Merge branch 'main' of https://github.com/pytorch/pytorch into peter/…

…cumulativeops

.ci/docker/ci_commit_pins/triton.txt

@@ -1 +1 @@
-9e3e10c5edb4a062cf547ae73e6ebfb19aad7bdf
+e6216047b8b0aef1fe8da6ca8667a3ad0a016411

.ci/docker/common/install_base.sh

@@ -77,20 +77,6 @@ install_ubuntu() {
   # see: https://github.com/pytorch/pytorch/issues/65931
   apt-get install -y libgnutls30
 
-  # cuda-toolkit does not work with gcc-11.2.0 which is default in Ubunutu 22.04
-  # see: https://github.com/NVlabs/instant-ngp/issues/119
-  if [[ "$UBUNTU_VERSION" == "22.04"* ]]; then
-    apt-get install -y g++-10
-    update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-10 30
-    update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-10 30
-    update-alternatives --install /usr/bin/gcov gcov /usr/bin/gcov-10 30
-
-    # https://www.spinics.net/lists/libreoffice/msg07549.html
-    sudo rm -rf /usr/lib/gcc/x86_64-linux-gnu/11
-    wget https://github.com/gcc-mirror/gcc/commit/2b2d97fc545635a0f6aa9c9ee3b017394bc494bf.patch -O noexecpt.patch
-    sudo patch  /usr/include/c++/10/bits/range_access.h noexecpt.patch
-  fi
-
   # Cleanup package manager
   apt-get autoclean && apt-get clean
   rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*

.ci/docker/common/install_docs_reqs.sh

@@ -7,7 +7,7 @@ if [ -n "$KATEX" ]; then
   # Ignore error if gpg-agent doesn't exist (for Ubuntu 16.04)
   apt-get install -y gpg-agent || :
 
-  curl --retry 3 -sL https://deb.nodesource.com/setup_12.x | sudo -E bash -
+  curl --retry 3 -sL https://deb.nodesource.com/setup_16.x | sudo -E bash -
   sudo apt-get install -y nodejs
 
   curl --retry 3 -sS https://dl.yarnpkg.com/debian/pubkey.gpg | sudo apt-key add -

.ci/pytorch/test.sh

@@ -267,7 +267,7 @@ test_dynamo_shard() {
       test_fx \
       test_package \
       test_legacy_vmap \
-      test_custom_op_testing \
+      test_custom_ops \
       test_content_store \
       export/test_db \
       functorch/test_dims \
@@ -666,7 +666,7 @@ test_distributed() {
   time python test/run_test.py --distributed-tests --shard "$SHARD_NUMBER" "$NUM_TEST_SHARDS" --verbose
   assert_git_not_dirty
 
-  if [[ "$BUILD_ENVIRONMENT" == *cuda* && "$SHARD_NUMBER" == 1 ]]; then
+  if [[ ("$BUILD_ENVIRONMENT" == *cuda* || "$BUILD_ENVIRONMENT" == *rocm*) && "$SHARD_NUMBER" == 1 ]]; then
     echo "Testing distributed C++ tests"
     ln -sf "$TORCH_LIB_DIR"/libtorch* "$TORCH_BIN_DIR"
     ln -sf "$TORCH_LIB_DIR"/libc10* "$TORCH_BIN_DIR"
@@ -678,30 +678,30 @@ test_distributed() {
     python test/run_test.py --cpp --verbose -i cpp/HashStoreTest
     python test/run_test.py --cpp --verbose -i cpp/TCPStoreTest
 
-    MPIEXEC=$(command -v mpiexec)
-    if [[ -n "$MPIEXEC" ]]; then
-      # NB: mpiexec only works directly with the C++ test binary here
-      MPICMD="${MPIEXEC} -np 2 $TORCH_BIN_DIR/ProcessGroupMPITest"
-      eval "$MPICMD"
-    fi
+    if [[ "$BUILD_ENVIRONMENT" == *cuda* ]]; then
+      MPIEXEC=$(command -v mpiexec)
+      if [[ -n "$MPIEXEC" ]]; then
+        # NB: mpiexec only works directly with the C++ test binary here
+        MPICMD="${MPIEXEC} -np 2 $TORCH_BIN_DIR/ProcessGroupMPITest"
+        eval "$MPICMD"
+      fi
 
-    python test/run_test.py --cpp --verbose -i cpp/ProcessGroupGlooTest
-    python test/run_test.py --cpp --verbose -i cpp/ProcessGroupNCCLTest
-    python test/run_test.py --cpp --verbose -i cpp/ProcessGroupNCCLErrorsTest
+      python test/run_test.py --cpp --verbose -i cpp/ProcessGroupGlooTest
+      python test/run_test.py --cpp --verbose -i cpp/ProcessGroupNCCLTest
+      python test/run_test.py --cpp --verbose -i cpp/ProcessGroupNCCLErrorsTest
+    fi
   fi
 }
 
 test_rpc() {
-  if [[ "$BUILD_ENVIRONMENT" != *rocm* ]]; then
-    echo "Testing RPC C++ tests"
-    # NB: the ending test_rpc must match the current function name for the current
-    # test reporting process to function as expected.
-    ln -sf "$TORCH_LIB_DIR"/libtorch* "$TORCH_BIN_DIR"
-    ln -sf "$TORCH_LIB_DIR"/libc10* "$TORCH_BIN_DIR"
-    ln -sf "$TORCH_LIB_DIR"/libtbb* "$TORCH_BIN_DIR"
+  echo "Testing RPC C++ tests"
+  # NB: the ending test_rpc must match the current function name for the current
+  # test reporting process to function as expected.
+  ln -sf "$TORCH_LIB_DIR"/libtorch* "$TORCH_BIN_DIR"
+  ln -sf "$TORCH_LIB_DIR"/libc10* "$TORCH_BIN_DIR"
+  ln -sf "$TORCH_LIB_DIR"/libtbb* "$TORCH_BIN_DIR"
 
-    CPP_TESTS_DIR="${TORCH_BIN_DIR}" python test/run_test.py --cpp --verbose -i cpp/test_cpp_rpc
-  fi
+  CPP_TESTS_DIR="${TORCH_BIN_DIR}" python test/run_test.py --cpp --verbose -i cpp/test_cpp_rpc
 }
 
 test_custom_backend() {

.clang-tidy

@@ -42,7 +42,6 @@ modernize-*,
 -modernize-use-trailing-return-type,
 -modernize-use-nodiscard,
 performance-*,
--performance-unnecessary-value-param,
 readability-container-size-empty,
 '
 HeaderFilterRegex: '^(c10/(?!test)|torch/csrc/(?!deploy/interpreter/cpython)).*$'

.git-blame-ignore-revs

@@ -30,3 +30,11 @@ e3900d2ba5c9f91a24a9ce34520794c8366d5c54
 30fb2c4abaaaa966999eab11674f25b18460e609
 # 2023-06-06 clang-format on Foreach / Multi-Tensor-Apply
 515c4279416f13fcc3c898e560f8ae8f15139a03
+# 2023-07-25 Apply UFMT to all files in benchmarks
+dd3a77bc965adf9fe8ba582ee13bb7f14c9661b0
+# 2023-07-26 Enable UFMT on a bunch of low traffic Python files outside of main files
+f70844bec783bfce43c950ccf180dc494e86f2bf
+# 2023-07-28 Apply UFMT to all non test/torch files
+e6ec0efaf87703c5f889cfc20b29be455885d58d
+# 2023-07-31 [optim][BE] split test file into logical parts: SWA, LR, optim
+a53cda1ddc15336dc1ff0ce1eff2a49cdc5f882e

.github/ci_commit_pins/vision.txt

@@ -1 +1 @@
-cc0f9d02e1e7549e9e8ef1081a0d3b5d72f1297d
+bdf16222f7e734f81af0f6ea6a5f413c7b353237

.github/ci_commit_pins/xla.txt

@@ -1 +1 @@
-f5edcb2088195db71bcd36d0f8f1b6a5e663afd8
+ca5eab87a71f80cd3168630511d02549cc7d2516

.github/scripts/gitutils.py

@@ -38,7 +38,7 @@ def fuzzy_list_to_dict(items: List[Tuple[str, str]]) -> Dict[str, List[str]]:
     """
     Converts list to dict preserving elements with duplicate keys
     """
-    rc: Dict[str, List[str]] = defaultdict(lambda: [])
+    rc: Dict[str, List[str]] = defaultdict(list)
     for key, val in items:
         rc[key].append(val)
     return dict(rc)
@@ -53,6 +53,9 @@ def _check_output(items: List[str], encoding: str = "utf-8") -> str:
         msg = f"Command `{' '.join(e.cmd)}` returned non-zero exit code {e.returncode}"
         stdout = e.stdout.decode(encoding) if e.stdout is not None else ""
         stderr = e.stderr.decode(encoding) if e.stderr is not None else ""
+        # These get swallowed up, so print them here for debugging
+        print(f"stdout: \n{stdout}")
+        print(f"stderr: \n{stderr}")
         if len(stderr) == 0:
             msg += f"\n```\n{stdout}```"
         else: