Update on "[jit][tensorexpr] Added aten::batch_norm into fuser when i…

…n inference mode"

Differential Revision: [D27134348](https://our.internmc.facebook.com/intern/diff/D27134348)

[ghstack-poisoned]

.circleci/cimodel/data/simple/docker_definitions.py

@@ -10,8 +10,6 @@
     "pytorch-linux-bionic-py3.6-clang9",
     "pytorch-linux-bionic-cuda10.2-cudnn7-py3.6-clang9",
     "pytorch-linux-bionic-py3.8-gcc9",
-    "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4",
-    "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7",
     "pytorch-linux-xenial-cuda10-cudnn7-py3-gcc7",
     "pytorch-linux-xenial-cuda10.1-cudnn7-py3-gcc7",
     "pytorch-linux-xenial-cuda10.2-cudnn7-py3-gcc7",

.circleci/config.yml

@@ -472,7 +472,7 @@ jobs:
 
           docker cp /home/circleci/project/. $id:/var/lib/jenkins/workspace
 
-          export COMMAND='((echo "sudo chown -R jenkins workspace && cd workspace && .jenkins/pytorch/build.sh && find ${BUILD_ROOT} -type f -name "*.a" -or -name "*.o" -delete") | docker exec -u jenkins -i "$id" bash) 2>&1'
+          export COMMAND='((echo "sudo chown -R jenkins workspace && export CIRCLE_JOB="$CIRCLE_JOB" && cd workspace && .jenkins/pytorch/build.sh && find ${BUILD_ROOT} -type f -name "*.a" -or -name "*.o" -delete") | docker exec -u jenkins -i "$id" bash) 2>&1'
 
           echo ${COMMAND} > ./command.sh && unbuffer bash ./command.sh | ts
 
@@ -603,6 +603,7 @@ jobs:
           # =================== The following code will be executed inside Docker container ===================
           set -ex
           export SCRIBE_GRAPHQL_ACCESS_TOKEN="${SCRIBE_GRAPHQL_ACCESS_TOKEN}"
+          export CIRCLE_JOB="$CIRCLE_JOB"
           ${PARALLEL_FLAGS}
           cd workspace
           EOL
@@ -696,6 +697,11 @@ jobs:
     executor: <<parameters.executor>>
     steps:
       - checkout
+      - run:
+          name: _HACK_ Install CUDA compatible cmath
+          no_output_timeout: 1m
+          command: |
+              powershell .circleci/scripts/vs_install_cmath.ps1
       - run:
           name: Install Cuda
           no_output_timeout: 30m
@@ -6729,12 +6735,6 @@ workflows:
       - docker_build_job:
           name: "docker-pytorch-linux-bionic-py3.8-gcc9"
           image_name: "pytorch-linux-bionic-py3.8-gcc9"
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
       - docker_build_job:
           name: "docker-pytorch-linux-xenial-cuda10-cudnn7-py3-gcc7"
           image_name: "pytorch-linux-xenial-cuda10-cudnn7-py3-gcc7"
@@ -9143,32 +9143,6 @@ workflows:
           vc_product: Community
           vc_version: ""
           vc_year: "2019"
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-      - pytorch_linux_build:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_build
-          requires:
-            - "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7-build"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-      - pytorch_linux_test:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_test
-          requires:
-            - pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_build
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7-test"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          use_cuda_docker_runtime: "1"
-          resource_class: gpu.medium
-      - pytorch_linux_build:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc5_4_build
-          requires:
-            - "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4-build"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
 
   # The following allows these jobs to run on ci-all and release branches
   debuggable-scheduled-ci:
@@ -9266,57 +9240,6 @@ workflows:
               only:
                 - /ci-all\/.*/
                 - /release\/.*/
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
-      - pytorch_linux_build:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_build
-          requires:
-            - "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7-build"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
-      - pytorch_linux_test:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_test
-          requires:
-            - pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_build
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7-test"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          use_cuda_docker_runtime: "1"
-          resource_class: gpu.medium
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
-      - pytorch_linux_build:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc5_4_build
-          requires:
-            - "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4-build"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
   ecr_gc:
     triggers:
       - schedule:

.circleci/docker/build.sh

@@ -101,24 +101,6 @@ case "$image" in
     DB=yes
     VISION=yes
     ;;
-  pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4)
-    CUDA_VERSION=9.2
-    CUDNN_VERSION=7
-    ANACONDA_PYTHON_VERSION=3.6
-    GCC_VERSION=5
-    PROTOBUF=yes
-    DB=yes
-    VISION=yes
-    ;;
-  pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7)
-    CUDA_VERSION=9.2
-    CUDNN_VERSION=7
-    ANACONDA_PYTHON_VERSION=3.6
-    GCC_VERSION=7
-    PROTOBUF=yes
-    DB=yes
-    VISION=yes
-    ;;
   pytorch-linux-xenial-cuda10-cudnn7-py3-gcc7)
     CUDA_VERSION=10.0
     CUDNN_VERSION=7

.circleci/docker/common/install_conda.sh

@@ -80,9 +80,8 @@ if [ -n "$ANACONDA_PYTHON_VERSION" ]; then
   else
     conda_install numpy=1.18.5 pyyaml mkl mkl-include setuptools cffi future six dataclasses typing_extensions
   fi
-  if [[ "$CUDA_VERSION" == 9.2* ]]; then
-    conda_install magma-cuda92 -c pytorch
-  elif [[ "$CUDA_VERSION" == 10.0* ]]; then
+
+  if [[ "$CUDA_VERSION" == 10.0* ]]; then
     conda_install magma-cuda100 -c pytorch
   elif [[ "$CUDA_VERSION" == 10.1* ]]; then
     conda_install magma-cuda101 -c pytorch

.circleci/docker/common/install_rocm.sh

@@ -9,7 +9,7 @@ install_magma() {
     cp make.inc-examples/make.inc.hip-mkl-gcc make.inc
     echo 'LIBDIR += -L$(MKLROOT)/lib' >> make.inc
     echo 'LIB += -Wl,--enable-new-dtags -Wl,--rpath,/opt/rocm/lib -Wl,--rpath,$(MKLROOT)/lib -Wl,--rpath,/opt/rocm/magma/lib' >> make.inc
-    echo 'DEVCCFLAGS += --amdgpu-target=gfx803 --amdgpu-target=gfx900 --amdgpu-target=gfx906 --amdgpu-target=gfx908' >> make.inc
+    echo 'DEVCCFLAGS += --amdgpu-target=gfx803 --amdgpu-target=gfx900 --amdgpu-target=gfx906 --amdgpu-target=gfx908 --gpu-max-threads-per-block=256' >> make.inc
     export PATH="${PATH}:/opt/rocm/bin"
     make -f make.gen.hipMAGMA -j $(nproc)
     make lib/libmagma.so -j $(nproc) MKLROOT=/opt/conda

.circleci/regenerate.sh

@@ -1,10 +1,17 @@
-#!/bin/bash -xe
+#!/bin/bash -e
 
 # Allows this script to be invoked from any directory:
 cd "$(dirname "$0")"
 
-OLD_FILE=$(mktemp)
-cp config.yml "$OLD_FILE"
+UNCOMMIT_CHANGE=$(git status -s | grep -c " config.yml")
+if [[ $UNCOMMIT_CHANGE != 0 ]]; then
+    OLD_FILE=$(mktemp)
+    cp config.yml "$OLD_FILE"
+    echo "Uncommitted change detected in .circleci/config.yml"
+    echo "It has been backed up to $OLD_FILE"
+fi
+
 NEW_FILE=$(mktemp)
 ./generate_config_yml.py > "$NEW_FILE"
 cp "$NEW_FILE" config.yml
+echo "New config generated in .circleci/config.yml"

.circleci/scripts/vs_install_cmath.ps1

@@ -0,0 +1,5 @@
+$CMATH_DOWNLOAD_LINK = "https://raw.githubusercontent.com/microsoft/STL/12c684bba78f9b032050526abdebf14f58ca26a3/stl/inc/cmath"
+$VC14_28_INSTALL_PATH="C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.28.29910\include"
+
+curl.exe --retry 3 -kL $CMATH_DOWNLOAD_LINK --output "$home\cmath"
+Move-Item -Path "$home\cmath" -Destination "$VC14_28_INSTALL_PATH" -Force

.circleci/verbatim-sources/job-specs/pytorch-job-specs.yml

@@ -34,7 +34,7 @@ jobs:
 
           docker cp /home/circleci/project/. $id:/var/lib/jenkins/workspace
 
-          export COMMAND='((echo "sudo chown -R jenkins workspace && cd workspace && .jenkins/pytorch/build.sh && find ${BUILD_ROOT} -type f -name "*.a" -or -name "*.o" -delete") | docker exec -u jenkins -i "$id" bash) 2>&1'
+          export COMMAND='((echo "sudo chown -R jenkins workspace && export CIRCLE_JOB="$CIRCLE_JOB" && cd workspace && .jenkins/pytorch/build.sh && find ${BUILD_ROOT} -type f -name "*.a" -or -name "*.o" -delete") | docker exec -u jenkins -i "$id" bash) 2>&1'
 
           echo ${COMMAND} > ./command.sh && unbuffer bash ./command.sh | ts
 
@@ -165,6 +165,7 @@ jobs:
           # =================== The following code will be executed inside Docker container ===================
           set -ex
           export SCRIBE_GRAPHQL_ACCESS_TOKEN="${SCRIBE_GRAPHQL_ACCESS_TOKEN}"
+          export CIRCLE_JOB="$CIRCLE_JOB"
           ${PARALLEL_FLAGS}
           cd workspace
           EOL
@@ -258,6 +259,11 @@ jobs:
     executor: <<parameters.executor>>
     steps:
       - checkout
+      - run:
+          name: _HACK_ Install CUDA compatible cmath
+          no_output_timeout: 1m
+          command: |
+              powershell .circleci/scripts/vs_install_cmath.ps1
       - run:
           name: Install Cuda
           no_output_timeout: 30m

.circleci/verbatim-sources/workflows/workflows-scheduled-ci.yml

@@ -66,32 +66,6 @@
           vc_product: Community
           vc_version: ""
           vc_year: "2019"
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-      - pytorch_linux_build:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_build
-          requires:
-            - "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7-build"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-      - pytorch_linux_test:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_test
-          requires:
-            - pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_build
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7-test"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          use_cuda_docker_runtime: "1"
-          resource_class: gpu.medium
-      - pytorch_linux_build:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc5_4_build
-          requires:
-            - "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4-build"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
 
   # The following allows these jobs to run on ci-all and release branches
   debuggable-scheduled-ci:
@@ -189,54 +163,3 @@
               only:
                 - /ci-all\/.*/
                 - /release\/.*/
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
-      - docker_build_job:
-          name: "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          image_name: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
-      - pytorch_linux_build:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_build
-          requires:
-            - "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7-build"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
-      - pytorch_linux_test:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_test
-          requires:
-            - pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc7_build
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7-test"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc7"
-          use_cuda_docker_runtime: "1"
-          resource_class: gpu.medium
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/
-      - pytorch_linux_build:
-          name: pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc5_4_build
-          requires:
-            - "docker-pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          build_environment: "pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4-build"
-          docker_image: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4"
-          filters:
-            branches:
-              only:
-                - /ci-all\/.*/
-                - /release\/.*/

.gitmodules

@@ -54,7 +54,7 @@
     ignore = dirty
     path = third_party/zstd
     url = https://github.com/facebook/zstd.git
-[submodule "third-party/cpuinfo"]
+[submodule "third_party/cpuinfo"]
     ignore = dirty
     path = third_party/cpuinfo
     url = https://github.com/pytorch/cpuinfo.git
@@ -131,5 +131,5 @@
     path = third_party/tensorpipe
     url = https://github.com/pytorch/tensorpipe.git
 [submodule "third_party/kineto"]
-	path = third_party/kineto
-	url = https://github.com/pytorch/kineto
+    path = third_party/kineto
+    url = https://github.com/pytorch/kineto

.jenkins/pytorch/build-asan.sh

@@ -37,7 +37,6 @@ CC="clang" CXX="clang++" LDSHARED="clang --shared" \
   USE_ASAN=1 USE_CUDA=0 USE_MKLDNN=0 \
   python setup.py install
 
-
 # Test building via the sdist source tarball
 python setup.py sdist
 mkdir -p /tmp/tmp

.jenkins/pytorch/build.sh

@@ -206,7 +206,7 @@ else
     # ppc64le build fails when WERROR=1
     # set only when building other architectures
     # only use for "python setup.py install" line
-    if [[ "$BUILD_ENVIRONMENT" != *ppc64le*  && "$BUILD_ENVIRONMENT" != *clang* ]]; then
+    if [[ "$BUILD_ENVIRONMENT" != *ppc64le* && "$BUILD_ENVIRONMENT" != *clang* ]]; then
       WERROR=1 python setup.py bdist_wheel
       python -mpip install dist/*.whl
     else
@@ -319,3 +319,9 @@ if [[ "${BUILD_ENVIRONMENT}" == *xla* ]]; then
   popd
   assert_git_not_dirty
 fi
+
+if [[ "$BUILD_ENVIRONMENT" != *libtorch* && "$BUILD_ENVIRONMENT" != *bazel* ]]; then
+  # export test times so that potential sharded tests that'll branch off this build will use consistent data
+  # don't do this for libtorch as libtorch is C++ only and thus won't have python tests run on its build
+  python test/run_test.py --export-past-test-times
+fi

.jenkins/pytorch/test.sh

@@ -435,10 +435,6 @@ elif [[ "${BUILD_ENVIRONMENT}" == *vulkan-linux* ]]; then
   test_vulkan
 elif [[ "${BUILD_ENVIRONMENT}" == *-bazel-* ]]; then
   test_bazel
-elif [[ "${BUILD_ENVIRONMENT}" == pytorch-linux-xenial-cuda9.2-cudnn7-py3-gcc5.4* ]]; then
-  # test cpp extension for xenial + cuda 9.2 + gcc 5.4 to make sure
-  # cpp extension can be built correctly under this old env
-  test_cpp_extensions
 else
   install_torchvision
   test_python

.jenkins/pytorch/win-test-helpers/build_pytorch.bat

@@ -117,5 +117,8 @@ python setup.py install --cmake && sccache --show-stats && (
     echo NOTE: To run `import torch`, please make sure to activate the conda environment by running `call %CONDA_PARENT_DIR%\Miniconda3\Scripts\activate.bat %CONDA_PARENT_DIR%\Miniconda3` in Command Prompt before running Git Bash.
   ) else (
     7z a %TMP_DIR_WIN%\%IMAGE_COMMIT_TAG%.7z %CONDA_PARENT_DIR%\Miniconda3\Lib\site-packages\torch %CONDA_PARENT_DIR%\Miniconda3\Lib\site-packages\caffe2 && copy /Y "%TMP_DIR_WIN%\%IMAGE_COMMIT_TAG%.7z" "%PYTORCH_FINAL_PACKAGE_DIR%\"
+
+    :: export test times so that potential sharded tests that'll branch off this build will use consistent data
+    python test/run_test.py --export-past-test-times %PYTORCH_FINAL_PACKAGE_DIR%/.pytorch-test-times
   )
 )