7) % 9)) - 8)] : 0.000000e+00f);
+      if (((int)threadIdx.x) &lt; 18) {
+        pad_temp_shared[(((int)threadIdx.x) * 4)] = (((((1 &lt;= (ry_outer_outer + (((int)blockIdx.x) % 7))) &amp;&amp; ((ry_outer_outer + (((int)blockIdx.x) % 7)) &lt; 8)) &amp;&amp; (1 &lt;= ((((int)threadIdx.x) * 4) % 9))) &amp;&amp; (((((int)threadIdx.x) * 4) % 9) &lt; 8)) ? data[((((((rc_outer_outer * 392) + (((((int)threadIdx.x) * 4) / 9) * 49)) + (ry_outer_outer * 7)) + ((((int)blockIdx.x) % 7) * 7)) + ((((int)threadIdx.x) * 4) % 9)) - 8)] : 0.000000e+00f);
       }
-      kernel_shared[((int)threadIdx.x)] = kernel[((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
-      kernel_shared[(((((((int)threadIdx.x) + 112) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 112) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
-      kernel_shared[(((((((int)threadIdx.x) + 224) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 224) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
-      kernel_shared[(((int)threadIdx.x) + 336)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 64512)];
-      kernel_shared[(((((((int)threadIdx.x) + 448) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
-      kernel_shared[(((((((int)threadIdx.x) + 560) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 560) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
-      kernel_shared[(((int)threadIdx.x) + 672)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 129024)];
-      kernel_shared[(((((((int)threadIdx.x) + 784) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 784) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
-      kernel_shared[(((((((int)threadIdx.x) + 896) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
-      kernel_shared[(((int)threadIdx.x) + 1008)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 193536)];
-      kernel_shared[(((((((int)threadIdx.x) + 1120) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1120) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
-      kernel_shared[(((((((int)threadIdx.x) + 1232) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1232) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
-      kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[(((((((((int)blockIdx.x) * 294912) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 258048)];
-      if (((int)threadIdx.x) &lt; 80) {
-        kernel_shared[(((((((int)threadIdx.x) + 1456) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[((((((((int)blockIdx.x) * 294912) + (((((int)threadIdx.x) + 1456) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      if (((int)threadIdx.x) &lt; 18) {
+        pad_temp_shared[((((int)threadIdx.x) * 4) + 1)] = (((((1 &lt;= (ry_outer_outer + (((int)blockIdx.x) % 7))) &amp;&amp; ((ry_outer_outer + (((int)blockIdx.x) % 7)) &lt; 8)) &amp;&amp; (1 &lt;= (((((int)threadIdx.x) * 4) + 1) % 9))) &amp;&amp; ((((((int)threadIdx.x) * 4) + 1) % 9) &lt; 8)) ? data[((((((rc_outer_outer * 392) + ((((((int)threadIdx.x) * 4) + 1) / 9) * 49)) + (ry_outer_outer * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((((int)threadIdx.x) * 4) + 1) % 9)) - 8)] : 0.000000e+00f);
       }
+      if (((int)threadIdx.x) &lt; 18) {
+        pad_temp_shared[((((int)threadIdx.x) * 4) + 2)] = (((((1 &lt;= (ry_outer_outer + (((int)blockIdx.x) % 7))) &amp;&amp; ((ry_outer_outer + (((int)blockIdx.x) % 7)) &lt; 8)) &amp;&amp; (1 &lt;= (((((int)threadIdx.x) * 4) + 2) % 9))) &amp;&amp; ((((((int)threadIdx.x) * 4) + 2) % 9) &lt; 8)) ? data[((((((rc_outer_outer * 392) + ((((((int)threadIdx.x) * 4) + 2) / 9) * 49)) + (ry_outer_outer * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((((int)threadIdx.x) * 4) + 2) % 9)) - 8)] : 0.000000e+00f);
+      }
+      if (((int)threadIdx.x) &lt; 18) {
+        pad_temp_shared[((((int)threadIdx.x) * 4) + 3)] = (((((1 &lt;= (ry_outer_outer + (((int)blockIdx.x) % 7))) &amp;&amp; ((ry_outer_outer + (((int)blockIdx.x) % 7)) &lt; 8)) &amp;&amp; (1 &lt;= (((((int)threadIdx.x) * 4) + 3) % 9))) &amp;&amp; ((((((int)threadIdx.x) * 4) + 3) % 9) &lt; 8)) ? data[((((((rc_outer_outer * 392) + ((((((int)threadIdx.x) * 4) + 3) / 9) * 49)) + (ry_outer_outer * 7)) + ((((int)blockIdx.x) % 7) * 7)) + (((((int)threadIdx.x) * 4) + 3) % 9)) - 8)] : 0.000000e+00f);
+      }
+      kernel_shared[((int)threadIdx.x)] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 64) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 64) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 128) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 128) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 192)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 36864)];
+      kernel_shared[(((((((int)threadIdx.x) + 256) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 256) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 320) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 320) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 384)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 73728)];
+      kernel_shared[(((((((int)threadIdx.x) + 448) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 448) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 512) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 512) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 576)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 110592)];
+      kernel_shared[(((((((int)threadIdx.x) + 640) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 640) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 704) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 704) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 768)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 147456)];
+      kernel_shared[(((((((int)threadIdx.x) + 832) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 832) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 896) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 896) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 960)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 184320)];
+      kernel_shared[(((((((int)threadIdx.x) + 1024) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1024) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 1088) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1088) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 1152)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 221184)];
+      kernel_shared[(((((((int)threadIdx.x) + 1216) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1216) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 1280) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1280) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 1344)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 258048)];
+      kernel_shared[(((((((int)threadIdx.x) + 1408) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1408) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 1472) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1472) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 1536)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 294912)];
+      kernel_shared[(((((((int)threadIdx.x) + 1600) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1600) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 1664) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1664) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 1728)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 331776)];
+      kernel_shared[(((((((int)threadIdx.x) + 1792) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1792) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 1856) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1856) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 1920)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 368640)];
+      kernel_shared[(((((((int)threadIdx.x) + 1984) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 1984) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 2048) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2048) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 2112)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 405504)];
+      kernel_shared[(((((((int)threadIdx.x) + 2176) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2176) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 2240) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2240) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 2304)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 442368)];
+      kernel_shared[(((((((int)threadIdx.x) + 2368) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2368) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 2432) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2432) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 2496)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 479232)];
+      kernel_shared[(((((((int)threadIdx.x) + 2560) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2560) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 2624) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2624) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 2688)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 516096)];
+      kernel_shared[(((((((int)threadIdx.x) + 2752) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2752) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 2816) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2816) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
+      kernel_shared[(((int)threadIdx.x) + 2880)] = kernel[((((((((((int)blockIdx.x) / 7) * 589824) + ((((int)threadIdx.x) / 24) * 4608)) + (rc_outer_outer * 72)) + (((((int)threadIdx.x) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + (((int)threadIdx.x) % 3)) + 552960)];
+      kernel_shared[(((((((int)threadIdx.x) + 2944) / 24) * 24) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 1) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 2944) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 16) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 1) % 3))];
+      kernel_shared[(((((((int)threadIdx.x) + 3008) / 24) * 24) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 3)) + ((((int)threadIdx.x) + 2) % 3))] = kernel[(((((((((int)blockIdx.x) / 7) * 589824) + (((((int)threadIdx.x) + 3008) / 24) * 4608)) + (rc_outer_outer * 72)) + ((((((int)threadIdx.x) + 8) % 24) / 3) * 9)) + (ry_outer_outer * 3)) + ((((int)threadIdx.x) + 2) % 3))];
       __syncthreads();
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 63)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 63)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 126)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 126)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[((((int)threadIdx.x) / 7) * 48)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 768)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 3)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 771)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 6)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 774)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 9)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 777)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[((((int)threadIdx.x) % 7) * 9)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 63)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 63)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 126)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 126)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 189)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 24)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 792)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 27)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 795)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 30)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 798)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 33)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 801)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 1)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 769)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 4)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 772)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 7)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 775)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 10)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 778)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 1)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 64)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 127)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 190)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 25)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 793)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 28)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 796)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 31)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 799)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 34)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 802)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 2)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 770)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 71)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 5)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 71)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 773)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 134)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 8)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 134)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 776)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 11)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 779)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 2)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 65)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 128)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 191)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 3)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 66)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 129)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 192)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 4)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 67)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 130)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 193)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 5)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 68)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 131)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 194)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 6)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 69)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 132)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 195)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 7)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 70)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 133)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 196)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 26)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 8)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 794)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 71)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 29)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 71)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 797)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 134)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 32)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 134)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 800)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 35)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 197)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 803)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 315)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 315)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 441)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 441)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 12)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 780)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 15)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 783)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 18)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 786)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 21)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 789)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 252)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 315)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 315)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 378)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 441)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 441)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 36)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 804)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 39)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 807)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 42)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 810)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 45)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 813)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 13)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 781)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 16)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 784)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 19)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 787)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 22)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 790)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 253)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 316)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 379)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 442)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 37)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 805)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 40)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 808)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 43)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 811)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 46)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 814)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-      conv2d_nchw[14] = (conv2d_nchw[14] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-      conv2d_nchw[15] = (conv2d_nchw[15] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-      conv2d_nchw[16] = (conv2d_nchw[16] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-      conv2d_nchw[17] = (conv2d_nchw[17] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-      conv2d_nchw[18] = (conv2d_nchw[18] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-      conv2d_nchw[19] = (conv2d_nchw[19] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 14)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 782)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 323)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 17)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 323)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 785)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 20)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 788)]));
-      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 449)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 23)]));
-      conv2d_nchw[20] = (conv2d_nchw[20] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 449)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 791)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 254)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 317)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 380)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-      conv2d_nchw[21] = (conv2d_nchw[21] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 443)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 255)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 318)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 381)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-      conv2d_nchw[22] = (conv2d_nchw[22] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 444)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 256)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 319)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 382)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-      conv2d_nchw[23] = (conv2d_nchw[23] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 445)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 257)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 320)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 383)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-      conv2d_nchw[24] = (conv2d_nchw[24] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 446)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 258)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 321)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 384)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-      conv2d_nchw[25] = (conv2d_nchw[25] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 447)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 259)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 322)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 385)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-      conv2d_nchw[26] = (conv2d_nchw[26] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 448)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 38)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 260)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 806)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 323)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 41)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 323)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 809)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 44)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 386)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 812)]));
-      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 449)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 47)]));
-      conv2d_nchw[27] = (conv2d_nchw[27] + (pad_temp_shared[(((((int)threadIdx.x) % 7) * 9) + 449)] * kernel_shared[(((((int)threadIdx.x) / 7) * 48) + 815)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[0] * kernel_shared[(((int)threadIdx.x) * 48)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[9] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[1] * kernel_shared[(((int)threadIdx.x) * 48)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[2] * kernel_shared[(((int)threadIdx.x) * 48)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[3] * kernel_shared[(((int)threadIdx.x) * 48)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[4] * kernel_shared[(((int)threadIdx.x) * 48)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[5] * kernel_shared[(((int)threadIdx.x) * 48)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[6] * kernel_shared[(((int)threadIdx.x) * 48)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 3)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[0] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[9] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[1] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 24)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 27)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[1] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 48) + 1)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 48) + 4)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[1] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[10] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 48) + 25)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 48) + 28)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[8] * kernel_shared[((((int)threadIdx.x) * 48) + 2)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[17] * kernel_shared[((((int)threadIdx.x) * 48) + 5)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[2] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[11] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[3] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[12] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[4] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[13] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[5] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[14] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[6] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[15] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[7] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[16] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[8] * kernel_shared[((((int)threadIdx.x) * 48) + 26)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[17] * kernel_shared[((((int)threadIdx.x) * 48) + 29)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[18] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[27] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 6)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 9)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[18] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[27] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 30)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 33)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 48) + 7)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 48) + 10)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[19] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[28] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 48) + 31)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 48) + 34)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[26] * kernel_shared[((((int)threadIdx.x) * 48) + 8)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[35] * kernel_shared[((((int)threadIdx.x) * 48) + 11)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[20] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[29] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[21] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[30] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[22] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[31] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[23] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[32] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[24] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[33] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[25] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[34] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[26] * kernel_shared[((((int)threadIdx.x) * 48) + 32)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[35] * kernel_shared[((((int)threadIdx.x) * 48) + 35)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[36] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[45] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 12)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 15)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[36] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[45] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 36)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 39)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 48) + 13)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 48) + 16)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[37] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[46] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 48) + 37)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 48) + 40)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[44] * kernel_shared[((((int)threadIdx.x) * 48) + 14)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[53] * kernel_shared[((((int)threadIdx.x) * 48) + 17)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[38] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[47] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[39] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[48] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[40] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[49] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[41] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[50] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[42] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[51] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[43] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[52] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[44] * kernel_shared[((((int)threadIdx.x) * 48) + 38)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[53] * kernel_shared[((((int)threadIdx.x) * 48) + 41)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[54] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[63] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 18)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 21)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[54] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[63] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 42)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 45)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 48) + 19)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 48) + 22)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[55] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[64] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 48) + 43)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 48) + 46)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+      conv2d_nchw[0] = (conv2d_nchw[0] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+      conv2d_nchw[1] = (conv2d_nchw[1] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+      conv2d_nchw[2] = (conv2d_nchw[2] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+      conv2d_nchw[3] = (conv2d_nchw[3] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+      conv2d_nchw[4] = (conv2d_nchw[4] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+      conv2d_nchw[5] = (conv2d_nchw[5] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[62] * kernel_shared[((((int)threadIdx.x) * 48) + 20)]));
+      conv2d_nchw[6] = (conv2d_nchw[6] + (pad_temp_shared[71] * kernel_shared[((((int)threadIdx.x) * 48) + 23)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[56] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+      conv2d_nchw[7] = (conv2d_nchw[7] + (pad_temp_shared[65] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[57] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+      conv2d_nchw[8] = (conv2d_nchw[8] + (pad_temp_shared[66] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[58] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+      conv2d_nchw[9] = (conv2d_nchw[9] + (pad_temp_shared[67] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[59] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+      conv2d_nchw[10] = (conv2d_nchw[10] + (pad_temp_shared[68] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[60] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+      conv2d_nchw[11] = (conv2d_nchw[11] + (pad_temp_shared[69] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[61] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+      conv2d_nchw[12] = (conv2d_nchw[12] + (pad_temp_shared[70] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[62] * kernel_shared[((((int)threadIdx.x) * 48) + 44)]));
+      conv2d_nchw[13] = (conv2d_nchw[13] + (pad_temp_shared[71] * kernel_shared[((((int)threadIdx.x) * 48) + 47)]));
     }
   }
   for (int i1_inner = 0; i1_inner &lt; 2; ++i1_inner) {
     for (int i3_inner = 0; i3_inner &lt; 7; ++i3_inner) {
-      compute[(((((((int)blockIdx.x) * 3136) + ((((int)threadIdx.x) / 7) * 98)) + (i1_inner * 49)) + ((((int)threadIdx.x) % 7) * 7)) + i3_inner)] = max((conv2d_nchw[((i1_inner * 7) + i3_inner)] + bias[(((((int)blockIdx.x) * 64) + ((((int)threadIdx.x) / 7) * 2)) + i1_inner)]), 0.000000e+00f);
-      compute[((((((((int)blockIdx.x) * 3136) + ((((int)threadIdx.x) / 7) * 98)) + (i1_inner * 49)) + ((((int)threadIdx.x) % 7) * 7)) + i3_inner) + 1568)] = max((conv2d_nchw[(((i1_inner * 7) + i3_inner) + 14)] + bias[((((((int)blockIdx.x) * 64) + ((((int)threadIdx.x) / 7) * 2)) + i1_inner) + 32)]), 0.000000e+00f);
+      compute[((((((((int)blockIdx.x) / 7) * 6272) + (((int)threadIdx.x) * 98)) + (i1_inner * 49)) + ((((int)blockIdx.x) % 7) * 7)) + i3_inner)] = max((conv2d_nchw[((i1_inner * 7) + i3_inner)] + bias[((((((int)blockIdx.x) / 7) * 128) + (((int)threadIdx.x) * 2)) + i1_inner)]), 0.000000e+00f);
     }
   }
 }
@@ -2181,7 +1577,7 @@ In the example below we resume the status and do more 5 trials.</p>
 Get devices for measurement successfully!
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 6 minutes  10.917 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 6 minutes  12.035 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 6df0562702..7a4c557ef6 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html
@@ -921,7 +921,7 @@ so we can read the log file and load the best schedules.</p>
 Evaluate inference time cost...
 Execution time summary:
  mean (ms)   median (ms)    max (ms)     min (ms)     std (ms)
-   7.8953       7.8927       7.9128       7.8804       0.0134
+   7.8844       7.8839       7.8919       7.8774       0.0060
 </pre></div>
 </div>
 </div>
@@ -943,7 +943,7 @@ to learn how to use the RPC Tracker and RPC Server.
 To use the RPC Tracker in auto-scheduler, replace the runner in <code class="code docutils literal notranslate"><span class="pre">TuningOptions</span></code>
 with <a class="reference internal" href="../../reference/api/python/auto_scheduler.html#tvm.auto_scheduler.RPCRunner" title="tvm.auto_scheduler.RPCRunner"><code class="xref any py py-class docutils literal notranslate"><span class="pre">auto_scheduler.RPCRunner</span></code></a>.</p></li>
 </ol>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  7.853 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  8.262 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 22bce25da3..1792824a17 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_network_x86.html
@@ -940,7 +940,7 @@ so we can read the log file and load the best schedules.</p>
 Evaluate inference time cost...
 Execution time summary:
  mean (ms)   median (ms)    max (ms)     min (ms)     std (ms)
-  754.7261     756.4411     756.4888     751.2483      2.4592
+  752.8096     752.8915     753.3811     752.1563      0.5033
 </pre></div>
 </div>
 </div>
@@ -962,7 +962,7 @@ to learn how to use the RPC Tracker and RPC Server.
 To use the RPC Tracker in auto-scheduler, replace the runner in <code class="code docutils literal notranslate"><span class="pre">TuningOptions</span></code>
 with <a class="reference internal" href="../../reference/api/python/auto_scheduler.html#tvm.auto_scheduler.RPCRunner" title="tvm.auto_scheduler.RPCRunner"><code class="xref any py py-class docutils literal notranslate"><span class="pre">auto_scheduler.RPCRunner</span></code></a>.</p></li>
 </ol>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  42.867 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  42.823 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 228f2b70a4..ba00fa4f98 100644
--- a/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
+++ b/docs/how_to/tune_with_autoscheduler/tune_sparse_x86.html
@@ -637,27 +637,149 @@ class Module:
     @T.prim_func
     def main(placeholder: T.Buffer((128, 256), &quot;float32&quot;), placeholder_1: T.Buffer((4916, 16, 1), &quot;float32&quot;), placeholder_2: T.Buffer((4916,), &quot;int32&quot;), placeholder_3: T.Buffer((33,), &quot;int32&quot;), placeholder_4: T.Buffer((128, 512), &quot;float32&quot;), compute: T.Buffer((128, 512), &quot;float32&quot;)):
         T.func_attr({&quot;from_legacy_te_schedule&quot;: True, &quot;global_symbol&quot;: &quot;main&quot;, &quot;tir.noalias&quot;: True})
-        for i0_outer_i1_outer_fused in T.parallel(64):
-            compute_1 = T.allocate([1024], &quot;float32&quot;, &quot;global&quot;)
-            compute_2 = T.Buffer((1024,), data=compute_1)
-            for i_outer_inner in range(8):
-                for i_inner_init, j_init in T.grid(8, 16):
-                    compute_2[i_outer_inner * 128 + i_inner_init * 16 + j_init] = T.float32(0)
-                for elem_idx, i_inner, j in T.grid(T.Let(placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1], where={cse_var_1: i0_outer_i1_outer_fused % 32}), 8, 16):
-                    cse_var_1 = T.int32()
+        for i0_outer_i1_outer_fused in T.parallel(512):
+            compute_1 = T.allocate([128], &quot;float32&quot;, &quot;global&quot;)
+            compute_2 = T.Buffer((128,), data=compute_1)
+            for i_outer_inner in range(4):
+                cse_var_2: T.int32 = i_outer_inner * 32
+                cse_var_1: T.int32 = i0_outer_i1_outer_fused % 64 // 2
+                compute_2[cse_var_2] = T.float32(0)
+                compute_2[cse_var_2 + 1] = T.float32(0)
+                compute_2[cse_var_2 + 2] = T.float32(0)
+                compute_2[cse_var_2 + 3] = T.float32(0)
+                compute_2[cse_var_2 + 4] = T.float32(0)
+                compute_2[cse_var_2 + 5] = T.float32(0)
+                compute_2[cse_var_2 + 6] = T.float32(0)
+                compute_2[cse_var_2 + 7] = T.float32(0)
+                compute_2[cse_var_2 + 8] = T.float32(0)
+                compute_2[cse_var_2 + 9] = T.float32(0)
+                compute_2[cse_var_2 + 10] = T.float32(0)
+                compute_2[cse_var_2 + 11] = T.float32(0)
+                compute_2[cse_var_2 + 12] = T.float32(0)
+                compute_2[cse_var_2 + 13] = T.float32(0)
+                compute_2[cse_var_2 + 14] = T.float32(0)
+                compute_2[cse_var_2 + 15] = T.float32(0)
+                compute_2[cse_var_2 + 16] = T.float32(0)
+                compute_2[cse_var_2 + 17] = T.float32(0)
+                compute_2[cse_var_2 + 18] = T.float32(0)
+                compute_2[cse_var_2 + 19] = T.float32(0)
+                compute_2[cse_var_2 + 20] = T.float32(0)
+                compute_2[cse_var_2 + 21] = T.float32(0)
+                compute_2[cse_var_2 + 22] = T.float32(0)
+                compute_2[cse_var_2 + 23] = T.float32(0)
+                compute_2[cse_var_2 + 24] = T.float32(0)
+                compute_2[cse_var_2 + 25] = T.float32(0)
+                compute_2[cse_var_2 + 26] = T.float32(0)
+                compute_2[cse_var_2 + 27] = T.float32(0)
+                compute_2[cse_var_2 + 28] = T.float32(0)
+                compute_2[cse_var_2 + 29] = T.float32(0)
+                compute_2[cse_var_2 + 30] = T.float32(0)
+                compute_2[cse_var_2 + 31] = T.float32(0)
+                for elem_idx in range(placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
                     placeholder_5 = T.Buffer((33,), &quot;int32&quot;, data=placeholder_3.data)
-                    cse_var_2: T.int32 = i0_outer_i1_outer_fused % 32
-                    if T.likely(elem_idx &lt; placeholder_5[cse_var_2 + 1] - placeholder_5[cse_var_2]):
-                        placeholder_6 = T.Buffer((78656,), data=placeholder_1.data)
-                        placeholder_7 = T.Buffer((32768,), data=placeholder.data)
-                        placeholder_8 = T.Buffer((4916,), &quot;int32&quot;, data=placeholder_2.data)
-                        cse_var_3: T.int32 = i_outer_inner * 128 + i_inner * 16 + j
-                        compute_2[cse_var_3] = compute_2[cse_var_3] + placeholder_6[placeholder_5[cse_var_2] * 16 + elem_idx * 16 + j] * T.max(placeholder_7[i0_outer_i1_outer_fused // 32 * 16384 + i_outer_inner * 2048 + i_inner * 256 + placeholder_8[placeholder_5[cse_var_2] + elem_idx]], T.float32(0))
-            for i0_inner in range(64):
-                cse_var_4: T.int32 = i0_outer_i1_outer_fused // 32 * 32768 + i0_inner * 512 + i0_outer_i1_outer_fused % 32 * 16
+                    placeholder_6 = T.Buffer((78656,), data=placeholder_1.data)
+                    placeholder_7 = T.Buffer((32768,), data=placeholder.data)
+                    placeholder_8 = T.Buffer((4916,), &quot;int32&quot;, data=placeholder_2.data)
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        compute_2[cse_var_2] = compute_2[cse_var_2] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_3: T.int32 = cse_var_2 + 1
+                        compute_2[cse_var_3] = compute_2[cse_var_3] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 1] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_4: T.int32 = cse_var_2 + 2
+                        compute_2[cse_var_4] = compute_2[cse_var_4] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 2] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_5: T.int32 = cse_var_2 + 3
+                        compute_2[cse_var_5] = compute_2[cse_var_5] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 3] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_6: T.int32 = cse_var_2 + 4
+                        compute_2[cse_var_6] = compute_2[cse_var_6] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 4] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_7: T.int32 = cse_var_2 + 5
+                        compute_2[cse_var_7] = compute_2[cse_var_7] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 5] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_8: T.int32 = cse_var_2 + 6
+                        compute_2[cse_var_8] = compute_2[cse_var_8] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 6] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_9: T.int32 = cse_var_2 + 7
+                        compute_2[cse_var_9] = compute_2[cse_var_9] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 7] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx]], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_10: T.int32 = cse_var_2 + 8
+                        compute_2[cse_var_10] = compute_2[cse_var_10] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_11: T.int32 = cse_var_2 + 9
+                        compute_2[cse_var_11] = compute_2[cse_var_11] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 1] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_12: T.int32 = cse_var_2 + 10
+                        compute_2[cse_var_12] = compute_2[cse_var_12] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 2] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_13: T.int32 = cse_var_2 + 11
+                        compute_2[cse_var_13] = compute_2[cse_var_13] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 3] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_14: T.int32 = cse_var_2 + 12
+                        compute_2[cse_var_14] = compute_2[cse_var_14] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 4] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_15: T.int32 = cse_var_2 + 13
+                        compute_2[cse_var_15] = compute_2[cse_var_15] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 5] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_16: T.int32 = cse_var_2 + 14
+                        compute_2[cse_var_16] = compute_2[cse_var_16] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 6] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_17: T.int32 = cse_var_2 + 15
+                        compute_2[cse_var_17] = compute_2[cse_var_17] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 7] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 256], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_18: T.int32 = cse_var_2 + 16
+                        compute_2[cse_var_18] = compute_2[cse_var_18] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_19: T.int32 = cse_var_2 + 17
+                        compute_2[cse_var_19] = compute_2[cse_var_19] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 1] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_20: T.int32 = cse_var_2 + 18
+                        compute_2[cse_var_20] = compute_2[cse_var_20] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 2] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_21: T.int32 = cse_var_2 + 19
+                        compute_2[cse_var_21] = compute_2[cse_var_21] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 3] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_22: T.int32 = cse_var_2 + 20
+                        compute_2[cse_var_22] = compute_2[cse_var_22] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 4] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_23: T.int32 = cse_var_2 + 21
+                        compute_2[cse_var_23] = compute_2[cse_var_23] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 5] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_24: T.int32 = cse_var_2 + 22
+                        compute_2[cse_var_24] = compute_2[cse_var_24] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 6] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_25: T.int32 = cse_var_2 + 23
+                        compute_2[cse_var_25] = compute_2[cse_var_25] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 7] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 512], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_26: T.int32 = cse_var_2 + 24
+                        compute_2[cse_var_26] = compute_2[cse_var_26] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_27: T.int32 = cse_var_2 + 25
+                        compute_2[cse_var_27] = compute_2[cse_var_27] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 1] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_28: T.int32 = cse_var_2 + 26
+                        compute_2[cse_var_28] = compute_2[cse_var_28] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 2] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_29: T.int32 = cse_var_2 + 27
+                        compute_2[cse_var_29] = compute_2[cse_var_29] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 3] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_30: T.int32 = cse_var_2 + 28
+                        compute_2[cse_var_30] = compute_2[cse_var_30] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 4] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_31: T.int32 = cse_var_2 + 29
+                        compute_2[cse_var_31] = compute_2[cse_var_31] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 5] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_32: T.int32 = cse_var_2 + 30
+                        compute_2[cse_var_32] = compute_2[cse_var_32] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 6] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+                    if T.likely(elem_idx &lt; placeholder_5[cse_var_1 + 1] - placeholder_5[cse_var_1]):
+                        cse_var_33: T.int32 = cse_var_2 + 31
+                        compute_2[cse_var_33] = compute_2[cse_var_33] + placeholder_6[placeholder_5[cse_var_1] * 16 + elem_idx * 16 + i0_outer_i1_outer_fused % 2 * 8 + 7] * T.max(placeholder_7[i0_outer_i1_outer_fused // 64 * 4096 + i_outer_inner * 1024 + placeholder_8[placeholder_5[cse_var_1] + elem_idx] + 768], T.float32(0))
+            for i0_inner in range(16):
+                cse_var_34: T.int32 = i0_outer_i1_outer_fused // 64 * 8192 + i0_inner * 512 + i0_outer_i1_outer_fused % 64 * 8
                 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 + 16] = T.max(compute_2[i0_inner * 16:i0_inner * 16 + 16] + placeholder_5[cse_var_4:cse_var_4 + 16], T.Broadcast(T.float32(0), 16))
+                compute_3[cse_var_34:cse_var_34 + 8] = T.max(compute_2[i0_inner * 8:i0_inner * 8 + 8] + placeholder_5[cse_var_34:cse_var_34 + 8], T.Broadcast(T.float32(0), 8))
 </pre></div>
 </div>
 </div>
@@ -691,7 +813,7 @@ class Module:
 <span class="p">)</span>
 </pre></div>
 </div>
-<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 1.546 ms
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time of this operator: 1.918 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 2e26759e36..ba818f7651 100644
--- a/docs/how_to/tune_with_autotvm/sg_execution_times.html
+++ b/docs/how_to/tune_with_autotvm/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-tune-with-autotvm-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:43.881</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
+<p><strong>00:36.058</strong> total execution time for <strong>how_to_tune_with_autotvm</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 84%" />
@@ -354,22 +354,22 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="tune_conv2d_cuda.html#sphx-glr-how-to-tune-with-autotvm-tune-conv2d-cuda-py"><span class="std std-ref">Tuning High Performance Convolution on NVIDIA GPUs</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_conv2d_cuda.py</span></code>)</p></td>
-<td><p>00:43.847</p></td>
+<td><p>00:36.024</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>00:00.021</p></td>
+<td><p>00:00.020</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="tune_relay_cuda.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-cuda-py"><span class="std std-ref">Auto-tuning a Convolutional Network for NVIDIA GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_cuda.py</span></code>)</p></td>
 <td><p>00:00.005</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
-<tr class="row-even"><td><p><a class="reference internal" href="tune_relay_mobile_gpu.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-mobile-gpu-py"><span class="std std-ref">Auto-tuning a Convolutional Network for Mobile GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_mobile_gpu.py</span></code>)</p></td>
-<td><p>00:00.005</p></td>
+<tr class="row-even"><td><p><a class="reference internal" href="tune_relay_arm.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-arm-py"><span class="std std-ref">Auto-tuning a Convolutional Network for ARM CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_arm.py</span></code>)</p></td>
+<td><p>00:00.004</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
-<tr class="row-odd"><td><p><a class="reference internal" href="tune_relay_arm.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-arm-py"><span class="std std-ref">Auto-tuning a Convolutional Network for ARM CPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_arm.py</span></code>)</p></td>
+<tr class="row-odd"><td><p><a class="reference internal" href="tune_relay_mobile_gpu.html#sphx-glr-how-to-tune-with-autotvm-tune-relay-mobile-gpu-py"><span class="std std-ref">Auto-tuning a Convolutional Network for Mobile GPU</span></a> (<code class="docutils literal notranslate"><span class="pre">tune_relay_mobile_gpu.py</span></code>)</p></td>
 <td><p>00:00.004</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
diff --git a/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html b/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
index f2a45daeec..54998fce45 100644
--- a/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
+++ b/docs/how_to/tune_with_autotvm/tune_conv2d_cuda.html
@@ -573,7 +573,8 @@ for this template</p>
 waiting for device...
 device available
 Get devices for measurement successfully!
-No: 1   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+No: 1   GFLOPS: 142.96/142.96   result: MeasureResult(costs=(0.0016193390869565216,), error_no=MeasureErrorNo.NO_ERROR, all_cost=7.002098798751831, timestamp=1679387620.600002)        [(&#39;tile_f&#39;, [-1, 2, 8, 2]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 4, 8]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,10364940
+No: 2   GFLOPS: 0.00/142.96     result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -695,8 +696,10 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 4, 2]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 16, 16]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 0)],None,3233414
-No: 2   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 2, 1, 256]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 1]), (&#39;tile_rc&#39;, [-1, 8, 4]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,9177077
+No: 3   GFLOPS: 0.98/142.96     result: MeasureResult(costs=(0.23536286925,), error_no=MeasureErrorNo.NO_ERROR, all_cost=7.865373849868774, timestamp=1679387625.744464)        [(&#39;tile_f&#39;, [-1, 1, 1, 256]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 4, 8]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 1)],None,7074976
+No: 4   GFLOPS: 1278.20/1278.20 result: MeasureResult(costs=(0.00018111530568181819,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.7274179458618164, timestamp=1679387626.8660135)     [(&#39;tile_f&#39;, [-1, 2, 8, 1]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 4, 2]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 1)],None,8563748
+No: 5   GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -818,8 +821,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 1, 1, 128]), (&#39;tile_y&#39;, [-1, 1, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 2, 128]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 0)],None,1727650
-No: 3   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 128, 2, 2]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 64, 4]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 0)],None,2414351
+No: 6   GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -941,8 +944,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 8, 8, 2]), (&#39;tile_y&#39;, [-1, 1, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 1]), (&#39;tile_rc&#39;, [-1, 2, 128]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,10436882
-No: 4   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 1, 128]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 2, 256]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,9093912
+No: 7   GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -1064,8 +1067,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 16, 2]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 2, 8]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 1)],None,6875967
-No: 5   GFLOPS: 0.00/0.00       result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 16, 2, 2]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 8, 64]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,9075948
+No: 8   GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -1187,9 +1190,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 2, 1, 16]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 2, 32]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 1)],None,8665085
-No: 6   GFLOPS: 25.12/25.12     result: MeasureResult(costs=(0.009215251235294116,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.549553871154785, timestamp=1679386605.377113) [(&#39;tile_f&#39;, [-1, 2, 1, 2]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 8, 2]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 1)],None,7986716
-No: 7   GFLOPS: 0.00/25.12      result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 128, 1]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 4, 16]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 1)],None,6518431
+No: 9   GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -1311,8 +1313,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 2, 4, 4]), (&#39;tile_y&#39;, [-1, 1, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 1]), (&#39;tile_rc&#39;, [-1, 4, 128]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 1)],None,8698036
-No: 8   GFLOPS: 0.00/25.12      result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 32, 16, 1]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 4, 64]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 0)],None,361059
+No: 10  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -1434,28 +1436,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 512, 1, 1]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 32, 4]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 0)],None,1055789
-No: 9   GFLOPS: 3.66/25.12      result: MeasureResult(costs=(0.06323308175,), error_no=MeasureErrorNo.NO_ERROR, all_cost=3.009866237640381, timestamp=1679386618.4774506)       [(&#39;tile_f&#39;, [-1, 32, 2, 1]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 2, 1]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 0)],None,779475
-No: 10  GFLOPS: 0.00/25.12      result: Traceback (most recent call last):
-  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 142, in build
-    res = future.result()
-  File &quot;/usr/lib/python3.7/concurrent/futures/_base.py&quot;, line 435, in result
-    return self.__get_result()
-  File &quot;/usr/lib/python3.7/concurrent/futures/_base.py&quot;, line 384, in __get_result
-    raise self._exception
-  File &quot;/usr/lib/python3.7/concurrent/futures/thread.py&quot;, line 57, in run
-    result = self.fn(*self.args, **self.kwargs)
-  File &quot;/workspace/python/tvm/contrib/popen_pool.py&quot;, line 432, in &lt;lambda&gt;
-    worker = lambda *args: self._worker_run(*args)
-  File &quot;/workspace/python/tvm/contrib/popen_pool.py&quot;, line 401, in _worker_run
-    return proc.recv()
-  File &quot;/workspace/python/tvm/contrib/popen_pool.py&quot;, line 309, in recv
-    raise TimeoutError()
-TimeoutError
-
-        [(&#39;tile_f&#39;, [-1, 16, 4, 8]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 2, 2]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 0)],None,4687253
-No: 11  GFLOPS: 195.97/195.97   result: MeasureResult(costs=(0.0011813070823529412,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.7828633785247803, timestamp=1679386619.329636)       [(&#39;tile_f&#39;, [-1, 1, 64, 2]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 1, 1, 1]), (&#39;tile_rc&#39;, [-1, 32, 1]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,8730354
-No: 12  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 8, 1, 32]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 1]), (&#39;tile_rc&#39;, [-1, 4, 32]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 0)],None,3826868
+No: 11  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -1577,8 +1559,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 1, 4, 32]), (&#39;tile_y&#39;, [-1, 1, 1, 1]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 1, 64]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 1)],None,7516274
-No: 13  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 32, 1]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 16, 32]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 1)],None,6546582
+No: 12  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -1700,8 +1682,9 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 1, 4, 128]), (&#39;tile_y&#39;, [-1, 1, 1, 1]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 1, 128]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 1)],None,5787975
-No: 14  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 2, 16, 4]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 1, 128]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,10242887
+No: 13  GFLOPS: 523.22/1278.20  result: MeasureResult(costs=(0.00044245605217391306,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.484649658203125, timestamp=1679387632.097078)       [(&#39;tile_f&#39;, [-1, 1, 8, 1]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 4, 2]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 0)],None,819087
+No: 14  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -1823,8 +1806,9 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 2, 2]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 1]), (&#39;tile_rc&#39;, [-1, 64, 4]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 0)],None,2024286
-No: 15  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 2, 16, 4]), (&#39;tile_y&#39;, [-1, 1, 1, 1]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 64, 2]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 0)],None,4316527
+No: 15  GFLOPS: 35.46/1278.20   result: MeasureResult(costs=(0.006527616,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.4571244716644287, timestamp=1679387632.9939733)        [(&#39;tile_f&#39;, [-1, 2, 2, 2]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 64, 1]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 0)],None,216325
+No: 16  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -1946,8 +1930,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 2, 128, 1]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 4, 8]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 1)],None,7654070
-No: 16  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 32, 8, 2]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 128, 4]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 0)],None,3384124
+No: 17  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -2069,8 +2053,8 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 128, 1, 4]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 32, 16]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 0)],None,4399447
-No: 17  GFLOPS: 0.00/195.97     result: Traceback (most recent call last):
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 32, 1, 8]), (&#39;tile_y&#39;, [-1, 7, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 32, 16]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 0)],None,1108281
+No: 18  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
     func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
@@ -2192,10 +2176,253 @@ Traceback (most recent call last):
   File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
   File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
     raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
-tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 32, 4]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 1]), (&#39;tile_rc&#39;, [-1, 4, 8]), (&#39;tile_ry&#39;, [-1, 3, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 1)],None,5523452
-No: 18  GFLOPS: 109.98/195.97   result: MeasureResult(costs=(0.0021049593125000003,), error_no=MeasureErrorNo.NO_ERROR, all_cost=2.143249034881592, timestamp=1679386623.6207368)       [(&#39;tile_f&#39;, [-1, 1, 8, 1]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 128, 1]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 0)],None,3511667
-No: 19  GFLOPS: 116.11/195.97   result: MeasureResult(costs=(0.001993747470588235,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.6696715354919434, timestamp=1679386624.47567) [(&#39;tile_f&#39;, [-1, 16, 8, 1]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 2, 4]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 0)],None,2783031
-No: 20  GFLOPS: 129.42/195.97   result: MeasureResult(costs=(0.0017887947857142855,), error_no=MeasureErrorNo.NO_ERROR, all_cost=1.4931731224060059, timestamp=1679386625.3033702)      [(&#39;tile_f&#39;, [-1, 1, 16, 8]), (&#39;tile_y&#39;, [-1, 1, 1, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 2, 2]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 1)],None,6816638
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 2, 4, 2]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 128, 1]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 3, 1]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 0)],None,2350333
+No: 19  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
+  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
+    func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
+  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
+    func = build(s, args, target=target, runtime=runtime)
+  File &quot;/workspace/python/tvm/driver/build_module.py&quot;, line 227, in build
+    input_mod = lower(inputs, args, name=name, binds=binds)
+  File &quot;/workspace/python/tvm/driver/build_module.py&quot;, line 134, in lower
+    return ffi.lower_schedule(inp, args, name, binds, simple_mode)
+  File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
+  File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 276, in tvm._ffi._cy3.core.FuncCall
+  File &quot;tvm/_ffi/_cython/./base.pxi&quot;, line 181, in tvm._ffi._cy3.core.CHECK_CALL
+tvm._ffi.base.TVMError: Traceback (most recent call last):
+  24: TVMFuncCall
+        at ../src/runtime/c_runtime_api.cc:477
+  23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+        at ../include/tvm/runtime/packed_func.h:1217
+  22: Call
+        at ../include/tvm/runtime/packed_func.h:1213
+  21: operator()
+        at ../include/tvm/runtime/packed_func.h:1734
+  20: unpack_call&lt;tvm::IRModule, 5, tvm::&lt;lambda(tvm::te::Schedule, const tvm::runtime::Array&lt;tvm::runtime::ObjectRef&gt;&amp;, const tvm::runtime::String&amp;, const tvm::runtime::Map&lt;tvm::te::Tensor, tvm::tir::Buffer&gt;&amp;, bool)&gt; &gt;
+        at ../include/tvm/runtime/packed_func.h:1674
+  19: run&lt;&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  18: run&lt;tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  17: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  16: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  15: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  14: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1649
+  13: operator()
+        at ../src/driver/driver_api.cc:402
+  12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array&lt;tvm::runtime::ObjectRef, void&gt; const&amp;, std::__cxx11::basic_string&lt;char, std::char_traits&lt;char&gt;, std::allocator&lt;char&gt; &gt; const&amp;, std::unordered_map&lt;tvm::te::Tensor, tvm::tir::Buffer, std::hash&lt;tvm::te::Tensor&gt;, std::equal_to&lt;tvm::te::Tensor&gt;, std::allocator&lt;std::pair&lt;tvm::te::Tensor const, tvm::tir::Buffer&gt; &gt; &gt; const&amp;, tvm::GlobalVarSupply, bool)
+        at ../src/driver/driver_api.cc:388
+  11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array&lt;tvm::transform::Pass, void&gt;)
+        at ../src/driver/driver_api.cc:283
+  10: tvm::transform::Pass::operator()(tvm::IRModule) const
+        at ../src/ir/transform.cc:258
+  9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:274
+  8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:451
+  7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:274
+  6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/tir/ir/transform.cc:100
+  5: tvm::runtime::TypedPackedFunc&lt;tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)&gt;::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
+        at ../include/tvm/runtime/packed_func.h:1753
+  4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher&lt;tvm::tir::PrimFunc&gt;::run&lt;tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext&gt;(tvm::runtime::PackedFunc const&amp;, tvm::tir::PrimFunc&amp;&amp;, tvm::IRModule&amp;&amp;, tvm::transform::PassContext&amp;&amp;)
+        at ../include/tvm/runtime/packed_func.h:1697
+  3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()&lt;tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext&gt;(tvm::tir::PrimFunc&amp;&amp;, tvm::IRModule&amp;&amp;, tvm::transform::PassContext&amp;&amp;) const
+        at ../include/tvm/runtime/packed_func.h:1621
+  2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+        at ../include/tvm/runtime/packed_func.h:1217
+  1: Call
+        at ../include/tvm/runtime/packed_func.h:1213
+  0: operator()
+        at ../src/runtime/c_runtime_api.cc:534
+  File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
+  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
+    raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
+
+Traceback (most recent call last):
+  24: TVMFuncCall
+        at ../src/runtime/c_runtime_api.cc:477
+  23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+        at ../include/tvm/runtime/packed_func.h:1217
+  22: Call
+        at ../include/tvm/runtime/packed_func.h:1213
+  21: operator()
+        at ../include/tvm/runtime/packed_func.h:1734
+  20: unpack_call&lt;tvm::IRModule, 5, tvm::&lt;lambda(tvm::te::Schedule, const tvm::runtime::Array&lt;tvm::runtime::ObjectRef&gt;&amp;, const tvm::runtime::String&amp;, const tvm::runtime::Map&lt;tvm::te::Tensor, tvm::tir::Buffer&gt;&amp;, bool)&gt; &gt;
+        at ../include/tvm/runtime/packed_func.h:1674
+  19: run&lt;&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  18: run&lt;tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  17: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  16: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  15: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  14: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1649
+  13: operator()
+        at ../src/driver/driver_api.cc:402
+  12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array&lt;tvm::runtime::ObjectRef, void&gt; const&amp;, std::__cxx11::basic_string&lt;char, std::char_traits&lt;char&gt;, std::allocator&lt;char&gt; &gt; const&amp;, std::unordered_map&lt;tvm::te::Tensor, tvm::tir::Buffer, std::hash&lt;tvm::te::Tensor&gt;, std::equal_to&lt;tvm::te::Tensor&gt;, std::allocator&lt;std::pair&lt;tvm::te::Tensor const, tvm::tir::Buffer&gt; &gt; &gt; const&amp;, tvm::GlobalVarSupply, bool)
+        at ../src/driver/driver_api.cc:388
+  11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array&lt;tvm::transform::Pass, void&gt;)
+        at ../src/driver/driver_api.cc:283
+  10: tvm::transform::Pass::operator()(tvm::IRModule) const
+        at ../src/ir/transform.cc:258
+  9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:274
+  8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:451
+  7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:274
+  6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/tir/ir/transform.cc:100
+  5: tvm::runtime::TypedPackedFunc&lt;tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)&gt;::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
+        at ../include/tvm/runtime/packed_func.h:1753
+  4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher&lt;tvm::tir::PrimFunc&gt;::run&lt;tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext&gt;(tvm::runtime::PackedFunc const&amp;, tvm::tir::PrimFunc&amp;&amp;, tvm::IRModule&amp;&amp;, tvm::transform::PassContext&amp;&amp;)
+        at ../include/tvm/runtime/packed_func.h:1697
+  3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()&lt;tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext&gt;(tvm::tir::PrimFunc&amp;&amp;, tvm::IRModule&amp;&amp;, tvm::transform::PassContext&amp;&amp;) const
+        at ../include/tvm/runtime/packed_func.h:1621
+  2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+        at ../include/tvm/runtime/packed_func.h:1217
+  1: Call
+        at ../include/tvm/runtime/packed_func.h:1213
+  0: operator()
+        at ../src/runtime/c_runtime_api.cc:534
+  File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
+  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
+    raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 2, 1]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 7, 1]), (&#39;tile_rc&#39;, [-1, 8, 32]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 0), (&#39;unroll_explicit&#39;, 0)],None,1702372
+No: 20  GFLOPS: 0.00/1278.20    result: Traceback (most recent call last):
+  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 592, in __call__
+    func, arg_info = _build_func_common(measure_input, self.runtime, **kwargs)
+  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 544, in _build_func_common
+    func = build(s, args, target=target, runtime=runtime)
+  File &quot;/workspace/python/tvm/driver/build_module.py&quot;, line 227, in build
+    input_mod = lower(inputs, args, name=name, binds=binds)
+  File &quot;/workspace/python/tvm/driver/build_module.py&quot;, line 134, in lower
+    return ffi.lower_schedule(inp, args, name, binds, simple_mode)
+  File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 331, in tvm._ffi._cy3.core.PackedFuncBase.__call__
+  File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 276, in tvm._ffi._cy3.core.FuncCall
+  File &quot;tvm/_ffi/_cython/./base.pxi&quot;, line 181, in tvm._ffi._cy3.core.CHECK_CALL
+tvm._ffi.base.TVMError: Traceback (most recent call last):
+  24: TVMFuncCall
+        at ../src/runtime/c_runtime_api.cc:477
+  23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+        at ../include/tvm/runtime/packed_func.h:1217
+  22: Call
+        at ../include/tvm/runtime/packed_func.h:1213
+  21: operator()
+        at ../include/tvm/runtime/packed_func.h:1734
+  20: unpack_call&lt;tvm::IRModule, 5, tvm::&lt;lambda(tvm::te::Schedule, const tvm::runtime::Array&lt;tvm::runtime::ObjectRef&gt;&amp;, const tvm::runtime::String&amp;, const tvm::runtime::Map&lt;tvm::te::Tensor, tvm::tir::Buffer&gt;&amp;, bool)&gt; &gt;
+        at ../include/tvm/runtime/packed_func.h:1674
+  19: run&lt;&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  18: run&lt;tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  17: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  16: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  15: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  14: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1649
+  13: operator()
+        at ../src/driver/driver_api.cc:402
+  12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array&lt;tvm::runtime::ObjectRef, void&gt; const&amp;, std::__cxx11::basic_string&lt;char, std::char_traits&lt;char&gt;, std::allocator&lt;char&gt; &gt; const&amp;, std::unordered_map&lt;tvm::te::Tensor, tvm::tir::Buffer, std::hash&lt;tvm::te::Tensor&gt;, std::equal_to&lt;tvm::te::Tensor&gt;, std::allocator&lt;std::pair&lt;tvm::te::Tensor const, tvm::tir::Buffer&gt; &gt; &gt; const&amp;, tvm::GlobalVarSupply, bool)
+        at ../src/driver/driver_api.cc:388
+  11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array&lt;tvm::transform::Pass, void&gt;)
+        at ../src/driver/driver_api.cc:283
+  10: tvm::transform::Pass::operator()(tvm::IRModule) const
+        at ../src/ir/transform.cc:258
+  9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:274
+  8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:451
+  7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:274
+  6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/tir/ir/transform.cc:100
+  5: tvm::runtime::TypedPackedFunc&lt;tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)&gt;::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
+        at ../include/tvm/runtime/packed_func.h:1753
+  4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher&lt;tvm::tir::PrimFunc&gt;::run&lt;tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext&gt;(tvm::runtime::PackedFunc const&amp;, tvm::tir::PrimFunc&amp;&amp;, tvm::IRModule&amp;&amp;, tvm::transform::PassContext&amp;&amp;)
+        at ../include/tvm/runtime/packed_func.h:1697
+  3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()&lt;tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext&gt;(tvm::tir::PrimFunc&amp;&amp;, tvm::IRModule&amp;&amp;, tvm::transform::PassContext&amp;&amp;) const
+        at ../include/tvm/runtime/packed_func.h:1621
+  2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+        at ../include/tvm/runtime/packed_func.h:1217
+  1: Call
+        at ../include/tvm/runtime/packed_func.h:1213
+  0: operator()
+        at ../src/runtime/c_runtime_api.cc:534
+  File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
+  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
+    raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel
+
+Traceback (most recent call last):
+  24: TVMFuncCall
+        at ../src/runtime/c_runtime_api.cc:477
+  23: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+        at ../include/tvm/runtime/packed_func.h:1217
+  22: Call
+        at ../include/tvm/runtime/packed_func.h:1213
+  21: operator()
+        at ../include/tvm/runtime/packed_func.h:1734
+  20: unpack_call&lt;tvm::IRModule, 5, tvm::&lt;lambda(tvm::te::Schedule, const tvm::runtime::Array&lt;tvm::runtime::ObjectRef&gt;&amp;, const tvm::runtime::String&amp;, const tvm::runtime::Map&lt;tvm::te::Tensor, tvm::tir::Buffer&gt;&amp;, bool)&gt; &gt;
+        at ../include/tvm/runtime/packed_func.h:1674
+  19: run&lt;&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  18: run&lt;tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  17: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  16: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  15: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1634
+  14: run&lt;tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_, tvm::runtime::TVMMovableArgValueWithContext_&gt;
+        at ../include/tvm/runtime/packed_func.h:1649
+  13: operator()
+        at ../src/driver/driver_api.cc:402
+  12: tvm::LowerSchedule(tvm::te::Schedule, tvm::runtime::Array&lt;tvm::runtime::ObjectRef, void&gt; const&amp;, std::__cxx11::basic_string&lt;char, std::char_traits&lt;char&gt;, std::allocator&lt;char&gt; &gt; const&amp;, std::unordered_map&lt;tvm::te::Tensor, tvm::tir::Buffer, std::hash&lt;tvm::te::Tensor&gt;, std::equal_to&lt;tvm::te::Tensor&gt;, std::allocator&lt;std::pair&lt;tvm::te::Tensor const, tvm::tir::Buffer&gt; &gt; &gt; const&amp;, tvm::GlobalVarSupply, bool)
+        at ../src/driver/driver_api.cc:388
+  11: tvm::LowerWithPassList(tvm::IRModule, tvm::runtime::Array&lt;tvm::transform::Pass, void&gt;)
+        at ../src/driver/driver_api.cc:283
+  10: tvm::transform::Pass::operator()(tvm::IRModule) const
+        at ../src/ir/transform.cc:258
+  9: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:274
+  8: tvm::transform::SequentialNode::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:451
+  7: tvm::transform::Pass::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/ir/transform.cc:274
+  6: tvm::tir::transform::PrimFuncPassNode::operator()(tvm::IRModule, tvm::transform::PassContext const&amp;) const
+        at ../src/tir/ir/transform.cc:100
+  5: tvm::runtime::TypedPackedFunc&lt;tvm::tir::PrimFunc (tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext)&gt;::operator()(tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext) const
+        at ../include/tvm/runtime/packed_func.h:1753
+  4: tvm::tir::PrimFunc tvm::runtime::detail::typed_packed_call_dispatcher&lt;tvm::tir::PrimFunc&gt;::run&lt;tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext&gt;(tvm::runtime::PackedFunc const&amp;, tvm::tir::PrimFunc&amp;&amp;, tvm::IRModule&amp;&amp;, tvm::transform::PassContext&amp;&amp;)
+        at ../include/tvm/runtime/packed_func.h:1697
+  3: tvm::runtime::TVMRetValue tvm::runtime::PackedFunc::operator()&lt;tvm::tir::PrimFunc, tvm::IRModule, tvm::transform::PassContext&gt;(tvm::tir::PrimFunc&amp;&amp;, tvm::IRModule&amp;&amp;, tvm::transform::PassContext&amp;&amp;) const
+        at ../include/tvm/runtime/packed_func.h:1621
+  2: tvm::runtime::PackedFuncObj::CallPacked(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const
+        at ../include/tvm/runtime/packed_func.h:1217
+  1: Call
+        at ../include/tvm/runtime/packed_func.h:1213
+  0: operator()
+        at ../src/runtime/c_runtime_api.cc:534
+  File &quot;tvm/_ffi/_cython/./packed_func.pxi&quot;, line 56, in tvm._ffi._cy3.core.tvm_callback
+  File &quot;/workspace/python/tvm/autotvm/measure/measure_methods.py&quot;, line 875, in verify_pass
+    raise InstantiationError(&quot;Skipped because of invalid gpu kernel&quot;)
+tvm.autotvm.task.space.InstantiationError: Skipped because of invalid gpu kernel        [(&#39;tile_f&#39;, [-1, 4, 32, 4]), (&#39;tile_y&#39;, [-1, 1, 1, 1]), (&#39;tile_x&#39;, [-1, 7, 1, 1]), (&#39;tile_rc&#39;, [-1, 1, 32]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,8853812
 </pre></div>
 </div>
 <p>Finally we can inspect the best config from log file, check correctness,
@@ -2234,9 +2461,9 @@ 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:
-[(&#39;tile_f&#39;, [-1, 1, 64, 2]), (&#39;tile_y&#39;, [-1, 1, 1, 7]), (&#39;tile_x&#39;, [-1, 1, 1, 1]), (&#39;tile_rc&#39;, [-1, 32, 1]), (&#39;tile_ry&#39;, [-1, 1, 1]), (&#39;tile_rx&#39;, [-1, 1, 1]), (&#39;auto_unroll_max_step&#39;, 1500), (&#39;unroll_explicit&#39;, 1)],None,8730354
+[(&#39;tile_f&#39;, [-1, 2, 8, 1]), (&#39;tile_y&#39;, [-1, 1, 7, 1]), (&#39;tile_x&#39;, [-1, 1, 1, 7]), (&#39;tile_rc&#39;, [-1, 4, 2]), (&#39;tile_ry&#39;, [-1, 1, 3]), (&#39;tile_rx&#39;, [-1, 1, 3]), (&#39;auto_unroll_max_step&#39;, 512), (&#39;unroll_explicit&#39;, 1)],None,8563748
 Finish loading 20 records
-Time cost of this operator: 0.001574
+Time cost of this operator: 0.000404
 </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 ed2dccf13e..7f5db4da50 100644
--- a/docs/how_to/work_with_microtvm/micro_autotune.html
+++ b/docs/how_to/work_with_microtvm/micro_autotune.html
@@ -649,10 +649,10 @@ the tuned operator.</p>
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>########## Build without Autotuning ##########
 Node Name                                     Ops                                           Time(us)  Time(%)  Shape              Inputs  Outputs  Measurements(us)
 ---------                                     ---                                           --------  -------  -----              ------  -------  ----------------
-tvmgen_default_fused_nn_contrib_conv2d_NCHWc  tvmgen_default_fused_nn_contrib_conv2d_NCHWc  315.7     98.728   (1, 2, 10, 10, 3)  2       1        [315.7]
-tvmgen_default_fused_layout_transform_1       tvmgen_default_fused_layout_transform_1       3.101     0.97     (1, 6, 10, 10)     1       1        [3.101]
-tvmgen_default_fused_layout_transform         tvmgen_default_fused_layout_transform         0.966     0.302    (1, 1, 10, 10, 3)  1       1        [0.966]
-Total_time                                    -                                             319.766   -        -                  -       -        -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc  tvmgen_default_fused_nn_contrib_conv2d_NCHWc  314.5     98.74    (1, 2, 10, 10, 3)  2       1        [314.5]
+tvmgen_default_fused_layout_transform_1       tvmgen_default_fused_layout_transform_1       3.048     0.957    (1, 6, 10, 10)     1       1        [3.048]
+tvmgen_default_fused_layout_transform         tvmgen_default_fused_layout_transform         0.966     0.303    (1, 1, 10, 10, 3)  1       1        [0.966]
+Total_time                                    -                                             318.514   -        -                  -       -        -
 </pre></div>
 </div>
 </div>
@@ -704,13 +704,13 @@ Total_time                                    -
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>########## Build with Autotuning ##########
 Node Name                                     Ops                                           Time(us)  Time(%)  Shape              Inputs  Outputs  Measurements(us)
 ---------                                     ---                                           --------  -------  -----              ------  -------  ----------------
-tvmgen_default_fused_nn_contrib_conv2d_NCHWc  tvmgen_default_fused_nn_contrib_conv2d_NCHWc  103.5     97.496   (1, 6, 10, 10, 1)  2       1        [103.5]
-tvmgen_default_fused_layout_transform_1       tvmgen_default_fused_layout_transform_1       1.798     1.693    (1, 6, 10, 10)     1       1        [1.798]
-tvmgen_default_fused_layout_transform         tvmgen_default_fused_layout_transform         0.86      0.81     (1, 3, 10, 10, 1)  1       1        [0.86]
-Total_time                                    -                                             106.158   -        -                  -       -        -
+tvmgen_default_fused_nn_contrib_conv2d_NCHWc  tvmgen_default_fused_nn_contrib_conv2d_NCHWc  105.0     97.564   (1, 6, 10, 10, 1)  2       1        [105.0]
+tvmgen_default_fused_layout_transform_1       tvmgen_default_fused_layout_transform_1       1.78      1.654    (1, 6, 10, 10)     1       1        [1.78]
+tvmgen_default_fused_layout_transform         tvmgen_default_fused_layout_transform         0.842     0.782    (1, 3, 10, 10, 1)  1       1        [0.842]
+Total_time                                    -                                             107.622   -        -                  -       -        -
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  21.994 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  22.179 seconds)</p>
 <div class="sphx-glr-footer sphx-glr-footer-example docutils container" id="sphx-glr-download-how-to-work-with-microtvm-micro-autotune-py">
 <div class="sphx-glr-download sphx-glr-download-python docutils container">
 <p><a class="reference download internal" download="" href="../../_downloads/9ccca8fd489a1486ac71b55a55c320c5/micro_autotune.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">micro_autotune.py</span></code></a></p>
diff --git a/docs/how_to/work_with_microtvm/micro_pytorch.html b/docs/how_to/work_with_microtvm/micro_pytorch.html
index eeaae63eb0..0a97c93443 100644
--- a/docs/how_to/work_with_microtvm/micro_pytorch.html
+++ b/docs/how_to/work_with_microtvm/micro_pytorch.html
@@ -460,7 +460,8 @@ download a cat image and preprocess it to use as the model input.</p>
 Downloading: &quot;https://download.pytorch.org/models/quantized/mobilenet_v2_qnnpack_37f702c5.pth&quot; to /workspace/.cache/torch/hub/checkpoints/mobilenet_v2_qnnpack_37f702c5.pth
 
   0%|          | 0.00/3.42M [00:00&lt;?, ?B/s]
-100%|##########| 3.42M/3.42M [00:00&lt;00:00, 53.9MB/s]
+ 61%|######    | 2.09M/3.42M [00:00&lt;00:00, 11.9MB/s]
+100%|##########| 3.42M/3.42M [00:00&lt;00:00, 18.6MB/s]
 /workspace/python/tvm/relay/frontend/pytorch_utils.py:47: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
   return LooseVersion(torch_ver) &gt; 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.
@@ -586,7 +587,7 @@ via the host <cite>main.cc`</cite> or if a Zephyr emulated board is selected as
 Torch top-1 id: 282, class name: tiger cat
 </pre></div>
 </div>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  18.461 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 1 minutes  19.009 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 fff03981bb..d2d5350056 100644
--- a/docs/how_to/work_with_microtvm/micro_train.html
+++ b/docs/how_to/work_with_microtvm/micro_train.html
@@ -529,7 +529,7 @@ take about <strong>2 minutes</strong> to download the Stanford Cars, while COCO
 <a href="https://docs.python.org/3/library/shutil.html#shutil.move" title="shutil.move" class="sphx-glr-backref-module-shutil sphx-glr-backref-type-py-function"><span class="n">shutil</span><span class="o">.</span><span class="n">move</span></a><span class="p">(</span><span class="sa">f</span><span class="s2">&quot;</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>&#39;/tmp/tmp_rfyuq_w/images/random&#39;
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>&#39;/tmp/tmpn9bjfr1u/images/random&#39;
 </pre></div>
 </div>
 </div>
@@ -589,8 +589,8 @@ objects to other stuff? We can display some examples from our datasets using <co
     <span class="n">plt</span><span class="o">.</span><span class="n">axis</span><span class="p">(</span><span class="s2">&quot;off&quot;</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], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [0.0, 1.0], [1.0, 0.0], [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/tmp_rfyuq_w/images/target contains 8144 images
-/tmp/tmp_rfyuq_w/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], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [1.0, 0.0], [0.0, 1.0], [0.0, 1.0]" class = "sphx-glr-single-img"/><div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>/tmp/tmpn9bjfr1u/images/target contains 8144 images
+/tmp/tmpn9bjfr1u/images/random contains 5000 images
 </pre></div>
 </div>
 </div>
@@ -702,13 +702,13 @@ the time on our validation set).</p>
 </pre></div>
 </div>
 <div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Epoch 1/3
-328/328 - 47s - loss: 0.2213 - accuracy: 0.9245 - val_loss: 0.1149 - val_accuracy: 0.9588 - 47s/epoch - 145ms/step
+328/328 - 47s - loss: 0.2282 - accuracy: 0.9221 - val_loss: 0.1165 - val_accuracy: 0.9592 - 47s/epoch - 144ms/step
 Epoch 2/3
-328/328 - 43s - loss: 0.1003 - accuracy: 0.9627 - val_loss: 0.0947 - val_accuracy: 0.9649 - 43s/epoch - 132ms/step
+328/328 - 44s - loss: 0.0989 - accuracy: 0.9645 - val_loss: 0.1235 - val_accuracy: 0.9600 - 44s/epoch - 133ms/step
 Epoch 3/3
-328/328 - 43s - loss: 0.0658 - accuracy: 0.9743 - val_loss: 0.0970 - val_accuracy: 0.9656 - 43s/epoch - 131ms/step
+328/328 - 43s - loss: 0.0618 - accuracy: 0.9753 - val_loss: 0.1142 - val_accuracy: 0.9683 - 43s/epoch - 132ms/step
 
-&lt;keras.callbacks.History object at 0x7fa816b52a10&gt;
+&lt;keras.callbacks.History object at 0x7f68b6ad09d0&gt;
 </pre></div>
 </div>
 </div>
@@ -972,7 +972,7 @@ as intended.</p>
 <p>From here, we could modify the model to read live images from the camera - we have another
 Arduino tutorial for how to do that <a class="reference external" href="https://github.com/guberti/tvm-arduino-demos/tree/master/examples/person_detection">on GitHub</a>. Alternatively, we could also
 <a class="reference external" href="https://tvm.apache.org/docs/how_to/work_with_microtvm/micro_autotune.html">use TVM’s autotuning capabilities</a> to dramatically improve the model’s performance.</p>
-<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 4 minutes  33.636 seconds)</p>
+<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 4 minutes  34.489 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 d5ef668a6a..279477b19f 100644
--- a/docs/how_to/work_with_microtvm/sg_execution_times.html
+++ b/docs/how_to/work_with_microtvm/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-work-with-microtvm-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>07:39.956</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
+<p><strong>07:41.083</strong> total execution time for <strong>how_to_work_with_microtvm</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 82%" />
@@ -354,27 +354,27 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="micro_train.html#sphx-glr-how-to-work-with-microtvm-micro-train-py"><span class="std std-ref">5. Training Vision Models for microTVM on Arduino</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_train.py</span></code>)</p></td>
-<td><p>04:33.636</p></td>
+<td><p>04:34.489</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="micro_autotune.html#sphx-glr-how-to-work-with-microtvm-micro-autotune-py"><span class="std std-ref">6. Model Tuning with microTVM</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_autotune.py</span></code>)</p></td>
-<td><p>01:21.994</p></td>
+<td><p>01:22.179</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="micro_pytorch.html#sphx-glr-how-to-work-with-microtvm-micro-pytorch-py"><span class="std std-ref">4. microTVM PyTorch Tutorial</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_pytorch.py</span></code>)</p></td>
-<td><p>01:18.461</p></td>
+<td><p>01:19.009</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="micro_aot.html#sphx-glr-how-to-work-with-microtvm-micro-aot-py"><span class="std std-ref">3. microTVM Ahead-of-Time (AOT) Compilation</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_aot.py</span></code>)</p></td>
-<td><p>00:10.341</p></td>
+<td><p>00:10.186</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="micro_custom_ide.html#sphx-glr-how-to-work-with-microtvm-micro-custom-ide-py"><span class="std std-ref">9. Bring microTVM to your own development environment</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_custom_ide.py</span></code>)</p></td>
-<td><p>00:08.017</p></td>
+<td><p>00:07.951</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="micro_tflite.html#sphx-glr-how-to-work-with-microtvm-micro-tflite-py"><span class="std std-ref">2. microTVM TFLite Tutorial</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_tflite.py</span></code>)</p></td>
-<td><p>00:07.507</p></td>
+<td><p>00:07.270</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="micro_ethosu.html#sphx-glr-how-to-work-with-microtvm-micro-ethosu-py"><span class="std std-ref">7. Running TVM on bare metal Arm(R) Cortex(R)-M55 CPU and Ethos(TM)-U55 NPU with CMSIS-NN</span></a> (<code class="docutils literal notranslate"><span class="pre">micro_ethosu.py</span></code>)</p></td>
diff --git a/docs/how_to/work_with_relay/sg_execution_times.html b/docs/how_to/work_with_relay/sg_execution_times.html
index 986f5cead1..c21f08bc54 100644
--- a/docs/how_to/work_with_relay/sg_execution_times.html
+++ b/docs/how_to/work_with_relay/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-work-with-relay-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:46.763</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
+<p><strong>00:46.557</strong> total execution time for <strong>how_to_work_with_relay</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 84%" />
@@ -354,15 +354,15 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="using_pipeline_executor.html#sphx-glr-how-to-work-with-relay-using-pipeline-executor-py"><span class="std std-ref">Using Pipeline Executor in Relay</span></a> (<code class="docutils literal notranslate"><span class="pre">using_pipeline_executor.py</span></code>)</p></td>
-<td><p>00:34.578</p></td>
+<td><p>00:34.426</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>00:10.588</p></td>
+<td><p>00:10.330</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>00:01.591</p></td>
+<td><p>00:01.795</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 3858d1e8fd..935e5060e8 100644
--- a/docs/how_to/work_with_schedules/intrin_math.html
+++ b/docs/how_to/work_with_schedules/intrin_math.html
@@ -540,7 +540,7 @@ The following example customizes CUDA lowering rule for <code class="code docuti
 <a href="../../reference/api/python/ir.html#tvm.ir.register_intrin_lowering" title="tvm.ir.register_intrin_lowering" class="sphx-glr-backref-module-tvm-ir sphx-glr-backref-type-py-function"><span class="n">register_intrin_lowering</span></a><span class="p">(</span><span class="s2">&quot;tir.exp&quot;</span><span class="p">,</span> <span class="n">target</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</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>&lt;function my_cuda_math_rule at 0x7fa6bc15f290&gt;
+<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>&lt;function my_cuda_math_rule at 0x7f67574853b0&gt;
 </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 690d7e5b73..8bafe407bc 100644
--- a/docs/how_to/work_with_schedules/sg_execution_times.html
+++ b/docs/how_to/work_with_schedules/sg_execution_times.html
@@ -345,7 +345,7 @@
             
   <div class="section" id="computation-times">
 <span id="sphx-glr-how-to-work-with-schedules-sg-execution-times"></span><h1>Computation times<a class="headerlink" href="#computation-times" title="Permalink to this headline">¶</a></h1>
-<p><strong>00:06.608</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
+<p><strong>00:08.138</strong> total execution time for <strong>how_to_work_with_schedules</strong> files:</p>
 <table class="docutils align-default">
 <colgroup>
 <col style="width: 83%" />
@@ -354,27 +354,27 @@
 </colgroup>
 <tbody>
 <tr class="row-odd"><td><p><a class="reference internal" href="intrin_math.html#sphx-glr-how-to-work-with-schedules-intrin-math-py"><span class="std std-ref">Intrinsics and Math Functions</span></a> (<code class="docutils literal notranslate"><span class="pre">intrin_math.py</span></code>)</p></td>
-<td><p>00:04.026</p></td>
+<td><p>00:05.598</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>00:01.177</p></td>
+<td><p>00:01.130</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>00:00.582</p></td>
+<td><p>00:00.588</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>00:00.565</p></td>
+<td><p>00:00.566</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>00:00.118</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-even"><td><p><a class="reference internal" href="tedd.html#sphx-glr-how-to-work-with-schedules-tedd-py"><span class="std std-ref">Use Tensor Expression Debug Display (TEDD) for Visualization</span></a> (<code class="docutils literal notranslate"><span class="pre">tedd.py</span></code>)</p></td>
-<td><p>00:00.060</p></td>
+<td><p>00:00.059</p></td>
 <td><p>0.0 MB</p></td>
 </tr>
 <tr class="row-odd"><td><p><a class="reference internal" href="schedule_primitives.html#sphx-glr-how-to-work-with-schedules-schedule-primitives-py"><span class="std std-ref">Schedule Primitives in TVM</span></a> (<code class="docutils literal notranslate"><span class="pre">schedule_primitives.py</span></code>)</p></td>
diff --git a/docs/install/nnpack.html b/docs/install/nnpack.html
index e285b427a5..1515c4b747 100644
--- a/docs/install/nnpack.html
+++ b/docs/install/nnpack.html
@@ -234,17 +234,7 @@
               <p class="caption" role="heading"><span class="caption-text">Getting Started</span></p>
 <ul class="current">
 <li class="toctree-l1 current"><a class="reference internal" href="index.html">Installing TVM</a><ul class="current">
-<li class="toctree-l2 current"><a class="reference internal" href="from_source.html">Install from Source</a><ul class="current">
-<li class="toctree-l3"><a class="reference internal" href="from_source.html#developers-get-source-from-github">Developers: Get Source from Github</a></li>
-<li class="toctree-l3"><a class="reference internal" href="from_source.html#build-the-shared-library">Build the Shared Library</a></li>
-<li class="toctree-l3"><a class="reference internal" href="from_source.html#python-package-installation">Python Package Installation</a></li>
-<li class="toctree-l3 current"><a class="reference internal" href="from_source.html#install-contrib-libraries">Install Contrib Libraries</a><ul class="current">
-<li class="toctree-l4 current"><a class="current reference internal" href="#">NNPACK Contrib Installation</a></li>
-</ul>
-</li>
-<li class="toctree-l3"><a class="reference internal" href="from_source.html#enable-c-tests">Enable C++ Tests</a></li>
-</ul>
-</li>
+<li class="toctree-l2"><a class="reference internal" href="from_source.html">Install from Source</a></li>
 <li class="toctree-l2"><a class="reference internal" href="docker.html">Docker Images</a></li>
 <li class="toctree-l2 current"><a class="current reference internal" href="#">NNPACK Contrib Installation</a><ul>
 <li class="toctree-l3"><a class="reference internal" href="#conditions">Conditions</a></li>
diff --git a/docs/reference/api/python/auto_scheduler.html b/docs/reference/api/python/auto_scheduler.html
index 3ef6e2365e..25e5a9511c 100644
--- a/docs/reference/api/python/auto_scheduler.html
+++ b/docs/reference/api/python/auto_scheduler.html
@@ -1621,7 +1621,7 @@ history states as starting point to perform Evolutionary Search).</p></li>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="tvm.auto_scheduler.SketchPolicy">
-<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">tvm.auto_scheduler.</span></span><span class="sig-name descname"><span class="pre">SketchPolicy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">program_cost_model</span></span><span class="o"><span class="pre">=</span></span><span class="defau [...]
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">tvm.auto_scheduler.</span></span><span class="sig-name descname"><span class="pre">SketchPolicy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">program_cost_model</span></span><span class="o"><span class="pre">=</span></span><span class="defau [...]
 <dd><p>The search policy that searches in a hierarchical search space defined by sketches.
 The policy randomly samples programs from the space defined by sketches and use evolutionary
 search to fine-tune them.</p>
@@ -1905,7 +1905,7 @@ Candidates:
 
 <dl class="py function">
 <dt class="sig sig-object py" id="tvm.auto_scheduler.auto_schedule">
-<span class="sig-prename descclassname"><span class="pre">tvm.auto_scheduler.</span></span><span class="sig-name descname"><span class="pre">auto_schedule</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_policy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em clas [...]
+<span class="sig-prename descclassname"><span class="pre">tvm.auto_scheduler.</span></span><span class="sig-name descname"><span class="pre">auto_schedule</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">task</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">search_policy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em clas [...]
 <dd><p>THIS API IS DEPRECATED.</p>
 <p>Run auto scheduling search for a task.</p>
 <dl class="field-list simple">
diff --git a/docs/reference/api/typedoc/classes/bytestreamreader.html b/docs/reference/api/typedoc/classes/bytestreamreader.html
index 8e1c53f272..4368df9bff 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/7f6da0905/web/src/rpc_server.ts#L45">rpc_server.ts:45</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/rpc_server.ts#L45">rpc_server.ts:45</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-parameters-title">Parameters</h4>
@@ -141,7 +141,7 @@
 					<div class="tsd-signature tsd-kind-icon">bytes<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Uint8Array</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/rpc_server.ts#L45">rpc_server.ts:45</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/rpc_server.ts#L45">rpc_server.ts:45</a></li>
 						</ul>
 					</aside>
 				</section>
@@ -151,7 +151,7 @@
 					<div class="tsd-signature tsd-kind-icon">offset<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span><span class="tsd-signature-symbol"> = 0</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/rpc_server.ts#L44">rpc_server.ts:44</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/rpc_server.ts#L44">rpc_server.ts:44</a></li>
 						</ul>
 					</aside>
 				</section>
@@ -168,7 +168,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/rpc_server.ts#L65">rpc_server.ts:65</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/rpc_server.ts#L65">rpc_server.ts:65</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">Uint8Array</span></h4>
@@ -185,7 +185,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/rpc_server.ts#L51">rpc_server.ts:51</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/rpc_server.ts#L51">rpc_server.ts:51</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
@@ -202,7 +202,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/rpc_server.ts#L59">rpc_server.ts:59</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/rpc_server.ts#L59">rpc_server.ts:59</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
diff --git a/docs/reference/api/typedoc/classes/cachedcallstack.html b/docs/reference/api/typedoc/classes/cachedcallstack.html
index e8b7cb8450..7f0b1c10a2 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/7f6da0905/web/src/memory.ts#L223">memory.ts:223</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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">&lt;</span><a href="../interfaces/disposable.html" class="tsd-signature-type">Disposable</a><span class="tsd-signature-symbol">&gt;</span><span class="tsd-signature-symbol"> = []</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/memory.ts#L208">memory.ts:208</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L312">memory.ts:312</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L284">memory.ts:284</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L388">memory.ts:388</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L376">memory.ts:376</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L267">memory.ts:267</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L243">memory.ts:243</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L321">memory.ts:321</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L252">memory.ts:252</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L359">memory.ts:359</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L342">memory.ts:342</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L350">memory.ts:350</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L326">memory.ts:326</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L363">memory.ts:363</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L346">memory.ts:346</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L334">memory.ts:334</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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 8e88a932ee..f8dc4ada45 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/7f6da0905/web/src/runtime.ts#L359">runtime.ts:359</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L359">runtime.ts:359</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-parameters-title">Parameters</h4>
@@ -147,7 +147,7 @@
 					<div class="tsd-signature tsd-kind-icon">bits<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L357">runtime.ts:357</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L357">runtime.ts:357</a></li>
 						</ul>
 					</aside>
 					<div class="tsd-comment tsd-typography">
@@ -162,7 +162,7 @@
 					<div class="tsd-signature tsd-kind-icon">code<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L355">runtime.ts:355</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L355">runtime.ts:355</a></li>
 						</ul>
 					</aside>
 					<div class="tsd-comment tsd-typography">
@@ -177,7 +177,7 @@
 					<div class="tsd-signature tsd-kind-icon">lanes<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L359">runtime.ts:359</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L359">runtime.ts:359</a></li>
 						</ul>
 					</aside>
 					<div class="tsd-comment tsd-typography">
@@ -199,7 +199,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L376">runtime.ts:376</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L376">runtime.ts:376</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
@@ -216,7 +216,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L367">runtime.ts:367</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L367">runtime.ts:367</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">string</span></h4>
diff --git a/docs/reference/api/typedoc/classes/dldevice.html b/docs/reference/api/typedoc/classes/dldevice.html
index 75e990f7aa..14b49500d7 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/7f6da0905/web/src/runtime.ts#L299">runtime.ts:299</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L299">runtime.ts:299</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-parameters-title">Parameters</h4>
@@ -146,7 +146,7 @@
 					<div class="tsd-signature tsd-kind-icon">device<wbr>Id<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L297">runtime.ts:297</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L297">runtime.ts:297</a></li>
 						</ul>
 					</aside>
 					<div class="tsd-comment tsd-typography">
@@ -161,7 +161,7 @@
 					<div class="tsd-signature tsd-kind-icon">device<wbr>Type<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">number</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L295">runtime.ts:295</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L295">runtime.ts:295</a></li>
 						</ul>
 					</aside>
 					<div class="tsd-comment tsd-typography">
@@ -183,7 +183,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L320">runtime.ts:320</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L320">runtime.ts:320</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -205,7 +205,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L327">runtime.ts:327</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L327">runtime.ts:327</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">string</span></h4>
diff --git a/docs/reference/api/typedoc/classes/environment.html b/docs/reference/api/typedoc/classes/environment.html
index f79240d200..6ad6b37465 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/7f6da0905/web/src/environment.ts#L86">environment.ts:86</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/environment.ts#L70">environment.ts:70</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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"> =&gt; </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/7f6da0905/web/src/environment.ts#L69">environment.ts:69</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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">&lt;</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">&gt;</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/7f6da0905/web/src/environment.ts#L78">environment.ts:78</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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">&lt;</span><span class="tsd-signature-type">number</span><span class="tsd-signature-symbol">&gt;</span><span class="tsd-signature-symbol"> = []</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/environment.ts#L84">environment.ts:84</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/environment.ts#L105">environment.ts:105</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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 de239c9d3a..0d18217f5b 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/7f6da0905/web/src/runtime.ts#L50">runtime.ts:50</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L50">runtime.ts:50</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-parameters-title">Parameters</h4>
@@ -156,7 +156,7 @@
 					<div class="tsd-signature tsd-kind-icon">exports<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Record</span><span class="tsd-signature-symbol">&lt;</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">&gt;</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L47">runtime.ts:47</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L47">runtime.ts:47</a></li>
 						</ul>
 					</aside>
 				</section>
@@ -166,7 +166,7 @@
 					<div class="tsd-signature tsd-kind-icon">memory<span class="tsd-signature-symbol">:</span> <a href="memory.html" class="tsd-signature-type">Memory</a></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L46">runtime.ts:46</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L46">runtime.ts:46</a></li>
 						</ul>
 					</aside>
 				</section>
@@ -176,7 +176,7 @@
 					<div class="tsd-signature tsd-kind-icon">wasm32<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">boolean</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L45">runtime.ts:45</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L45">runtime.ts:45</a></li>
 						</ul>
 					</aside>
 				</section>
@@ -186,7 +186,7 @@
 					<div class="tsd-signature tsd-kind-icon">webGPUContext<span class="tsd-signature-symbol">:</span> <a href="webgpucontext.html" class="tsd-signature-type">WebGPUContext</a></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L48">runtime.ts:48</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L48">runtime.ts:48</a></li>
 						</ul>
 					</aside>
 				</section>
@@ -203,7 +203,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L77">runtime.ts:77</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L77">runtime.ts:77</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-parameters-title">Parameters</h4>
@@ -226,7 +226,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L67">runtime.ts:67</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L67">runtime.ts:67</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -243,7 +243,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L85">runtime.ts:85</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L85">runtime.ts:85</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <a href="cachedcallstack.html" class="tsd-signature-type">CachedCallStack</a></h4>
@@ -260,7 +260,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L96">runtime.ts:96</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L96">runtime.ts:96</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-parameters-title">Parameters</h4>
@@ -283,7 +283,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L73">runtime.ts:73</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L73">runtime.ts:73</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">number</span></h4>
diff --git a/docs/reference/api/typedoc/classes/instance.html b/docs/reference/api/typedoc/classes/instance.html
index 3ed6cec8b8..1d44397431 100644
--- a/docs/reference/api/typedoc/classes/instance.html
+++ b/docs/reference/api/typedoc/classes/instance.html
@@ -161,7 +161,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L844">runtime.ts:844</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L844">runtime.ts:844</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -224,7 +224,7 @@
 					<div class="tsd-signature tsd-kind-icon">exports<span class="tsd-signature-symbol">:</span> <span class="tsd-signature-type">Record</span><span class="tsd-signature-symbol">&lt;</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">&gt;</span></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L834">runtime.ts:834</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L834">runtime.ts:834</a></li>
 						</ul>
 					</aside>
 				</section>
@@ -234,7 +234,7 @@
 					<div class="tsd-signature tsd-kind-icon">memory<span class="tsd-signature-symbol">:</span> <a href="memory.html" class="tsd-signature-type">Memory</a></div>
 					<aside class="tsd-sources">
 						<ul>
-							<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L833">runtime.ts:833</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L833">runtime.ts:833</a></li>
 						</ul>
 					</aside>
 				</section>
@@ -251,7 +251,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L973">runtime.ts:973</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L973">runtime.ts:973</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -296,7 +296,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L932">runtime.ts:932</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L932">runtime.ts:932</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -318,7 +318,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L901">runtime.ts:901</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L901">runtime.ts:901</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -381,7 +381,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1215">runtime.ts:1215</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1215">runtime.ts:1215</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -412,7 +412,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1000">runtime.ts:1000</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1000">runtime.ts:1000</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -453,7 +453,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1207">runtime.ts:1207</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1207">runtime.ts:1207</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -491,7 +491,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L922">runtime.ts:922</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L922">runtime.ts:922</a></li>
 								</ul>
 							</aside>
 							<h4 class="tsd-returns-title">Returns <span class="tsd-signature-type">void</span></h4>
@@ -508,7 +508,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1235">runtime.ts:1235</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1235">runtime.ts:1235</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -552,7 +552,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L943">runtime.ts:943</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L943">runtime.ts:943</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -577,7 +577,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1088">runtime.ts:1088</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1088">runtime.ts:1088</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -609,7 +609,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1363">runtime.ts:1363</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1363">runtime.ts:1363</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -640,7 +640,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1123">runtime.ts:1123</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1123">runtime.ts:1123</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -672,7 +672,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1016">runtime.ts:1016</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1016">runtime.ts:1016</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -695,7 +695,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1281">runtime.ts:1281</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1281">runtime.ts:1281</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -729,7 +729,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L986">runtime.ts:986</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L986">runtime.ts:986</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -769,7 +769,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1341">runtime.ts:1341</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1341">runtime.ts:1341</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -817,7 +817,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1055">runtime.ts:1055</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1055">runtime.ts:1055</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -857,7 +857,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1320">runtime.ts:1320</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1320">runtime.ts:1320</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -900,7 +900,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1197">runtime.ts:1197</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1197">runtime.ts:1197</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -938,7 +938,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1491">runtime.ts:1491</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1491">runtime.ts:1491</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -990,7 +990,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1009">runtime.ts:1009</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1009">runtime.ts:1009</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -1014,7 +1014,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1151">runtime.ts:1151</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1151">runtime.ts:1151</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -1046,7 +1046,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1134">runtime.ts:1134</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1134">runtime.ts:1134</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -1078,7 +1078,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1292">runtime.ts:1292</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1292">runtime.ts:1292</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -1110,7 +1110,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L1223">runtime.ts:1223</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L1223">runtime.ts:1223</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
@@ -1141,7 +1141,7 @@
 						<li class="tsd-description">
 							<aside class="tsd-sources">
 								<ul>
-									<li>Defined in <a href="https://github.com/apache/tvm/blob/7f6da0905/web/src/runtime.ts#L957">runtime.ts:957</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/web/src/runtime.ts#L957">runtime.ts:957</a></li>
 								</ul>
 							</aside>
 							<div class="tsd-comment tsd-typography">
diff --git a/docs/reference/api/typedoc/classes/memory.html b/docs/reference/api/typedoc/classes/memory.html
index 3e5797b625..42238c06fc 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/7f6da0905/web/src/memory.ts#L40">memory.ts:40</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L32">memory.ts:32</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L33">memory.ts:33</a></li>
+							<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L154">memory.ts:154</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L90">memory.ts:90</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L97">memory.ts:97</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L74">memory.ts:74</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L81">memory.ts:81</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L104">memory.ts:104</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L132">memory.ts:132</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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/7f6da0905/web/src/memory.ts#L145">memory.ts:145</a></li>
+									<li>Defined in <a href="https://github.com/apache/tvm/blob/481930080/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">
... 2385 lines suppressed ...