Updated docs to add the error case for torch.multinomial

Summary: Updated docs to add the error condition for torch.multinomial

Test Plan: No change in code

Reviewers:

Subscribers: @drisspg

Tasks:

Tags:

.ci/docker/aotriton_version.txt

@@ -1,5 +0,0 @@
-0.6b
-manylinux_2_17
-rocm6
-04b5df8c8123f90cba3ede7e971e6fbc6040d506
-3db6ecbc915893ff967abd6e1b43bd5f54949868873be60dc802086c3863e648

.ci/docker/build.sh

@@ -84,30 +84,16 @@ fi
 # CMake 3.18 is needed to support CUDA17 language variant
 CMAKE_VERSION=3.18.5
 
-_UCX_COMMIT=7bb2722ff2187a0cad557ae4a6afa090569f83fb
-_UCC_COMMIT=20eae37090a4ce1b32bcce6144ccad0b49943e0b
+_UCX_COMMIT=00bcc6bb18fc282eb160623b4c0d300147f579af
+_UCC_COMMIT=7cb07a76ccedad7e56ceb136b865eb9319c258ea
 
 # It's annoying to rename jobs every time you want to rewrite a
 # configuration, so we hardcode everything here rather than do it
 # from scratch
 case "$image" in
-  pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9)
-    CUDA_VERSION=12.4.0
-    CUDNN_VERSION=9
-    ANACONDA_PYTHON_VERSION=3.10
-    GCC_VERSION=9
-    PROTOBUF=yes
-    DB=yes
-    VISION=yes
-    KATEX=yes
-    UCX_COMMIT=${_UCX_COMMIT}
-    UCC_COMMIT=${_UCC_COMMIT}
-    CONDA_CMAKE=yes
-    TRITON=yes
-    ;;
-  pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9)
+  pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9)
     CUDA_VERSION=12.1.1
-    CUDNN_VERSION=9
+    CUDNN_VERSION=8
     ANACONDA_PYTHON_VERSION=3.10
     GCC_VERSION=9
     PROTOBUF=yes
@@ -119,24 +105,9 @@ case "$image" in
     CONDA_CMAKE=yes
     TRITON=yes
     ;;
-  pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks)
-    CUDA_VERSION=12.4.0
-    CUDNN_VERSION=9
-    ANACONDA_PYTHON_VERSION=3.10
-    GCC_VERSION=9
-    PROTOBUF=yes
-    DB=yes
-    VISION=yes
-    KATEX=yes
-    UCX_COMMIT=${_UCX_COMMIT}
-    UCC_COMMIT=${_UCC_COMMIT}
-    CONDA_CMAKE=yes
-    TRITON=yes
-    INDUCTOR_BENCHMARKS=yes
-    ;;
-  pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks)
+  pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9-inductor-benchmarks)
     CUDA_VERSION=12.1.1
-    CUDNN_VERSION=9
+    CUDNN_VERSION=8
     ANACONDA_PYTHON_VERSION=3.10
     GCC_VERSION=9
     PROTOBUF=yes
@@ -149,39 +120,9 @@ case "$image" in
     TRITON=yes
     INDUCTOR_BENCHMARKS=yes
     ;;
-  pytorch-linux-focal-cuda12.1-cudnn9-py3.12-gcc9-inductor-benchmarks)
-    CUDA_VERSION=12.1.1
-    CUDNN_VERSION=9
-    ANACONDA_PYTHON_VERSION=3.12
-    GCC_VERSION=9
-    PROTOBUF=yes
-    DB=yes
-    VISION=yes
-    KATEX=yes
-    UCX_COMMIT=${_UCX_COMMIT}
-    UCC_COMMIT=${_UCC_COMMIT}
-    CONDA_CMAKE=yes
-    TRITON=yes
-    INDUCTOR_BENCHMARKS=yes
-    ;;
-  pytorch-linux-focal-cuda12.4-cudnn9-py3.12-gcc9-inductor-benchmarks)
-    CUDA_VERSION=12.4.0
-    CUDNN_VERSION=9
-    ANACONDA_PYTHON_VERSION=3.12
-    GCC_VERSION=9
-    PROTOBUF=yes
-    DB=yes
-    VISION=yes
-    KATEX=yes
-    UCX_COMMIT=${_UCX_COMMIT}
-    UCC_COMMIT=${_UCC_COMMIT}
-    CONDA_CMAKE=yes
-    TRITON=yes
-    INDUCTOR_BENCHMARKS=yes
-    ;;
-  pytorch-linux-focal-cuda11.8-cudnn9-py3-gcc9)
+  pytorch-linux-focal-cuda11.8-cudnn8-py3-gcc9)
     CUDA_VERSION=11.8.0
-    CUDNN_VERSION=9
+    CUDNN_VERSION=8
     ANACONDA_PYTHON_VERSION=3.10
     GCC_VERSION=9
     PROTOBUF=yes
@@ -193,37 +134,9 @@ case "$image" in
     CONDA_CMAKE=yes
     TRITON=yes
     ;;
-  pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9)
-    CUDA_VERSION=12.4.0
-    CUDNN_VERSION=9
-    ANACONDA_PYTHON_VERSION=3.10
-    GCC_VERSION=9
-    PROTOBUF=yes
-    DB=yes
-    VISION=yes
-    KATEX=yes
-    UCX_COMMIT=${_UCX_COMMIT}
-    UCC_COMMIT=${_UCC_COMMIT}
-    CONDA_CMAKE=yes
-    TRITON=yes
-    ;;
-  pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9)
+  pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9)
     CUDA_VERSION=12.1.1
-    CUDNN_VERSION=9
-    ANACONDA_PYTHON_VERSION=3.10
-    GCC_VERSION=9
-    PROTOBUF=yes
-    DB=yes
-    VISION=yes
-    KATEX=yes
-    UCX_COMMIT=${_UCX_COMMIT}
-    UCC_COMMIT=${_UCC_COMMIT}
-    CONDA_CMAKE=yes
-    TRITON=yes
-    ;;
-  pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9)
-    CUDA_VERSION=12.4.0
-    CUDNN_VERSION=9
+    CUDNN_VERSION=8
     ANACONDA_PYTHON_VERSION=3.10
     GCC_VERSION=9
     PROTOBUF=yes
@@ -291,7 +204,7 @@ case "$image" in
     PROTOBUF=yes
     DB=yes
     VISION=yes
-    ROCM_VERSION=6.0
+    ROCM_VERSION=5.7
     NINJA_VERSION=1.9.0
     CONDA_CMAKE=yes
     TRITON=yes
@@ -302,7 +215,7 @@ case "$image" in
     PROTOBUF=yes
     DB=yes
     VISION=yes
-    ROCM_VERSION=6.1
+    ROCM_VERSION=6.0
     NINJA_VERSION=1.9.0
     CONDA_CMAKE=yes
     TRITON=yes
@@ -313,7 +226,7 @@ case "$image" in
     PROTOBUF=yes
     DB=yes
     VISION=yes
-    XPU_VERSION=0.5
+    BASEKIT_VERSION=2024.0.0-49522
     NINJA_VERSION=1.9.0
     CONDA_CMAKE=yes
     TRITON=yes
@@ -330,10 +243,10 @@ case "$image" in
     DOCS=yes
     INDUCTOR_BENCHMARKS=yes
     ;;
-  pytorch-linux-jammy-cuda11.8-cudnn9-py3.8-clang12)
+  pytorch-linux-jammy-cuda11.8-cudnn8-py3.8-clang12)
     ANACONDA_PYTHON_VERSION=3.8
     CUDA_VERSION=11.8
-    CUDNN_VERSION=9
+    CUDNN_VERSION=8
     CLANG_VERSION=12
     PROTOBUF=yes
     DB=yes
@@ -380,7 +293,7 @@ case "$image" in
     ANACONDA_PYTHON_VERSION=3.9
     CONDA_CMAKE=yes
     ;;
-  pytorch-linux-jammy-cuda11.8-cudnn9-py3.9-linter)
+  pytorch-linux-jammy-cuda11.8-cudnn8-py3.9-linter)
     ANACONDA_PYTHON_VERSION=3.9
     CUDA_VERSION=11.8
     CONDA_CMAKE=yes
@@ -393,12 +306,6 @@ case "$image" in
     DB=yes
     VISION=yes
     CONDA_CMAKE=yes
-    # snadampal: skipping sccache due to the following issue
-    # https://github.com/pytorch/pytorch/issues/121559
-    SKIP_SCCACHE_INSTALL=yes
-    # snadampal: skipping llvm src build install because the current version
-    # from pytorch/llvm:9.0.1 is x86 specific
-    SKIP_LLVM_SRC_BUILD_INSTALL=yes
     ;;
   *)
     # Catch-all for builds that are not hardcoded.
@@ -447,13 +354,13 @@ tmp_tag=$(basename "$(mktemp -u)" | tr '[:upper:]' '[:lower:]')
 #when using cudnn version 8 install it separately from cuda
 if [[ "$image" == *cuda*  && ${OS} == "ubuntu" ]]; then
   IMAGE_NAME="nvidia/cuda:${CUDA_VERSION}-cudnn${CUDNN_VERSION}-devel-ubuntu${UBUNTU_VERSION}"
-  if [[ ${CUDNN_VERSION} == 9 ]]; then
+  if [[ ${CUDNN_VERSION} == 8 ]]; then
     IMAGE_NAME="nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION}"
   fi
 fi
 
 # Build image
-docker build \
+DOCKER_BUILDKIT=1 docker build \
        --no-cache \
        --progress=plain \
        --build-arg "BUILD_ENVIRONMENT=${image}" \
@@ -490,16 +397,14 @@ docker build \
        --build-arg "DOCS=${DOCS}" \
        --build-arg "INDUCTOR_BENCHMARKS=${INDUCTOR_BENCHMARKS}" \
        --build-arg "EXECUTORCH=${EXECUTORCH}" \
-       --build-arg "XPU_VERSION=${XPU_VERSION}" \
+       --build-arg "BASEKIT_VERSION=${BASEKIT_VERSION}" \
        --build-arg "ACL=${ACL:-}" \
-       --build-arg "SKIP_SCCACHE_INSTALL=${SKIP_SCCACHE_INSTALL:-}" \
-       --build-arg "SKIP_LLVM_SRC_BUILD_INSTALL=${SKIP_LLVM_SRC_BUILD_INSTALL:-}" \
        -f $(dirname ${DOCKERFILE})/Dockerfile \
        -t "$tmp_tag" \
        "$@" \
        .
 
-# NVIDIA dockers for RC releases use tag names like `11.0-cudnn9-devel-ubuntu18.04-rc`,
+# NVIDIA dockers for RC releases use tag names like `11.0-cudnn8-devel-ubuntu18.04-rc`,
 # for this case we will set UBUNTU_VERSION to `18.04-rc` so that the Dockerfile could
 # find the correct image. As a result, here we have to replace the
 #   "$UBUNTU_VERSION" == "18.04-rc"

.ci/docker/centos-rocm/Dockerfile

@@ -62,7 +62,7 @@ RUN if [ -n "${DB}" ]; then bash ./install_db.sh; fi
 RUN rm install_db.sh
 ENV INSTALLED_DB ${DB}
 
-# (optional) Install vision packages like OpenCV
+# (optional) Install vision packages like OpenCV and ffmpeg
 ARG VISION
 COPY ./common/install_vision.sh ./common/cache_vision_models.sh ./common/common_utils.sh ./
 RUN if [ -n "${VISION}" ]; then bash ./install_vision.sh; fi
@@ -77,9 +77,6 @@ RUN rm install_rocm.sh
 COPY ./common/install_rocm_magma.sh install_rocm_magma.sh
 RUN bash ./install_rocm_magma.sh
 RUN rm install_rocm_magma.sh
-COPY ./common/install_amdsmi.sh install_amdsmi.sh
-RUN bash ./install_amdsmi.sh
-RUN rm install_amdsmi.sh
 ENV PATH /opt/rocm/bin:$PATH
 ENV PATH /opt/rocm/hcc/bin:$PATH
 ENV PATH /opt/rocm/hip/bin:$PATH
@@ -113,13 +110,6 @@ COPY triton_version.txt triton_version.txt
 RUN if [ -n "${TRITON}" ]; then bash ./install_triton.sh; fi
 RUN rm install_triton.sh common_utils.sh triton-rocm.txt triton_version.txt
 
-# Install AOTriton (Early fail)
-COPY ./aotriton_version.txt aotriton_version.txt
-COPY ./common/common_utils.sh common_utils.sh
-COPY ./common/install_aotriton.sh install_aotriton.sh
-RUN ["/bin/bash", "-c", "./install_aotriton.sh /opt/rocm && rm -rf install_aotriton.sh aotriton_version.txt common_utils.sh"]
-ENV AOTRITON_INSTALLED_PREFIX /opt/rocm/aotriton
-
 # Install ccache/sccache (do this last, so we get priority in PATH)
 COPY ./common/install_cache.sh install_cache.sh
 ENV PATH /opt/cache/bin:$PATH

.ci/docker/ci_commit_pins/triton-rocm.txt

@@ -1 +1 @@
-01cbe5045a6898c9a925f01435c8277b2fe6afcc
+bbe6246e37d8aa791c67daaf9d9d61b26c9ccfdc

.ci/docker/common/install_acl.sh

@@ -1,6 +1,6 @@
 set -euo pipefail
 
-readonly version=v24.04
+readonly version=v23.08
 readonly src_host=https://review.mlplatform.org/ml
 readonly src_repo=ComputeLibrary
 

.ci/docker/common/install_amdsmi.sh

@@ -1,5 +0,0 @@
-#!/bin/bash
-
-set -ex
-
-cd /opt/rocm/share/amd_smi && pip install .

.ci/docker/common/install_aotriton.sh

@@ -1,23 +0,0 @@
-#!/bin/bash
-
-set -ex
-
-source "$(dirname "${BASH_SOURCE[0]}")/common_utils.sh"
-
-TARBALL='aotriton.tar.bz2'
-# This read command alwasy returns with exit code 1
-read -d "\n" VER MANYLINUX ROCMBASE PINNED_COMMIT SHA256 < aotriton_version.txt || true
-ARCH=$(uname -m)
-AOTRITON_INSTALL_PREFIX="$1"
-AOTRITON_URL="https://github.com/ROCm/aotriton/releases/download/${VER}/aotriton-${VER}-${MANYLINUX}_${ARCH}-${ROCMBASE}.tar.bz2"
-
-cd "${AOTRITON_INSTALL_PREFIX}"
-# Must use -L to follow redirects
-curl -L --retry 3 -o "${TARBALL}" "${AOTRITON_URL}"
-ACTUAL_SHA256=$(sha256sum "${TARBALL}" | cut -d " " -f 1)
-if [ "${SHA256}" != "${ACTUAL_SHA256}" ]; then
-  echo -n "Error: The SHA256 of downloaded tarball is ${ACTUAL_SHA256},"
-  echo " which does not match the expected value ${SHA256}."
-  exit
-fi
-tar xf "${TARBALL}" && rm -rf "${TARBALL}"

.ci/docker/common/install_base.sh

@@ -3,7 +3,7 @@
 set -ex
 
 install_ubuntu() {
-  # NVIDIA dockers for RC releases use tag names like `11.0-cudnn9-devel-ubuntu18.04-rc`,
+  # NVIDIA dockers for RC releases use tag names like `11.0-cudnn8-devel-ubuntu18.04-rc`,
   # for this case we will set UBUNTU_VERSION to `18.04-rc` so that the Dockerfile could
   # find the correct image. As a result, here we have to check for
   #   "$UBUNTU_VERSION" == "18.04"*
@@ -113,6 +113,7 @@ install_centos() {
     glibc-devel \
     glibc-headers \
     glog-devel \
+    hiredis-devel \
     libstdc++-devel \
     libsndfile-devel \
     make \

.ci/docker/common/install_cudnn.sh

@@ -1,18 +1,20 @@
 #!/bin/bash
 
-if [[ -n "${CUDNN_VERSION}" ]]; then
+if [[ ${CUDNN_VERSION} == 8 ]]; then
     # cuDNN license: https://developer.nvidia.com/cudnn/license_agreement
     mkdir tmp_cudnn
     pushd tmp_cudnn
-    if [[ ${CUDA_VERSION:0:2} == "12" ]]; then
-        CUDNN_NAME="cudnn-linux-x86_64-9.1.0.70_cuda12-archive"
-    elif [[ ${CUDA_VERSION:0:2} == "11" ]]; then
-        CUDNN_NAME="cudnn-linux-x86_64-9.1.0.70_cuda11-archive"
+    if [[ ${CUDA_VERSION:0:4} == "12.1" ]]; then
+        CUDNN_NAME="cudnn-linux-x86_64-8.9.2.26_cuda12-archive"
+        curl --retry 3 -OLs https://developer.download.nvidia.com/compute/cudnn/redist/cudnn/linux-x86_64/${CUDNN_NAME}.tar.xz
+    elif [[ ${CUDA_VERSION:0:4} == "11.8" ]]; then
+        CUDNN_NAME="cudnn-linux-x86_64-8.7.0.84_cuda11-archive"
+        curl --retry 3 -OLs https://developer.download.nvidia.com/compute/redist/cudnn/v8.7.0/local_installers/11.8/${CUDNN_NAME}.tar.xz
     else
         print "Unsupported CUDA version ${CUDA_VERSION}"
         exit 1
     fi
-    curl --retry 3 -OLs https://developer.download.nvidia.com/compute/cudnn/redist/cudnn/linux-x86_64/${CUDNN_NAME}.tar.xz
+
     tar xf ${CUDNN_NAME}.tar.xz
     cp -a ${CUDNN_NAME}/include/* /usr/local/cuda/include/
     cp -a ${CUDNN_NAME}/lib/* /usr/local/cuda/lib64/

.ci/docker/common/install_cusparselt.sh

@@ -5,14 +5,9 @@ set -ex
 # cuSPARSELt license: https://docs.nvidia.com/cuda/cusparselt/license.html
 mkdir tmp_cusparselt && cd tmp_cusparselt
 
-if [[ ${CUDA_VERSION:0:4} =~ ^12\.[1-4]$ ]]; then
-    arch_path='sbsa'
-    export TARGETARCH=${TARGETARCH:-$(uname -m)}
-    if [ ${TARGETARCH} = 'amd64' ] || [ "${TARGETARCH}" = 'x86_64' ]; then
-        arch_path='x86_64'
-    fi
-    CUSPARSELT_NAME="libcusparse_lt-linux-${arch_path}-0.5.2.1-archive"
-    curl --retry 3 -OLs https://developer.download.nvidia.com/compute/cusparselt/redist/libcusparse_lt/linux-${arch_path}/${CUSPARSELT_NAME}.tar.xz
+if [[ ${CUDA_VERSION:0:4} == "12.1" ]]; then
+    CUSPARSELT_NAME="libcusparse_lt-linux-x86_64-0.5.2.1-archive"
+    curl --retry 3 -OLs https://developer.download.nvidia.com/compute/cusparselt/redist/libcusparse_lt/linux-x86_64/${CUSPARSELT_NAME}.tar.xz
 elif [[ ${CUDA_VERSION:0:4} == "11.8" ]]; then
     CUSPARSELT_NAME="libcusparse_lt-linux-x86_64-0.4.0.7-archive"
     curl --retry 3 -OLs https://developer.download.nvidia.com/compute/cusparselt/redist/libcusparse_lt/linux-x86_64/${CUSPARSELT_NAME}.tar.xz

.ci/docker/common/install_db.sh

@@ -4,6 +4,11 @@ set -ex
 
 install_ubuntu() {
   apt-get update
+  apt-get install -y --no-install-recommends \
+          libhiredis-dev \
+          libleveldb-dev \
+          liblmdb-dev \
+          libsnappy-dev
 
   # Cleanup
   apt-get autoclean && apt-get clean
@@ -15,6 +20,12 @@ install_centos() {
   # See http://fedoraproject.org/wiki/EPEL
   yum --enablerepo=extras install -y epel-release
 
+  yum install -y \
+      hiredis-devel \
+      leveldb-devel \
+      lmdb-devel \
+      snappy-devel
+
   # Cleanup
   yum clean all
   rm -rf /var/cache/yum

.ci/docker/common/install_onnx.sh

@@ -30,10 +30,10 @@ pip_install \
 
 pip_install coloredlogs packaging
 
-pip_install onnxruntime==1.18
-pip_install onnx==1.16.0
+pip_install onnxruntime==1.17.0
+pip_install onnx==1.15.0
 # pip_install "onnxscript@git+https://github.com/microsoft/onnxscript@3e869ef8ccf19b5ebd21c10d3e9c267c9a9fa729" --no-deps
-pip_install onnxscript==0.1.0.dev20240523 --no-deps
+pip_install onnxscript==0.1.0.dev20240315 --no-deps
 
 # Cache the transformers model to be used later by ONNX tests. We need to run the transformers
 # package to download the model. By default, the model is cached at ~/.cache/huggingface/hub/

.ci/docker/common/install_rocm.sh

@@ -6,6 +6,9 @@ ver() {
     printf "%3d%03d%03d%03d" $(echo "$1" | tr '.' ' ');
 }
 
+# Map ROCm version to AMDGPU version
+declare -A AMDGPU_VERSIONS=( ["5.0"]="21.50" ["5.1.1"]="22.10.1" ["5.2"]="22.20" )
+
 install_ubuntu() {
     apt-get update
     if [[ $UBUNTU_VERSION == 18.04 ]]; then
@@ -23,14 +26,31 @@ install_ubuntu() {
     apt-get install -y libc++1
     apt-get install -y libc++abi1
 
-    # Add amdgpu repository
-    UBUNTU_VERSION_NAME=`cat /etc/os-release | grep UBUNTU_CODENAME | awk -F= '{print $2}'`
-    echo "deb [arch=amd64] https://repo.radeon.com/amdgpu/${ROCM_VERSION}/ubuntu ${UBUNTU_VERSION_NAME} main" > /etc/apt/sources.list.d/amdgpu.list
+    if [[ $(ver $ROCM_VERSION) -ge $(ver 4.5) ]]; then
+        # Add amdgpu repository
+        UBUNTU_VERSION_NAME=`cat /etc/os-release | grep UBUNTU_CODENAME | awk -F= '{print $2}'`
+        local amdgpu_baseurl
+        if [[ $(ver $ROCM_VERSION) -ge $(ver 5.3) ]]; then
+          amdgpu_baseurl="https://repo.radeon.com/amdgpu/${ROCM_VERSION}/ubuntu"
+        else
+          amdgpu_baseurl="https://repo.radeon.com/amdgpu/${AMDGPU_VERSIONS[$ROCM_VERSION]}/ubuntu"
+        fi
+        echo "deb [arch=amd64] ${amdgpu_baseurl} ${UBUNTU_VERSION_NAME} main" > /etc/apt/sources.list.d/amdgpu.list
+    fi
+
+    ROCM_REPO="ubuntu"
+    if [[ $(ver $ROCM_VERSION) -lt $(ver 4.2) ]]; then
+        ROCM_REPO="xenial"
+    fi
+
+    if [[ $(ver $ROCM_VERSION) -ge $(ver 5.3) ]]; then
+        ROCM_REPO="${UBUNTU_VERSION_NAME}"
+    fi
 
     # Add rocm repository
     wget -qO - http://repo.radeon.com/rocm/rocm.gpg.key | apt-key add -
     local rocm_baseurl="http://repo.radeon.com/rocm/apt/${ROCM_VERSION}"
-    echo "deb [arch=amd64] ${rocm_baseurl} ${UBUNTU_VERSION_NAME} main" > /etc/apt/sources.list.d/rocm.list
+    echo "deb [arch=amd64] ${rocm_baseurl} ${ROCM_REPO} main" > /etc/apt/sources.list.d/rocm.list
     apt-get update --allow-insecure-repositories
 
     DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated \
@@ -39,28 +59,34 @@ install_ubuntu() {
                    rocm-libs \
                    rccl \
                    rocprofiler-dev \
-                   roctracer-dev \
-                   amd-smi-lib
-
-    if [[ $(ver $ROCM_VERSION) -ge $(ver 6.1) ]]; then
-        DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated rocm-llvm-dev
-    fi
+                   roctracer-dev
 
     # precompiled miopen kernels added in ROCm 3.5, renamed in ROCm 5.5
     # search for all unversioned packages
     # if search fails it will abort this script; use true to avoid case where search fails
-    MIOPENHIPGFX=$(apt-cache search --names-only miopen-hip-gfx | awk '{print $1}' | grep -F -v . || true)
-    if [[ "x${MIOPENHIPGFX}" = x ]]; then
-      echo "miopen-hip-gfx package not available" && exit 1
+    if [[ $(ver $ROCM_VERSION) -ge $(ver 5.5) ]]; then
+        MIOPENHIPGFX=$(apt-cache search --names-only miopen-hip-gfx | awk '{print $1}' | grep -F -v . || true)
+        if [[ "x${MIOPENHIPGFX}" = x ]]; then
+          echo "miopen-hip-gfx package not available" && exit 1
+        else
+          DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated ${MIOPENHIPGFX}
+        fi
     else
-      DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated ${MIOPENHIPGFX}
+        MIOPENKERNELS=$(apt-cache search --names-only miopenkernels | awk '{print $1}' | grep -F -v . || true)
+        if [[ "x${MIOPENKERNELS}" = x ]]; then
+          echo "miopenkernels package not available" && exit 1
+        else
+          DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated ${MIOPENKERNELS}
+        fi
     fi
 
     # ROCm 6.0 had a regression where journal_mode was enabled on the kdb files resulting in permission errors at runtime
-    for kdb in /opt/rocm/share/miopen/db/*.kdb
-    do
-        sqlite3 $kdb "PRAGMA journal_mode=off; PRAGMA VACUUM;"
-    done
+    if [[ $(ver $ROCM_VERSION) -ge $(ver 6.0) ]]; then
+        for kdb in /opt/rocm/share/miopen/db/*.kdb
+        do
+            sqlite3 $kdb "PRAGMA journal_mode=off; PRAGMA VACUUM;"
+        done
+    fi
 
     # Cleanup
     apt-get autoclean && apt-get clean
@@ -77,19 +103,25 @@ install_centos() {
   yum install -y epel-release
   yum install -y dkms kernel-headers-`uname -r` kernel-devel-`uname -r`
 
-  # Add amdgpu repository
-  local amdgpu_baseurl
-  if [[ $OS_VERSION == 9 ]]; then
-      amdgpu_baseurl="https://repo.radeon.com/amdgpu/${ROCM_VERSION}/rhel/9.0/main/x86_64"
-  else
-      amdgpu_baseurl="https://repo.radeon.com/amdgpu/${ROCM_VERSION}/rhel/7.9/main/x86_64"
+  if [[ $(ver $ROCM_VERSION) -ge $(ver 4.5) ]]; then
+      # Add amdgpu repository
+      local amdgpu_baseurl
+      if [[ $OS_VERSION == 9 ]]; then
+          amdgpu_baseurl="https://repo.radeon.com/amdgpu/${AMDGPU_VERSIONS[$ROCM_VERSION]}/rhel/9.0/main/x86_64"
+      else
+        if [[ $(ver $ROCM_VERSION) -ge $(ver 5.3) ]]; then
+          amdgpu_baseurl="https://repo.radeon.com/amdgpu/${ROCM_VERSION}/rhel/7.9/main/x86_64"
+        else
+          amdgpu_baseurl="https://repo.radeon.com/amdgpu/${AMDGPU_VERSIONS[$ROCM_VERSION]}/rhel/7.9/main/x86_64"
+        fi
+      fi
+      echo "[AMDGPU]" > /etc/yum.repos.d/amdgpu.repo
+      echo "name=AMDGPU" >> /etc/yum.repos.d/amdgpu.repo
+      echo "baseurl=${amdgpu_baseurl}" >> /etc/yum.repos.d/amdgpu.repo
+      echo "enabled=1" >> /etc/yum.repos.d/amdgpu.repo
+      echo "gpgcheck=1" >> /etc/yum.repos.d/amdgpu.repo
+      echo "gpgkey=http://repo.radeon.com/rocm/rocm.gpg.key" >> /etc/yum.repos.d/amdgpu.repo
   fi
-  echo "[AMDGPU]" > /etc/yum.repos.d/amdgpu.repo
-  echo "name=AMDGPU" >> /etc/yum.repos.d/amdgpu.repo
-  echo "baseurl=${amdgpu_baseurl}" >> /etc/yum.repos.d/amdgpu.repo
-  echo "enabled=1" >> /etc/yum.repos.d/amdgpu.repo
-  echo "gpgcheck=1" >> /etc/yum.repos.d/amdgpu.repo
-  echo "gpgkey=http://repo.radeon.com/rocm/rocm.gpg.key" >> /etc/yum.repos.d/amdgpu.repo
 
   local rocm_baseurl="http://repo.radeon.com/rocm/yum/${ROCM_VERSION}"
   echo "[ROCm]" > /etc/yum.repos.d/rocm.repo
@@ -107,23 +139,33 @@ install_centos() {
                    rocm-libs \
                    rccl \
                    rocprofiler-dev \
-                   roctracer-dev \
-                   amd-smi-lib
+                   roctracer-dev
 
   # precompiled miopen kernels; search for all unversioned packages
   # if search fails it will abort this script; use true to avoid case where search fails
-  MIOPENHIPGFX=$(yum -q search miopen-hip-gfx | grep miopen-hip-gfx | awk '{print $1}'| grep -F kdb. || true)
-  if [[ "x${MIOPENHIPGFX}" = x ]]; then
-    echo "miopen-hip-gfx package not available" && exit 1
+  if [[ $(ver $ROCM_VERSION) -ge $(ver 5.5) ]]; then
+      MIOPENHIPGFX=$(yum -q search miopen-hip-gfx | grep miopen-hip-gfx | awk '{print $1}'| grep -F kdb. || true)
+      if [[ "x${MIOPENHIPGFX}" = x ]]; then
+        echo "miopen-hip-gfx package not available" && exit 1
+      else
+        yum install -y ${MIOPENHIPGFX}
+      fi
   else
-    yum install -y ${MIOPENHIPGFX}
+      MIOPENKERNELS=$(yum -q search miopenkernels | grep miopenkernels- | awk '{print $1}'| grep -F kdb. || true)
+      if [[ "x${MIOPENKERNELS}" = x ]]; then
+        echo "miopenkernels package not available" && exit 1
+      else
+        yum install -y ${MIOPENKERNELS}
+      fi
   fi
 
   # ROCm 6.0 had a regression where journal_mode was enabled on the kdb files resulting in permission errors at runtime
-  for kdb in /opt/rocm/share/miopen/db/*.kdb
-  do
-      sqlite3 $kdb "PRAGMA journal_mode=off; PRAGMA VACUUM;"
-  done
+  if [[ $(ver $ROCM_VERSION) -ge $(ver 6.0) ]]; then
+      for kdb in /opt/rocm/share/miopen/db/*.kdb
+      do
+          sqlite3 $kdb "PRAGMA journal_mode=off; PRAGMA VACUUM;"
+      done
+  fi
 
   # Cleanup
   yum clean all

.ci/docker/common/install_triton.sh

@@ -15,7 +15,7 @@ conda_reinstall() {
 if [ -n "${ROCM_VERSION}" ]; then
   TRITON_REPO="https://github.com/openai/triton"
   TRITON_TEXT_FILE="triton-rocm"
-elif [ -n "${XPU_VERSION}" ]; then
+elif [ -n "${BASEKIT_VERSION}" ]; then
   TRITON_REPO="https://github.com/intel/intel-xpu-backend-for-triton"
   TRITON_TEXT_FILE="triton-xpu"
 else

.ci/docker/common/install_vision.sh

@@ -5,7 +5,8 @@ set -ex
 install_ubuntu() {
   apt-get update
   apt-get install -y --no-install-recommends \
-          libopencv-dev
+          libopencv-dev \
+          libavcodec-dev
 
   # Cleanup
   apt-get autoclean && apt-get clean
@@ -18,7 +19,8 @@ install_centos() {
   yum --enablerepo=extras install -y epel-release
 
   yum install -y \
-      opencv-devel
+      opencv-devel \
+      ffmpeg-devel
 
   # Cleanup
   yum clean all

.ci/docker/common/install_xpu.sh

@@ -3,7 +3,10 @@ set -xe
 
 
 # Intel® software for general purpose GPU capabilities.
-# Refer to https://www.intel.com/content/www/us/en/developer/articles/tool/pytorch-prerequisites-for-intel-gpus.html
+# Refer to https://dgpu-docs.intel.com/releases/LTS_803.29_20240131.html
+
+# Intel® oneAPI Base Toolkit (version 2024.0.0) has been updated to include functional and security updates.
+# Refer to https://www.intel.com/content/www/us/en/developer/tools/oneapi/base-toolkit-download.html
 
 # Users should update to the latest version as it becomes available
 
@@ -14,16 +17,14 @@ function install_ubuntu() {
     # Set up the repository. To do this, download the key to the system keyring
     wget -qO - https://repositories.intel.com/gpu/intel-graphics.key \
         | gpg --dearmor --output /usr/share/keyrings/intel-graphics.gpg
-    wget -qO - https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB \
-        | gpg --dearmor --output /usr/share/keyrings/intel-for-pytorch-gpu-dev-keyring.gpg
+    wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB \
+        | gpg --dearmor | tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null
 
     # Add the signed entry to APT sources and configure the APT client to use the Intel repository
-    echo "deb [arch=amd64 signed-by=/usr/share/keyrings/intel-graphics.gpg] \
-        https://repositories.intel.com/gpu/ubuntu jammy/lts/2350 unified" \
+    echo "deb [arch=amd64 signed-by=/usr/share/keyrings/intel-graphics.gpg] https://repositories.intel.com/gpu/ubuntu jammy/lts/2350 unified" \
         | tee /etc/apt/sources.list.d/intel-gpu-jammy.list
-    echo "deb [signed-by=/usr/share/keyrings/intel-for-pytorch-gpu-dev-keyring.gpg] \
-        https://apt.repos.intel.com/intel-for-pytorch-gpu-dev all main" \
-        | tee /etc/apt/sources.list.d/intel-for-pytorch-gpu-dev.list
+    echo "deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main" \
+        | tee /etc/apt/sources.list.d/oneAPI.list
 
     # Update the packages list and repository index
     apt-get update
@@ -39,11 +40,11 @@ function install_ubuntu() {
         mesa-vdpau-drivers mesa-vulkan-drivers va-driver-all vainfo hwinfo clinfo
     # Development Packages
     apt-get install -y libigc-dev intel-igc-cm libigdfcl-dev libigfxcmrt-dev level-zero-dev
-    # Install Intel Support Packages
-    if [ -n "$XPU_VERSION" ]; then
-        apt-get install -y intel-for-pytorch-gpu-dev-${XPU_VERSION}
+    # Install Intel® oneAPI Base Toolkit
+    if [ -n "$BASEKIT_VERSION" ]; then
+        apt-get install intel-basekit=$BASEKIT_VERSION -y
     else
-        apt-get install -y intel-for-pytorch-gpu-dev
+        apt-get install intel-basekit -y
     fi
 
     # Cleanup

.ci/docker/requirements-ci.txt

@@ -263,10 +263,10 @@ unittest-xml-reporting<=3.2.0,>=2.0.0
 #Pinned versions:
 #test that import:
 
-#lintrunner is supported on aarch64-linux only from 0.12.4 version
-lintrunner==0.12.5
+#wheel not found on aarch64, and source build requires rust
+lintrunner==0.10.7 ; platform_machine == "x86_64"
 #Description: all about linters!
-#Pinned versions: 0.12.5
+#Pinned versions: 0.10.7
 #test that import:
 
 rockset==1.0.3
@@ -279,9 +279,9 @@ ghstack==0.8.0
 #Pinned versions: 0.8.0
 #test that import:
 
-jinja2==3.1.4
+jinja2==3.1.3
 #Description: jinja2 template engine
-#Pinned versions: 3.1.4
+#Pinned versions: 3.1.3
 #test that import:
 
 pytest-cpp==2.3.0
@@ -310,5 +310,3 @@ lxml==5.0.0.
 #Description: This is a requirement of unittest-xml-reporting
 
 # Python-3.9 binaries
-
-PyGithub==2.3.0

.ci/docker/ubuntu-cuda/Dockerfile

@@ -56,7 +56,7 @@ RUN if [ -n "${DB}" ]; then bash ./install_db.sh; fi
 RUN rm install_db.sh
 ENV INSTALLED_DB ${DB}
 
-# (optional) Install vision packages like OpenCV
+# (optional) Install vision packages like OpenCV and ffmpeg
 ARG VISION
 COPY ./common/install_vision.sh ./common/cache_vision_models.sh ./common/common_utils.sh ./
 RUN if [ -n "${VISION}" ]; then bash ./install_vision.sh; fi
@@ -139,7 +139,7 @@ COPY --from=pytorch/llvm:9.0.1 /opt/llvm /opt/llvm
 ARG CUDNN_VERSION
 ARG CUDA_VERSION
 COPY ./common/install_cudnn.sh install_cudnn.sh
-RUN if [ -n "${CUDNN_VERSION}" ]; then bash install_cudnn.sh; fi
+RUN if [ "${CUDNN_VERSION}" -eq 8 ]; then bash install_cudnn.sh; fi
 RUN rm install_cudnn.sh
 
 # Install CUSPARSELT
@@ -152,7 +152,6 @@ RUN rm install_cusparselt.sh
 RUN if [ -h /usr/local/cuda-11.6/cuda-11.6 ]; then rm /usr/local/cuda-11.6/cuda-11.6; fi
 RUN if [ -h /usr/local/cuda-11.7/cuda-11.7 ]; then rm /usr/local/cuda-11.7/cuda-11.7; fi
 RUN if [ -h /usr/local/cuda-12.1/cuda-12.1 ]; then rm /usr/local/cuda-12.1/cuda-12.1; fi
-RUN if [ -h /usr/local/cuda-12.4/cuda-12.4 ]; then rm /usr/local/cuda-12.4/cuda-12.4; fi
 
 USER jenkins
 CMD ["bash"]

.ci/docker/ubuntu-rocm/Dockerfile

@@ -53,7 +53,7 @@ RUN if [ -n "${DB}" ]; then bash ./install_db.sh; fi
 RUN rm install_db.sh
 ENV INSTALLED_DB ${DB}
 
-# (optional) Install vision packages like OpenCV
+# (optional) Install vision packages like OpenCV and ffmpeg
 ARG VISION
 COPY ./common/install_vision.sh ./common/cache_vision_models.sh ./common/common_utils.sh ./
 RUN if [ -n "${VISION}" ]; then bash ./install_vision.sh; fi
@@ -78,11 +78,6 @@ ENV MAGMA_HOME /opt/rocm/magma
 ENV LANG C.UTF-8
 ENV LC_ALL C.UTF-8
 
-# Install amdsmi
-COPY ./common/install_amdsmi.sh install_amdsmi.sh
-RUN bash ./install_amdsmi.sh
-RUN rm install_amdsmi.sh
-
 # (optional) Install non-default CMake version
 ARG CMAKE_VERSION
 COPY ./common/install_cmake.sh install_cmake.sh
@@ -105,13 +100,6 @@ COPY triton_version.txt triton_version.txt
 RUN if [ -n "${TRITON}" ]; then bash ./install_triton.sh; fi
 RUN rm install_triton.sh common_utils.sh triton-rocm.txt triton_version.txt
 
-# Install AOTriton
-COPY ./aotriton_version.txt aotriton_version.txt
-COPY ./common/common_utils.sh common_utils.sh
-COPY ./common/install_aotriton.sh install_aotriton.sh
-RUN ["/bin/bash", "-c", "./install_aotriton.sh /opt/rocm && rm -rf install_aotriton.sh aotriton_version.txt common_utils.sh"]
-ENV AOTRITON_INSTALLED_PREFIX /opt/rocm/aotriton
-
 # Install ccache/sccache (do this last, so we get priority in PATH)
 COPY ./common/install_cache.sh install_cache.sh
 ENV PATH /opt/cache/bin:$PATH

.ci/docker/ubuntu-xpu/Dockerfile

@@ -62,7 +62,7 @@ RUN if [ -n "${INDUCTOR_BENCHMARKS}" ]; then bash ./install_inductor_benchmark_d
 RUN rm install_inductor_benchmark_deps.sh common_utils.sh timm.txt huggingface.txt
 
 # Install XPU Dependencies
-ARG XPU_VERSION
+ARG BASEKIT_VERSION
 COPY ./common/install_xpu.sh install_xpu.sh
 RUN bash ./install_xpu.sh && rm install_xpu.sh
 
@@ -83,7 +83,7 @@ RUN if [ -n "${DB}" ]; then bash ./install_db.sh; fi
 RUN rm install_db.sh
 ENV INSTALLED_DB ${DB}
 
-# (optional) Install vision packages like OpenCV
+# (optional) Install vision packages like OpenCV and ffmpeg
 ARG VISION
 COPY ./common/install_vision.sh ./common/cache_vision_models.sh ./common/common_utils.sh ./
 RUN if [ -n "${VISION}" ]; then bash ./install_vision.sh; fi

.ci/docker/ubuntu/Dockerfile

@@ -80,7 +80,7 @@ RUN if [ -n "${DB}" ]; then bash ./install_db.sh; fi
 RUN rm install_db.sh
 ENV INSTALLED_DB ${DB}
 
-# (optional) Install vision packages like OpenCV
+# (optional) Install vision packages like OpenCV and ffmpeg
 ARG VISION
 COPY ./common/install_vision.sh ./common/cache_vision_models.sh ./common/common_utils.sh ./
 RUN if [ -n "${VISION}" ]; then bash ./install_vision.sh; fi
@@ -169,11 +169,9 @@ RUN rm install_acl.sh
 ENV INSTALLED_ACL ${ACL}
 
 # Install ccache/sccache (do this last, so we get priority in PATH)
-ARG SKIP_SCCACHE_INSTALL
 COPY ./common/install_cache.sh install_cache.sh
 ENV PATH /opt/cache/bin:$PATH
-RUN if [ -z "${SKIP_SCCACHE_INSTALL}" ]; then bash ./install_cache.sh; fi
-RUN rm install_cache.sh
+RUN bash ./install_cache.sh && rm install_cache.sh
 
 # Add jni.h for java host build
 COPY ./common/install_jni.sh install_jni.sh
@@ -190,9 +188,7 @@ ARG BUILD_ENVIRONMENT
 ENV BUILD_ENVIRONMENT ${BUILD_ENVIRONMENT}
 
 # Install LLVM dev version (Defined in the pytorch/builder github repository)
-ARG SKIP_LLVM_SRC_BUILD_INSTALL
 COPY --from=pytorch/llvm:9.0.1 /opt/llvm /opt/llvm
-RUN if [ -n "${SKIP_LLVM_SRC_BUILD_INSTALL}" ]; then set -eu; rm -rf /opt/llvm; fi
 
 # AWS specific CUDA build guidance
 ENV TORCH_CUDA_ARCH_LIST Maxwell

.ci/pytorch/build.sh

@@ -44,7 +44,15 @@ if [[ "$BUILD_ENVIRONMENT" == *cuda11* ]]; then
   fi
 fi
 
-if [[ ${BUILD_ENVIRONMENT} == *"parallelnative"* ]]; then
+if [[ ${BUILD_ENVIRONMENT} == *"caffe2"* ]]; then
+  echo "Caffe2 build is ON"
+  export BUILD_CAFFE2=ON
+fi
+
+if [[ ${BUILD_ENVIRONMENT} == *"paralleltbb"* ]]; then
+  export ATEN_THREADING=TBB
+  export USE_TBB=1
+elif [[ ${BUILD_ENVIRONMENT} == *"parallelnative"* ]]; then
   export ATEN_THREADING=NATIVE
 fi
 
@@ -73,22 +81,7 @@ if ! which conda; then
     export USE_MKLDNN=0
   fi
 else
-  # CMAKE_PREFIX_PATH precedences
-  # 1. $CONDA_PREFIX, if defined. This follows the pytorch official build instructions.
-  # 2. /opt/conda/envs/py_${ANACONDA_PYTHON_VERSION}, if ANACONDA_PYTHON_VERSION defined.
-  #    This is for CI, which defines ANACONDA_PYTHON_VERSION but not CONDA_PREFIX.
-  # 3. $(conda info --base). The fallback value of pytorch official build
-  #    instructions actually refers to this.
-  #    Commonly this is /opt/conda/
-  if [[ -v CONDA_PREFIX ]]; then
-    export CMAKE_PREFIX_PATH=${CONDA_PREFIX}
-  elif [[ -v ANACONDA_PYTHON_VERSION ]]; then
-    export CMAKE_PREFIX_PATH="/opt/conda/envs/py_${ANACONDA_PYTHON_VERSION}"
-  else
-    # already checked by `! which conda`
-    CMAKE_PREFIX_PATH="$(conda info --base)"
-    export CMAKE_PREFIX_PATH
-  fi
+  export CMAKE_PREFIX_PATH=/opt/conda
 
   # Workaround required for MKL library linkage
   # https://github.com/pytorch/pytorch/issues/119557
@@ -286,9 +279,6 @@ else
       fi
       WERROR=1 python setup.py bdist_wheel
     else
-      if [[ "$BUILD_ENVIRONMENT" == *xla* ]]; then
-        source .ci/pytorch/install_cache_xla.sh
-      fi
       python setup.py bdist_wheel
     fi
     pip_install_whl "$(echo dist/*.whl)"
@@ -330,7 +320,7 @@ else
     SITE_PACKAGES="$(python -c 'from distutils.sysconfig import get_python_lib; print(get_python_lib())')"
     mkdir -p "$CUSTOM_OP_BUILD"
     pushd "$CUSTOM_OP_BUILD"
-    cmake "$CUSTOM_OP_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES/torch" -DPython_EXECUTABLE="$(which python)" \
+    cmake "$CUSTOM_OP_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES/torch" -DPYTHON_EXECUTABLE="$(which python)" \
           -DCMAKE_MODULE_PATH="$CUSTOM_TEST_MODULE_PATH" -DUSE_ROCM="$CUSTOM_TEST_USE_ROCM"
     make VERBOSE=1
     popd
@@ -343,7 +333,7 @@ else
     SITE_PACKAGES="$(python -c 'from distutils.sysconfig import get_python_lib; print(get_python_lib())')"
     mkdir -p "$JIT_HOOK_BUILD"
     pushd "$JIT_HOOK_BUILD"
-    cmake "$JIT_HOOK_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES/torch" -DPython_EXECUTABLE="$(which python)" \
+    cmake "$JIT_HOOK_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES/torch" -DPYTHON_EXECUTABLE="$(which python)" \
           -DCMAKE_MODULE_PATH="$CUSTOM_TEST_MODULE_PATH" -DUSE_ROCM="$CUSTOM_TEST_USE_ROCM"
     make VERBOSE=1
     popd
@@ -355,7 +345,7 @@ else
     python --version
     mkdir -p "$CUSTOM_BACKEND_BUILD"
     pushd "$CUSTOM_BACKEND_BUILD"
-    cmake "$CUSTOM_BACKEND_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES/torch" -DPython_EXECUTABLE="$(which python)" \
+    cmake "$CUSTOM_BACKEND_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES/torch" -DPYTHON_EXECUTABLE="$(which python)" \
           -DCMAKE_MODULE_PATH="$CUSTOM_TEST_MODULE_PATH" -DUSE_ROCM="$CUSTOM_TEST_USE_ROCM"
     make VERBOSE=1
     popd
@@ -386,8 +376,4 @@ if [[ "$BUILD_ENVIRONMENT" != *libtorch* && "$BUILD_ENVIRONMENT" != *bazel* ]];
   python tools/stats/export_test_times.py
 fi
 
-# snadampal: skipping it till sccache support added for aarch64
-# https://github.com/pytorch/pytorch/issues/121559
-if [[ "$BUILD_ENVIRONMENT" != *aarch64* ]]; then
-  print_sccache_stats
-fi
+print_sccache_stats

.ci/pytorch/docs-test.sh

@@ -6,4 +6,4 @@ source "$(dirname "${BASH_SOURCE[0]}")/common.sh"
 echo "Testing pytorch docs"
 
 cd docs
-TERM=vt100 make doctest
+make doctest

.ci/pytorch/install_cache_xla.sh

@@ -1,37 +0,0 @@
-#!/bin/bash
-
-# Script for installing sccache on the xla build job, which uses xla's docker
-# image and doesn't have sccache installed on it.  This is mostly copied from
-# .ci/docker/install_cache.sh.  Changes are: removing checks that will always
-# return the same thing, ex checks for for rocm, CUDA, and changing the path
-# where sccache is installed, and not changing /etc/environment.
-
-set -ex
-
-install_binary() {
-  echo "Downloading sccache binary from S3 repo"
-  curl --retry 3 https://s3.amazonaws.com/ossci-linux/sccache -o /tmp/cache/bin/sccache
-}
-
-mkdir -p /tmp/cache/bin
-mkdir -p /tmp/cache/lib
-export PATH="/tmp/cache/bin:$PATH"
-
-install_binary
-chmod a+x /tmp/cache/bin/sccache
-
-function write_sccache_stub() {
-  # Unset LD_PRELOAD for ps because of asan + ps issues
-  # https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90589
-  # shellcheck disable=SC2086
-  # shellcheck disable=SC2059
-  printf "#!/bin/sh\nif [ \$(env -u LD_PRELOAD ps -p \$PPID -o comm=) != sccache ]; then\n  exec sccache $(which $1) \"\$@\"\nelse\n  exec $(which $1) \"\$@\"\nfi" > "/tmp/cache/bin/$1"
-  chmod a+x "/tmp/cache/bin/$1"
-}
-
-write_sccache_stub cc
-write_sccache_stub c++
-write_sccache_stub gcc
-write_sccache_stub g++
-write_sccache_stub clang
-write_sccache_stub clang++

.ci/pytorch/multigpu-test.sh

@@ -18,7 +18,6 @@ time python test/run_test.py --verbose -i distributed/test_c10d_gloo
 time python test/run_test.py --verbose -i distributed/test_c10d_nccl
 time python test/run_test.py --verbose -i distributed/test_c10d_spawn_gloo
 time python test/run_test.py --verbose -i distributed/test_c10d_spawn_nccl
-time python test/run_test.py --verbose -i distributed/test_cuda_p2p
 time python test/run_test.py --verbose -i distributed/test_store
 time python test/run_test.py --verbose -i distributed/test_pg_wrapper
 time python test/run_test.py --verbose -i distributed/rpc/cuda/test_tensorpipe_agent
@@ -46,13 +45,7 @@ time python test/run_test.py --verbose -i distributed/test_device_mesh
 time python test/run_test.py --verbose -i distributed/tensor/parallel/test_ddp_2d_parallel
 time python test/run_test.py --verbose -i distributed/tensor/parallel/test_fsdp_2d_parallel
 time python test/run_test.py --verbose -i distributed/tensor/parallel/test_tp_examples
-time python test/run_test.py --verbose -i distributed/tensor/parallel/test_tp_random_state
-
-# FSDP2 tests
-time python test/run_test.py --verbose -i distributed/_composable/fsdp/test_fully_shard_training -- -k test_2d_mlp_with_nd_mesh
-
-# Pipelining composability tests
-time python test/run_test.py --verbose -i distributed/pipelining/test_composability.py
+time python test/run_test.py --verbose -i distributed/tensor/parallel/test_tp_random_state.py
 
 # Other tests
 time python test/run_test.py --verbose -i test_cuda_primary_ctx

.ci/pytorch/test.sh

@@ -181,11 +181,6 @@ if [[ "$BUILD_ENVIRONMENT" != *-bazel-* ]] ; then
   export PATH="$HOME/.local/bin:$PATH"
 fi
 
-if [[ "$BUILD_ENVIRONMENT" == *aarch64* ]]; then
-  # TODO: revisit this once the CI is stabilized on aarch64 linux
-  export VALGRIND=OFF
-fi
-
 install_tlparse
 
 # DANGER WILL ROBINSON.  The LD_PRELOAD here could cause you problems
@@ -264,18 +259,6 @@ elif [[ $TEST_CONFIG == 'nogpu_AVX512' ]]; then
   export ATEN_CPU_CAPABILITY=avx2
 fi
 
-# temp workarounds for https://github.com/pytorch/pytorch/issues/126692, remove when fixed
-if [[ "$BUILD_ENVIRONMENT" != *-bazel-* ]]; then
-  pushd test
-  CUDA_VERSION=$(python -c "import torch; print(torch.version.cuda)")
-  if [ "$CUDA_VERSION" == "12.4" ]; then
-    ISCUDA124="cu124"
-  else
-    ISCUDA124=""
-  fi
-  popd
-fi
-
 test_python_legacy_jit() {
   time python test/run_test.py --include test_jit_legacy test_jit_fuser_legacy --verbose
   assert_git_not_dirty
@@ -322,24 +305,22 @@ test_dynamo_shard() {
 test_inductor_distributed() {
   # Smuggle a few multi-gpu tests here so that we don't have to request another large node
   echo "Testing multi_gpu tests in test_torchinductor"
-  python test/run_test.py -i inductor/test_torchinductor.py -k test_multi_gpu --verbose
-  python test/run_test.py -i inductor/test_aot_inductor.py -k test_non_default_cuda_device --verbose
-  python test/run_test.py -i inductor/test_aot_inductor.py -k test_replicate_on_devices --verbose
-  python test/run_test.py -i distributed/test_c10d_functional_native.py --verbose
-  python test/run_test.py -i distributed/_tensor/test_dtensor_compile.py --verbose
-  python test/run_test.py -i distributed/tensor/parallel/test_fsdp_2d_parallel.py --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_comm.py --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_multi_group --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_with_activation_checkpointing --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_2d_mlp --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_hsdp --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_2d_transformer_checkpoint_resume --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_training.py -k test_gradient_accumulation --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_frozen.py --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_mixed_precision.py -k test_compute_dtype --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_mixed_precision.py -k test_reduce_dtype --verbose
-  python test/run_test.py -i distributed/_composable/fsdp/test_fully_shard_clip_grad_norm_.py -k test_clip_grad_norm_2d --verbose
-  python test/run_test.py -i distributed/fsdp/test_fsdp_tp_integration.py -k test_fsdp_tp_integration --verbose
+  pytest test/inductor/test_torchinductor.py -k test_multi_gpu
+  pytest test/inductor/test_aot_inductor.py -k test_non_default_cuda_device
+  pytest test/inductor/test_aot_inductor.py -k test_replicate_on_devices
+  pytest test/distributed/test_c10d_functional_native.py
+  pytest test/distributed/_tensor/test_dtensor_compile.py
+  pytest test/distributed/tensor/parallel/test_fsdp_2d_parallel.py
+  pytest test/distributed/_composable/fsdp/test_fully_shard_comm.py
+  pytest test/distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_multi_group
+  pytest test/distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_with_activation_checkpointing
+  pytest test/distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_2d_mlp
+  pytest test/distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_hsdp
+  pytest test/distributed/_composable/fsdp/test_fully_shard_training.py -k test_train_parity_2d_transformer_checkpoint_resume
+  pytest test/distributed/_composable/fsdp/test_fully_shard_frozen.py
+  pytest test/distributed/_composable/fsdp/test_fully_shard_mixed_precision.py -k test_compute_dtype
+  pytest test/distributed/_composable/fsdp/test_fully_shard_mixed_precision.py -k test_reduce_dtype
+  pytest test/distributed/fsdp/test_fsdp_tp_integration.py -k test_fsdp_tp_integration
 
   # this runs on both single-gpu and multi-gpu instance. It should be smart about skipping tests that aren't supported
   # with if required # gpus aren't available
@@ -363,20 +344,10 @@ test_inductor() {
 
 test_inductor_cpp_wrapper_abi_compatible() {
   export TORCHINDUCTOR_ABI_COMPATIBLE=1
-  TEST_REPORTS_DIR=$(pwd)/test/test-reports
-  mkdir -p "$TEST_REPORTS_DIR"
-
   echo "Testing Inductor cpp wrapper mode with TORCHINDUCTOR_ABI_COMPATIBLE=1"
   # cpu stack allocation causes segfault and needs more investigation
-  PYTORCH_TESTING_DEVICE_ONLY_FOR="" python test/run_test.py --include inductor/test_cpu_cpp_wrapper
+  TORCHINDUCTOR_STACK_ALLOCATION=0 python test/run_test.py --include inductor/test_cpu_cpp_wrapper
   python test/run_test.py --include inductor/test_cuda_cpp_wrapper
-
-  TORCHINDUCTOR_CPP_WRAPPER=1 python benchmarks/dynamo/timm_models.py --device cuda --accuracy --amp \
-    --training --inductor --disable-cudagraphs --only vit_base_patch16_224 \
-    --output "$TEST_REPORTS_DIR/inductor_cpp_wrapper_training.csv"
-  python benchmarks/dynamo/check_accuracy.py \
-    --actual "$TEST_REPORTS_DIR/inductor_cpp_wrapper_training.csv" \
-    --expected "benchmarks/dynamo/ci_expected_accuracy/${ISCUDA124}/inductor_timm_training.csv"
 }
 
 # "Global" flags for inductor benchmarking controlled by TEST_CONFIG
@@ -538,18 +509,13 @@ test_single_dynamo_benchmark() {
       --output "$TEST_REPORTS_DIR/${name}_${suite}.csv"
     python benchmarks/dynamo/check_accuracy.py \
       --actual "$TEST_REPORTS_DIR/${name}_$suite.csv" \
-      --expected "benchmarks/dynamo/ci_expected_accuracy/${ISCUDA124}/${TEST_CONFIG}_${name}.csv"
+      --expected "benchmarks/dynamo/ci_expected_accuracy/${TEST_CONFIG}_${name}.csv"
     python benchmarks/dynamo/check_graph_breaks.py \
       --actual "$TEST_REPORTS_DIR/${name}_$suite.csv" \
-      --expected "benchmarks/dynamo/ci_expected_accuracy/${ISCUDA124}/${TEST_CONFIG}_${name}.csv"
+      --expected "benchmarks/dynamo/ci_expected_accuracy/${TEST_CONFIG}_${name}.csv"
   fi
 }
 
-test_inductor_micro_benchmark() {
-  TEST_REPORTS_DIR=$(pwd)/test/test-reports
-  python benchmarks/gpt_fast/benchmark.py --output "${TEST_REPORTS_DIR}/gpt_fast_benchmark.csv"
-}
-
 test_dynamo_benchmark() {
   # Usage: test_dynamo_benchmark huggingface 0
   TEST_REPORTS_DIR=$(pwd)/test/test-reports
@@ -565,11 +531,7 @@ test_dynamo_benchmark() {
     test_single_dynamo_benchmark "dashboard" "$suite" "$shard_id" "$@"
   else
     if [[ "${TEST_CONFIG}" == *cpu_inductor* ]]; then
-      if [[ "${TEST_CONFIG}" == *freezing* ]]; then
-        test_single_dynamo_benchmark "inference" "$suite" "$shard_id" --inference --float32 --freezing "$@"
-      else
-        test_single_dynamo_benchmark "inference" "$suite" "$shard_id" --inference --float32 "$@"
-      fi
+      test_single_dynamo_benchmark "inference" "$suite" "$shard_id" --inference --float32 "$@"
     elif [[ "${TEST_CONFIG}" == *aot_inductor* ]]; then
       test_single_dynamo_benchmark "inference" "$suite" "$shard_id" --inference --bfloat16 "$@"
     else
@@ -583,16 +545,12 @@ test_inductor_torchbench_smoketest_perf() {
   TEST_REPORTS_DIR=$(pwd)/test/test-reports
   mkdir -p "$TEST_REPORTS_DIR"
 
-  # Test some models in the cpp wrapper mode
-  TORCHINDUCTOR_ABI_COMPATIBLE=1 TORCHINDUCTOR_CPP_WRAPPER=1 python benchmarks/dynamo/torchbench.py --device cuda --accuracy \
-    --bfloat16 --inference --inductor --only hf_T5 --output "$TEST_REPORTS_DIR/inductor_cpp_wrapper_inference.csv"
-  TORCHINDUCTOR_ABI_COMPATIBLE=1 TORCHINDUCTOR_CPP_WRAPPER=1 python benchmarks/dynamo/torchbench.py --device cuda --accuracy \
-    --bfloat16 --inference --inductor --only llama --output "$TEST_REPORTS_DIR/inductor_cpp_wrapper_inference.csv"
-  TORCHINDUCTOR_ABI_COMPATIBLE=1 TORCHINDUCTOR_CPP_WRAPPER=1 python benchmarks/dynamo/torchbench.py --device cuda --accuracy \
-    --bfloat16 --inference --inductor --only moco --output "$TEST_REPORTS_DIR/inductor_cpp_wrapper_inference.csv"
+  # smoke test the cpp_wrapper mode
+  TORCHINDUCTOR_CPP_WRAPPER=1 python benchmarks/dynamo/torchbench.py --device cuda --accuracy --bfloat16 \
+    --inference --inductor --only hf_T5 --output "$TEST_REPORTS_DIR/inductor_cpp_wrapper_smoketest.csv"
   python benchmarks/dynamo/check_accuracy.py \
-    --actual "$TEST_REPORTS_DIR/inductor_cpp_wrapper_inference.csv" \
-    --expected "benchmarks/dynamo/ci_expected_accuracy/${ISCUDA124}/inductor_torchbench_inference.csv"
+      --actual "$TEST_REPORTS_DIR/inductor_cpp_wrapper_smoketest.csv" \
+      --expected "benchmarks/dynamo/ci_expected_accuracy/inductor_torchbench_inference.csv"
 
   python benchmarks/dynamo/torchbench.py --device cuda --performance --backend inductor --float16 --training \
     --batch-size-file "$(realpath benchmarks/dynamo/torchbench_models_list.txt)" --only hf_Bert \
@@ -607,13 +565,7 @@ test_inductor_torchbench_smoketest_perf() {
   # https://github.com/pytorch/pytorch/actions/runs/7158691360/job/19491437314,
   # and thus we lower its threshold to reduce flakiness. If this continues to be a problem,
   # we switch to use some other model.
-  # Use 4.7 for cuda 12.4, change back to 4.9 after fixing https://github.com/pytorch/pytorch/issues/126692
-  if [ "$CUDA_VERSION" == "12.4" ]; then
-    THRESHOLD=4.7
-  else
-    THRESHOLD=4.9
-  fi
-  python benchmarks/dynamo/check_perf_csv.py -f "$TEST_REPORTS_DIR/inductor_inference_smoketest.csv" -t $THRESHOLD
+  python benchmarks/dynamo/check_perf_csv.py -f "$TEST_REPORTS_DIR/inductor_inference_smoketest.csv" -t 4.9
 
   # Check memory compression ratio for a few models
   for test in hf_Albert timm_vision_transformer; do
@@ -625,15 +577,6 @@ test_inductor_torchbench_smoketest_perf() {
       "$TEST_REPORTS_DIR/inductor_training_smoketest_$test.csv" \
       --expected benchmarks/dynamo/expected_ci_perf_inductor_torchbench.csv
   done
-
-  # Perform some "warm-start" runs for a few huggingface models.
-  for test in AlbertForQuestionAnswering AllenaiLongformerBase DistilBertForMaskedLM DistillGPT2 GoogleFnet YituTechConvBert; do
-    python benchmarks/dynamo/huggingface.py --accuracy --training --amp --inductor --device cuda --warm-start-latency \
-      --only $test --output "$TEST_REPORTS_DIR/inductor_warm_start_smoketest_$test.csv"
-    python benchmarks/dynamo/check_accuracy.py \
-      --actual "$TEST_REPORTS_DIR/inductor_warm_start_smoketest_$test.csv" \
-      --expected "benchmarks/dynamo/ci_expected_accuracy/${ISCUDA124}/inductor_huggingface_training.csv"
-  done
 }
 
 test_inductor_torchbench_cpu_smoketest_perf(){
@@ -717,6 +660,7 @@ test_aten() {
   ${SUDO} ln -sf "$TORCH_LIB_DIR"/libmkldnn* "$TEST_BASE_DIR"
   ${SUDO} ln -sf "$TORCH_LIB_DIR"/libnccl* "$TEST_BASE_DIR"
   ${SUDO} ln -sf "$TORCH_LIB_DIR"/libtorch* "$TEST_BASE_DIR"
+  ${SUDO} ln -sf "$TORCH_LIB_DIR"/libtbb* "$TEST_BASE_DIR"
 
   ls "$TEST_BASE_DIR"
   aten/tools/run_tests.sh "$TEST_BASE_DIR"
@@ -741,6 +685,21 @@ test_without_numpy() {
   popd
 }
 
+# pytorch extensions require including torch/extension.h which includes all.h
+# which includes utils.h which includes Parallel.h.
+# So you can call for instance parallel_for() from your extension,
+# but the compilation will fail because of Parallel.h has only declarations
+# and definitions are conditionally included Parallel.h(see last lines of Parallel.h).
+# I tried to solve it #39612 and #39881 by including Config.h into Parallel.h
+# But if Pytorch is built with TBB it provides Config.h
+# that has AT_PARALLEL_NATIVE_TBB=1(see #3961 or #39881) and it means that if you include
+# torch/extension.h which transitively includes Parallel.h
+# which transitively includes tbb.h which is not available!
+if [[ "${BUILD_ENVIRONMENT}" == *tbb* ]]; then
+  sudo mkdir -p /usr/include/tbb
+  sudo cp -r "$PWD"/third_party/tbb/include/tbb/* /usr/include/tbb
+fi
+
 test_libtorch() {
   local SHARD="$1"
 
@@ -754,6 +713,7 @@ test_libtorch() {
     ln -sf "$TORCH_LIB_DIR"/libc10* "$TORCH_BIN_DIR"
     ln -sf "$TORCH_LIB_DIR"/libshm* "$TORCH_BIN_DIR"
     ln -sf "$TORCH_LIB_DIR"/libtorch* "$TORCH_BIN_DIR"
+    ln -sf "$TORCH_LIB_DIR"/libtbb* "$TORCH_BIN_DIR"
     ln -sf "$TORCH_LIB_DIR"/libnvfuser* "$TORCH_BIN_DIR"
 
     export CPP_TESTS_DIR="${TORCH_BIN_DIR}"
@@ -890,6 +850,7 @@ test_rpc() {
   # test reporting process to function as expected.
   ln -sf "$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
 }
@@ -1191,33 +1152,11 @@ test_executorch() {
   assert_git_not_dirty
 }
 
-test_linux_aarch64(){
-  python test/run_test.py --include test_modules test_mkldnn test_mkldnn_fusion test_openmp test_torch test_dynamic_shapes \
-       test_transformers test_multiprocessing test_numpy_interop --verbose
-
-  # Dynamo tests
-  python test/run_test.py --include dynamo/test_compile dynamo/test_backends dynamo/test_comptime dynamo/test_config \
-       dynamo/test_functions dynamo/test_fx_passes_pre_grad dynamo/test_interop dynamo/test_model_output dynamo/test_modules \
-       dynamo/test_optimizers dynamo/test_recompile_ux dynamo/test_recompiles --verbose
-
-  # Inductor tests
-  python test/run_test.py --include inductor/test_torchinductor inductor/test_benchmark_fusion inductor/test_codecache \
-       inductor/test_config inductor/test_control_flow inductor/test_coordinate_descent_tuner inductor/test_fx_fusion \
-       inductor/test_group_batch_fusion inductor/test_inductor_freezing inductor/test_inductor_utils \
-       inductor/test_inplacing_pass inductor/test_kernel_benchmark inductor/test_layout_optim \
-       inductor/test_max_autotune inductor/test_memory_planning inductor/test_metrics inductor/test_multi_kernel inductor/test_pad_mm \
-       inductor/test_pattern_matcher inductor/test_perf inductor/test_profiler inductor/test_select_algorithm inductor/test_smoke \
-       inductor/test_split_cat_fx_passes inductor/test_standalone_compile inductor/test_torchinductor \
-       inductor/test_torchinductor_codegen_dynamic_shapes inductor/test_torchinductor_dynamic_shapes --verbose
-}
-
 if ! [[ "${BUILD_ENVIRONMENT}" == *libtorch* || "${BUILD_ENVIRONMENT}" == *-bazel-* ]]; then
   (cd test && python -c "import torch; print(torch.__config__.show())")
   (cd test && python -c "import torch; print(torch.__config__.parallel_info())")
 fi
-if [[ "$BUILD_ENVIRONMENT" == *aarch64* ]]; then
-  test_linux_aarch64
-elif [[ "${TEST_CONFIG}" == *backward* ]]; then
+if [[ "${TEST_CONFIG}" == *backward* ]]; then
   test_forward_backward_compatibility
   # Do NOT add tests after bc check tests, see its comment.
 elif [[ "${TEST_CONFIG}" == *xla* ]]; then
@@ -1242,8 +1181,6 @@ elif [[ "$TEST_CONFIG" == deploy ]]; then
   test_torch_deploy
 elif [[ "${TEST_CONFIG}" == *inductor_distributed* ]]; then
   test_inductor_distributed
-elif [[ "${TEST_CONFIG}" == *inductor-micro-benchmark* ]]; then
-  test_inductor_micro_benchmark
 elif [[ "${TEST_CONFIG}" == *huggingface* ]]; then
   install_torchvision
   id=$((SHARD_NUMBER-1))
@@ -1297,10 +1234,6 @@ elif [[ "${TEST_CONFIG}" == *dynamo* && "${SHARD_NUMBER}" == 1 && $NUM_TEST_SHAR
 elif [[ "${TEST_CONFIG}" == *dynamo* && $SHARD_NUMBER -gt 1 && $NUM_TEST_SHARDS -gt 1 ]]; then
   install_torchvision
   test_dynamo_shard "${SHARD_NUMBER}"
-elif [[ "${BUILD_ENVIRONMENT}" == *rocm* && -n "$TESTS_TO_INCLUDE" ]]; then
-  install_torchvision
-  test_python_shard "$SHARD_NUMBER"
-  test_aten
 elif [[ "${SHARD_NUMBER}" == 1 && $NUM_TEST_SHARDS -gt 1 ]]; then
   test_without_numpy
   install_torchvision
@@ -1330,6 +1263,10 @@ elif [[ "${BUILD_ENVIRONMENT}" == *-mobile-lightweight-dispatch* ]]; then
   test_libtorch
 elif [[ "${TEST_CONFIG}" = docs_test ]]; then
   test_docs_test
+elif [[ "${BUILD_ENVIRONMENT}" == *rocm* && -n "$TESTS_TO_INCLUDE" ]]; then
+  install_torchvision
+  test_python
+  test_aten
 elif [[ "${BUILD_ENVIRONMENT}" == *xpu* ]]; then
   install_torchvision
   test_python

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

@@ -17,22 +17,22 @@ set PATH=C:\Program Files\CMake\bin;C:\Program Files\7-Zip;C:\ProgramData\chocol
 set INSTALLER_DIR=%SCRIPT_HELPERS_DIR%\installation-helpers
 
 call %INSTALLER_DIR%\install_magma.bat
-if errorlevel 1 goto fail
-if not errorlevel 0 goto fail
+if errorlevel 1 exit /b
+if not errorlevel 0 exit /b
 
 call %INSTALLER_DIR%\install_sccache.bat
-if errorlevel 1 goto fail
-if not errorlevel 0 goto fail
+if errorlevel 1 exit /b
+if not errorlevel 0 exit /b
 
 :: Miniconda has been installed as part of the Windows AMI with all the dependencies.
 :: We just need to activate it here
 call %INSTALLER_DIR%\activate_miniconda3.bat
-if errorlevel 1 goto fail
-if not errorlevel 0 goto fail
+if errorlevel 1 exit /b
+if not errorlevel 0 exit /b
 
 call pip install mkl-include==2021.4.0 mkl-devel==2021.4.0
-if errorlevel 1 goto fail
-if not errorlevel 0 goto fail
+if errorlevel 1 exit /b
+if not errorlevel 0 exit /b
 
 :: Override VS env here
 pushd .
@@ -41,8 +41,8 @@ if "%VC_VERSION%" == "" (
 ) else (
     call "C:\Program Files (x86)\Microsoft Visual Studio\%VC_YEAR%\%VC_PRODUCT%\VC\Auxiliary\Build\vcvarsall.bat" x64 -vcvars_ver=%VC_VERSION%
 )
-if errorlevel 1 goto fail
-if not errorlevel 0 goto fail
+if errorlevel 1 exit /b
+if not errorlevel 0 exit /b
 @echo on
 popd
 
@@ -52,12 +52,12 @@ set CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v%CUDA_VERSION%
 
 if x%CUDA_VERSION:.=%==x%CUDA_VERSION% (
     echo CUDA version %CUDA_VERSION% format isn't correct, which doesn't contain '.'
-    goto fail
+    exit /b 1
 )
 rem version transformer, for example 10.1 to 10_1.
 if x%CUDA_VERSION:.=%==x%CUDA_VERSION% (
     echo CUDA version %CUDA_VERSION% format isn't correct, which doesn't contain '.'
-    goto fail
+    exit /b 1
 )
 set VERSION_SUFFIX=%CUDA_VERSION:.=_%
 set CUDA_PATH_V%VERSION_SUFFIX%=%CUDA_PATH%
@@ -101,8 +101,8 @@ if "%USE_CUDA%"=="1" (
   :: CMake requires a single command as CUDA_NVCC_EXECUTABLE, so we push the wrappers
   :: randomtemp.exe and sccache.exe into a batch file which CMake invokes.
   curl -kL https://github.com/peterjc123/randomtemp-rust/releases/download/v0.4/randomtemp.exe --output %TMP_DIR_WIN%\bin\randomtemp.exe
-  if errorlevel 1 goto fail
-  if not errorlevel 0 goto fail
+  if errorlevel 1 exit /b
+  if not errorlevel 0 exit /b
   echo @"%TMP_DIR_WIN%\bin\randomtemp.exe" "%TMP_DIR_WIN%\bin\sccache.exe" "%CUDA_PATH%\bin\nvcc.exe" %%* > "%TMP_DIR%/bin/nvcc.bat"
   cat %TMP_DIR%/bin/nvcc.bat
   set CUDA_NVCC_EXECUTABLE=%TMP_DIR%/bin/nvcc.bat
@@ -114,8 +114,8 @@ if "%USE_CUDA%"=="1" (
 set
 
 python setup.py bdist_wheel
-if errorlevel 1 goto fail
-if not errorlevel 0 goto fail
+if errorlevel 1 exit /b
+if not errorlevel 0 exit /b
 sccache --show-stats
 python -c "import os, glob; os.system('python -mpip install --no-index --no-deps ' + glob.glob('dist/*.whl')[0])"
 (
@@ -135,8 +135,3 @@ python -c "import os, glob; os.system('python -mpip install --no-index --no-deps
 
 sccache --show-stats --stats-format json | jq .stats > sccache-stats-%BUILD_ENVIRONMENT%-%OUR_GITHUB_JOB_ID%.json
 sccache --stop-server
-
-exit /b 0
-
-:fail
-exit /b 1

.circleci/scripts/binary_linux_test.sh

@@ -96,13 +96,8 @@ if [[ "$PACKAGE_TYPE" == conda ]]; then
     conda install \${EXTRA_CONDA_FLAGS} -y "\$pkg" --offline
   )
 elif [[ "$PACKAGE_TYPE" != libtorch ]]; then
-  if [[ "\$BUILD_ENVIRONMENT" != *s390x* ]]; then
-    pip install "\$pkg" --index-url "https://download.pytorch.org/whl/\${CHANNEL}/${DESIRED_CUDA}"
-    retry pip install -q numpy protobuf typing-extensions
-  else
-    pip install "\$pkg"
-    retry pip install -q numpy protobuf typing-extensions
-  fi
+  pip install "\$pkg" --index-url "https://download.pytorch.org/whl/\${CHANNEL}/${DESIRED_CUDA}"
+  retry pip install -q numpy protobuf typing-extensions
 fi
 if [[ "$PACKAGE_TYPE" == libtorch ]]; then
   pkg="\$(ls /final_pkgs/*-latest.zip)"

.circleci/scripts/binary_populate_env.sh

@@ -76,8 +76,8 @@ TRITON_VERSION=$(cat $PYTORCH_ROOT/.ci/docker/triton_version.txt)
 
 # Here PYTORCH_EXTRA_INSTALL_REQUIREMENTS is already set for the all the wheel builds hence append TRITON_CONSTRAINT
 if [[ "$PACKAGE_TYPE" =~ .*wheel.* &&  -n "${PYTORCH_EXTRA_INSTALL_REQUIREMENTS:-}" ]]; then
-  # Only linux Python < 3.13 are supported wheels for triton
-  TRITON_CONSTRAINT="platform_system == 'Linux' and platform_machine == 'x86_64' and python_version < '3.13'"
+  # Only linux Python < 3.12 are supported wheels for triton
+  TRITON_CONSTRAINT="platform_system == 'Linux' and platform_machine == 'x86_64' and python_version < '3.12'"
   TRITON_REQUIREMENT="triton==${TRITON_VERSION}; ${TRITON_CONSTRAINT}"
   if [[ -n "$PYTORCH_BUILD_VERSION" && "$PYTORCH_BUILD_VERSION" =~ .*dev.* ]]; then
       TRITON_SHORTHASH=$(cut -c1-10 $PYTORCH_ROOT/.ci/docker/ci_commit_pins/triton.txt)

.clang-tidy

@@ -61,7 +61,6 @@ readability-simplify-subscript-expr,
 readability-string-compare,
 '
 HeaderFilterRegex: '^(aten/|c10/|torch/).*$'
+AnalyzeTemporaryDtors: false
 WarningsAsErrors: '*'
-CheckOptions:
-  misc-header-include-cycle.IgnoredFilesList: 'format.h;ivalue.h;custom_class.h;Dict.h;List.h'
 ...

.flake8

@@ -54,7 +54,6 @@ per-file-ignores =
     torch/ao/quantization/fx/_decomposed.py: TOR901
     torch/distributed/_functional_collectives.py: TOR901
     torch/distributed/_spmd/data_parallel.py: TOR901
-    torch/distributed/_tensor/_collective_utils.py: TOR901
 optional-ascii-coding = True
 exclude =
     ./.git,

.github/actionlint.yaml

@@ -1,12 +1,9 @@
 self-hosted-runner:
   labels:
-    # GitHub hosted x86 Linux runners
     - linux.20_04.4x
     - linux.20_04.16x
-    # Repo-specific LF hosted ARC runners
-    - linux.large.arc
-    # Organization-wide AWS Linux Runners
     - linux.large
+    - linux.large.arc
     - linux.2xlarge
     - linux.4xlarge
     - linux.12xlarge
@@ -16,34 +13,16 @@ self-hosted-runner:
     - linux.8xlarge.nvidia.gpu
     - linux.16xlarge.nvidia.gpu
     - linux.g5.4xlarge.nvidia.gpu
-    # Organization-wide AWS Linux Runners on Linux Foundation account
-    - lf.linux.large
-    - lf.linux.2xlarge
-    - lf.linux.4xlarge
-    - lf.linux.12xlarge
-    - lf.linux.24xlarge
-    - lf.linux.arm64.2xlarge
-    - lf.linux.4xlarge.nvidia.gpu
-    - lf.linux.8xlarge.nvidia.gpu
-    - lf.linux.16xlarge.nvidia.gpu
-    - lf.linux.g5.4xlarge.nvidia.gpu
-    # Repo-specific IBM hosted S390x runner
-    - linux.s390x
-    # Organization wide AWS Windows runners
     - windows.4xlarge.nonephemeral
     - windows.8xlarge.nvidia.gpu
     - windows.8xlarge.nvidia.gpu.nonephemeral
     - windows.g5.4xlarge.nvidia.gpu
-    # Organization-wide AMD hosted MI300 runners
+    - bm-runner
     - linux.rocm.gpu
-    # Repo-specific Apple hosted  runners
-    - macos-m1-ultra
-    - macos-m2-14
-    # Org wise AWS `mac2.metal` runners (2020 Mac mini hardware powered by Apple silicon M1 processors)
     - macos-m1-stable
     - macos-m1-13
-    - macos-m1-14
-    # GitHub-hosted MacOS runners
+    - macos-12-xl
+    - macos-12
+    - macos12.3-m1
     - macos-latest-xlarge
     - macos-13-xlarge
-    - macos-14-xlarge

.github/actions/filter-test-configs/action.yml

@@ -66,8 +66,7 @@ runs:
         command: |
           set -eux
           # PyYAML 6.0 doesn't work with MacOS x86 anymore
-          # This must run on Python-3.7 (AmazonLinux2) so can't use request=3.32.2
-          python3 -m pip install requests==2.27.1 pyyaml==6.0.1
+          python3 -m pip install requests==2.26.0 pyyaml==6.0.1
 
     - name: Parse ref
       id: parse-ref

.github/actions/test-pytorch-binary/action.yml

@@ -35,7 +35,7 @@ runs:
           "${DOCKER_IMAGE}"
         )
 
-        if [[ "${GPU_ARCH_TYPE}" != "rocm" && "${BUILD_ENVIRONMENT}" != "linux-aarch64-binary-manywheel" && "${BUILD_ENVIRONMENT}" != "linux-s390x-binary-manywheel" ]]; then
+        if [[ "${GPU_ARCH_TYPE}" != "rocm" && "${BUILD_ENVIRONMENT}" != "linux-aarch64-binary-manywheel" ]]; then
           # Propagate download.pytorch.org IP to container. This is only needed on Linux non aarch64 runner
           grep download.pytorch.org /etc/hosts | docker exec -i "${container_name}" bash -c "/bin/cat >> /etc/hosts"
         fi
@@ -44,12 +44,3 @@ runs:
         # Generate test script
         docker exec -t -w "${PYTORCH_ROOT}" -e OUTPUT_SCRIPT="/run.sh" "${container_name}" bash -c "bash .circleci/scripts/binary_linux_test.sh"
         docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash -x /run.sh"
-
-    - name: Cleanup docker
-      if: always() && env.BUILD_ENVIRONMENT == 'linux-s390x-binary-manywheel'
-      shell: bash
-      run: |
-        # on s390x stop the container for clean worker stop
-        # ignore expansion of "docker ps -q" since it could be empty
-        # shellcheck disable=SC2046
-        docker stop $(docker ps -q) || true

.github/ci_commit_pins/audio.txt

@@ -1 +1 @@
-b829e936f7cc61b48149f5f957a451a38bf2a178
+ea437b31ce316ea3d66fe73768c0dcb94edb79ad

.github/ci_commit_pins/vision.txt

@@ -1 +1 @@
-d23a6e1664d20707c11781299611436e1f0c104f
+2c4665ffbb64f03f5d18016d3398af4ac4da5f03

.github/ci_commit_pins/xla.txt

@@ -1 +1 @@
-6f0b61e5d782913a0fc7743812f2a8e522189111
+58a412cb271a3f98ae2e01fd1d24bdbb66645d4e

.github/labeler.yml

@@ -58,17 +58,6 @@
 - third_party/mkl-dnn.BUILD
 - torch/csrc/jit/codegen/onednn/**
 - test/test_jit_llga_fuser.py
-- test/test_mkldnn.py
-
-"ciflow/linux-aarch64":
-- third_party/ideep
-- caffe2/ideep/**
-- caffe2/python/ideep/**
-- cmake/Modules/FindMKLDNN.cmake
-- third_party/mkl-dnn.BUILD
-- torch/csrc/jit/codegen/onednn/**
-- test/test_jit_llga_fuser.py
-- test/test_mkldnn.py
 
 "module: amp (automated mixed precision)":
 - torch/amp/**

.github/lf-canary-scale-config.yml

@@ -1,154 +0,0 @@
-# Defines runner types that will be provisioned by by LF Self-hosted
-# runners for pytorch/pytorch-canary and their labels.
-#
-# Runners listed here will be available as self hosted runners.
-# Configuration is directly pulled from the main branch.
-#
-# Default values:
-#
-# runner_types:
-#   runner_label: # label to specify in the Github Actions workflow
-#     instance_type: m4.large
-#     os: linux
-#     max_available: 20
-#     disk_size: 50
-#     is_ephemeral: true
-
-runner_types:
-  lf.c.linux.12xlarge:
-    disk_size: 200
-    instance_type: c5.12xlarge
-    is_ephemeral: false
-    max_available: 1000
-    os: linux
-  lf.c.linux.24xl.spr-metal:
-    disk_size: 200
-    instance_type: c7i.metal-24xl
-    is_ephemeral: false
-    max_available: 30
-    os: linux
-  lf.c.linux.16xlarge.spr:
-    disk_size: 200
-    instance_type: c7i.16xlarge
-    is_ephemeral: false
-    max_available: 30
-    os: linux
-  lf.c.linux.12xlarge.ephemeral:
-    disk_size: 200
-    instance_type: c5.12xlarge
-    is_ephemeral: true
-    max_available: 300
-    os: linux
-  lf.c.linux.16xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g3.16xlarge
-    is_ephemeral: false
-    max_available: 30
-    os: linux
-  lf.c.linux.24xlarge:
-    disk_size: 150
-    instance_type: c5.24xlarge
-    is_ephemeral: false
-    max_available: 250
-    os: linux
-  lf.c.linux.2xlarge:
-    disk_size: 150
-    instance_type: c5.2xlarge
-    is_ephemeral: false
-    max_available: 3120
-    os: linux
-  lf.c.linux.4xlarge:
-    disk_size: 150
-    instance_type: c5.4xlarge
-    is_ephemeral: false
-    max_available: 1000
-    os: linux
-  lf.c.linux.4xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g3.4xlarge
-    is_ephemeral: false
-    max_available: 520
-    os: linux
-  lf.c.linux.8xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g3.8xlarge
-    is_ephemeral: false
-    max_available: 400
-    os: linux
-  lf.c.linux.g4dn.12xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g4dn.12xlarge
-    is_ephemeral: false
-    max_available: 50
-    os: linux
-  lf.c.linux.g4dn.metal.nvidia.gpu:
-    disk_size: 150
-    instance_type: g4dn.metal
-    is_ephemeral: false
-    max_available: 30
-    os: linux
-  lf.c.linux.g5.48xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g5.48xlarge
-    is_ephemeral: false
-    max_available: 20
-    os: linux
-  lf.c.linux.g5.12xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g5.12xlarge
-    is_ephemeral: false
-    max_available: 150
-    os: linux
-  lf.c.linux.g5.4xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g5.4xlarge
-    is_ephemeral: false
-    max_available: 1200
-    os: linux
-  lf.c.linux.large:
-    disk_size: 15
-    instance_type: c5.large
-    is_ephemeral: false
-    os: linux
-  lf.c.linux.arm64.2xlarge:
-    disk_size: 256
-    instance_type: t4g.2xlarge
-    is_ephemeral: false
-    max_available: 200
-    os: linux
-  lf.c.linux.arm64.m7g.2xlarge:
-    disk_size: 256
-    instance_type: m7g.2xlarge
-    is_ephemeral: false
-    max_available: 20
-    os: linux
-  lf.c.windows.4xlarge:
-    disk_size: 256
-    instance_type: c5d.4xlarge
-    is_ephemeral: true
-    max_available: 420
-    os: windows
-  lf.c.windows.4xlarge.nonephemeral:
-    disk_size: 256
-    instance_type: c5d.4xlarge
-    is_ephemeral: false
-    max_available: 420
-    os: windows
-  lf.c.windows.8xlarge.nvidia.gpu:
-    disk_size: 256
-    instance_type: p3.2xlarge
-    is_ephemeral: true
-    max_available: 150
-    os: windows
-  lf.c.windows.8xlarge.nvidia.gpu.nonephemeral:
-    disk_size: 256
-    instance_type: p3.2xlarge
-    is_ephemeral: false
-    max_available: 150
-    os: windows
-  lf.c.windows.g5.4xlarge.nvidia.gpu:
-    disk_size: 256
-    instance_type: g5.4xlarge
-    is_ephemeral: false
-    max_available: 250
-    os: windows

.github/lf-scale-config.yml

@@ -1,154 +0,0 @@
-# Defines runner types that will be provisioned by by LF Self-hosted
-# runners for pytorch/pytorch and their labels.
-#
-# Runners listed here will be available as self hosted runners.
-# Configuration is directly pulled from the main branch.
-#
-# Default values:
-#
-# runner_types:
-#   runner_label: # label to specify in the Github Actions workflow
-#     instance_type: m4.large
-#     os: linux
-#     max_available: 20
-#     disk_size: 50
-#     is_ephemeral: true
-
-runner_types:
-  lf.linux.12xlarge:
-    disk_size: 200
-    instance_type: c5.12xlarge
-    is_ephemeral: false
-    max_available: 1000
-    os: linux
-  lf.linux.24xl.spr-metal:
-    disk_size: 200
-    instance_type: c7i.metal-24xl
-    is_ephemeral: false
-    max_available: 30
-    os: linux
-  lf.linux.16xlarge.spr:
-    disk_size: 200
-    instance_type: c7i.16xlarge
-    is_ephemeral: false
-    max_available: 30
-    os: linux
-  lf.linux.12xlarge.ephemeral:
-    disk_size: 200
-    instance_type: c5.12xlarge
-    is_ephemeral: true
-    max_available: 300
-    os: linux
-  lf.linux.16xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g3.16xlarge
-    is_ephemeral: false
-    max_available: 30
-    os: linux
-  lf.linux.24xlarge:
-    disk_size: 150
-    instance_type: c5.24xlarge
-    is_ephemeral: false
-    max_available: 250
-    os: linux
-  lf.linux.2xlarge:
-    disk_size: 150
-    instance_type: c5.2xlarge
-    is_ephemeral: false
-    max_available: 3120
-    os: linux
-  lf.linux.4xlarge:
-    disk_size: 150
-    instance_type: c5.4xlarge
-    is_ephemeral: false
-    max_available: 1000
-    os: linux
-  lf.linux.4xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g3.4xlarge
-    is_ephemeral: false
-    max_available: 520
-    os: linux
-  lf.linux.8xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g3.8xlarge
-    is_ephemeral: false
-    max_available: 400
-    os: linux
-  lf.linux.g4dn.12xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g4dn.12xlarge
-    is_ephemeral: false
-    max_available: 50
-    os: linux
-  lf.linux.g4dn.metal.nvidia.gpu:
-    disk_size: 150
-    instance_type: g4dn.metal
-    is_ephemeral: false
-    max_available: 30
-    os: linux
-  lf.linux.g5.48xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g5.48xlarge
-    is_ephemeral: false
-    max_available: 20
-    os: linux
-  lf.linux.g5.12xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g5.12xlarge
-    is_ephemeral: false
-    max_available: 150
-    os: linux
-  lf.linux.g5.4xlarge.nvidia.gpu:
-    disk_size: 150
-    instance_type: g5.4xlarge
-    is_ephemeral: false
-    max_available: 1200
-    os: linux
-  lf.linux.large:
-    disk_size: 15
-    instance_type: c5.large
-    is_ephemeral: false
-    os: linux
-  lf.linux.arm64.2xlarge:
-    disk_size: 256
-    instance_type: t4g.2xlarge
-    is_ephemeral: false
-    max_available: 200
-    os: linux
-  lf.linux.arm64.m7g.2xlarge:
-    disk_size: 256
-    instance_type: m7g.2xlarge
-    is_ephemeral: false
-    max_available: 20
-    os: linux
-  lf.windows.4xlarge:
-    disk_size: 256
-    instance_type: c5d.4xlarge
-    is_ephemeral: true
-    max_available: 420
-    os: windows
-  lf.windows.4xlarge.nonephemeral:
-    disk_size: 256
-    instance_type: c5d.4xlarge
-    is_ephemeral: false
-    max_available: 420
-    os: windows
-  lf.windows.8xlarge.nvidia.gpu:
-    disk_size: 256
-    instance_type: p3.2xlarge
-    is_ephemeral: true
-    max_available: 150
-    os: windows
-  lf.windows.8xlarge.nvidia.gpu.nonephemeral:
-    disk_size: 256
-    instance_type: p3.2xlarge
-    is_ephemeral: false
-    max_available: 150
-    os: windows
-  lf.windows.g5.4xlarge.nvidia.gpu:
-    disk_size: 256
-    instance_type: g5.4xlarge
-    is_ephemeral: false
-    max_available: 250
-    os: windows

.github/merge_rules.yaml

@@ -29,12 +29,10 @@
   approved_by:
   - BowenBao
   - justinchuby
-  - liqunfu
   - shubhambhokare1
   - thiagocrepaldi
   - titaiwangms
   - wschin
-  - xadupre
   mandatory_checks_name:
   - EasyCLA
   - Lint
@@ -245,7 +243,6 @@
   - torch/xpu/**
   - test/xpu/**
   - third_party/xpu.txt
-  - .ci/docker/ci_commit_pins/triton-xpu.txt
   approved_by:
   - EikanWang
   - jgong5

.github/pytorch-probot.yml

@@ -1,5 +1,6 @@
 tracking_issue: 24422
 ciflow_tracking_issue: 64124
+TD_rollout_issue: 123120
 ciflow_push_tags:
 - ciflow/binaries
 - ciflow/binaries_conda
@@ -7,9 +8,6 @@ ciflow_push_tags:
 - ciflow/binaries_wheel
 - ciflow/inductor
 - ciflow/inductor-perf-compare
-- ciflow/inductor-micro-benchmark
-- ciflow/inductor-cu124
-- ciflow/linux-aarch64
 - ciflow/mps
 - ciflow/nightly
 - ciflow/periodic
@@ -20,6 +18,7 @@ ciflow_push_tags:
 - ciflow/xpu
 - ciflow/torchbench
 retryable_workflows:
+- lint
 - pull
 - trunk
 - linux-binary

.github/requirements-gha-cache.txt

@@ -5,11 +5,11 @@
 #   functorch/docs/requirements.txt
 #   .ci/docker/requirements-ci.txt
 boto3==1.19.12
-jinja2==3.1.4
+jinja2==3.1.3
 lintrunner==0.10.7
 ninja==1.10.0.post1
 nvidia-ml-py==11.525.84
 pyyaml==6.0
-requests==2.32.2
+requests==2.31.0
 rich==10.9.0
 rockset==1.0.3

.github/requirements/conda-env-Linux-X64.txt

@@ -4,5 +4,6 @@ mkl-include=2022.1.0
 ninja=1.10.2
 numpy=1.23.3
 pyyaml=6.0
+requests=2.31.0
 setuptools=68.2.2
-typing-extensions=4.9.0
+typing-extensions=4.3.0

.github/requirements/conda-env-iOS.txt

@@ -3,5 +3,6 @@ cmake=3.22.1
 ninja=1.10.2
 numpy=1.23.3
 pyyaml=6.0
+requests=2.31.0
 setuptools=68.2.2
-typing-extensions=4.9.0
+typing-extensions=4.3.0

.github/requirements/conda-env-macOS-ARM64

@@ -2,7 +2,7 @@ numpy=1.22.3
 pyyaml=6.0
 setuptools=61.2.0
 cmake=3.22.*
-typing-extensions=4.9.0
+typing-extensions=4.3.0
 dataclasses=0.8
 pip=22.2.2
 pillow=10.0.1

.github/requirements/conda-env-macOS-X64

@@ -4,7 +4,7 @@ numpy=1.21.2
 pyyaml=5.3
 setuptools=46.0.0
 cmake=3.22.*
-typing-extensions=4.9.0
+typing-extensions=4.3.0
 dataclasses=0.8
 pip=22.2.2
 pillow=10.0.1

.github/scripts/amd/patch_triton_wheel.sh

@@ -1,11 +1,7 @@
 #!/bin/bash
 set -x
 
-if [ -z "$1" ]; then
-    echo "Need wheel location argument" && exit 1
-fi
-
-WHEELHOUSE_DIR=$1
+WHEELHOUSE_DIR=/artifacts
 PATCHELF_BIN=patchelf
 ROCM_LIB=backends/amd/lib
 ROCM_LD=backends/amd/llvm/bin

.github/scripts/build_triton_wheel.py

@@ -157,10 +157,10 @@ def build_triton(
 
         if build_rocm:
             check_call(
-                [f"{SCRIPT_DIR}/amd/patch_triton_wheel.sh", Path.cwd()],
+                [f"{SCRIPT_DIR}/amd/patch_triton_wheel.sh"],
                 cwd=triton_basedir,
+                shell=True,
             )
-
         return Path.cwd() / whl_path.name
 
 

.github/scripts/delete_old_branches.py

@@ -2,7 +2,6 @@
 import os
 import re
 from datetime import datetime
-from functools import lru_cache
 from pathlib import Path
 from typing import Any, Callable, Dict, List, Set
 
@@ -188,17 +187,6 @@ def get_recent_prs() -> Dict[str, Any]:
     return prs_by_branch_base
 
 
-@lru_cache(maxsize=1)
-def get_open_prs() -> List[Dict[str, Any]]:
-    return paginate_graphql(
-        GRAPHQL_OPEN_PRS,
-        {"owner": "pytorch", "repo": "pytorch"},
-        lambda data: False,
-        lambda res: res["data"]["repository"]["pullRequests"]["nodes"],
-        lambda res: res["data"]["repository"]["pullRequests"]["pageInfo"],
-    )
-
-
 def get_branches_with_magic_label_or_open_pr() -> Set[str]:
     pr_infos: List[Dict[str, Any]] = paginate_graphql(
         GRAPHQL_NO_DELETE_BRANCH_LABEL,
@@ -208,7 +196,15 @@ def get_branches_with_magic_label_or_open_pr() -> Set[str]:
         lambda res: res["data"]["repository"]["label"]["pullRequests"]["pageInfo"],
     )
 
-    pr_infos.extend(get_open_prs())
+    pr_infos.extend(
+        paginate_graphql(
+            GRAPHQL_OPEN_PRS,
+            {"owner": "pytorch", "repo": "pytorch"},
+            lambda data: False,
+            lambda res: res["data"]["repository"]["pullRequests"]["nodes"],
+            lambda res: res["data"]["repository"]["pullRequests"]["pageInfo"],
+        )
+    )
 
     # Get the most recent PR for each branch base (group gh together)
     branch_bases = set()
@@ -274,41 +270,5 @@ def delete_branches() -> None:
         delete_branch(git_repo, branch)
 
 
-def delete_old_ciflow_tags() -> None:
-    # Deletes ciflow tags if they are associated with a closed PR or a specific
-    # commit.  Lightweight tags don't have information about the date they were
-    # created, so we can't check how old they are.  The script just assumes that
-    # ciflow tags should be deleted regardless of creation date.
-    git_repo = GitRepo(str(REPO_ROOT), "origin", debug=True)
-
-    def delete_tag(tag: str) -> None:
-        print(f"Deleting tag {tag}")
-        ESTIMATED_TOKENS[0] += 1
-        delete_branch(git_repo, f"refs/tags/{tag}")
-
-    tags = git_repo._run_git("tag").splitlines()
-    open_pr_numbers = [x["number"] for x in get_open_prs()]
-
-    for tag in tags:
-        try:
-            if ESTIMATED_TOKENS[0] > 400:
-                print("Estimated tokens exceeded, exiting")
-                break
-            if not tag.startswith("ciflow/"):
-                continue
-            re_match_pr = re.match(r"^ciflow\/.*\/(\d{5,6})$", tag)
-            re_match_sha = re.match(r"^ciflow\/.*\/([0-9a-f]{40})$", tag)
-            if re_match_pr:
-                pr_number = int(re_match_pr.group(1))
-                if pr_number in open_pr_numbers:
-                    continue
-                delete_tag(tag)
-            elif re_match_sha:
-                delete_tag(tag)
-        except Exception as e:
-            print(f"Failed to check tag {tag}: {e}")
-
-
 if __name__ == "__main__":
     delete_branches()
-    delete_old_ciflow_tags()

.github/scripts/docathon-label-sync.py

@@ -1,52 +0,0 @@
-import os
-import re
-import sys
-
-from github import Github
-
-
-def main() -> None:
-    token = os.environ.get("GITHUB_TOKEN")
-
-    repo_owner = "pytorch"
-    repo_name = "pytorch"
-    pull_request_number = int(sys.argv[1])
-
-    g = Github(token)
-    repo = g.get_repo(f"{repo_owner}/{repo_name}")
-    pull_request = repo.get_pull(pull_request_number)
-    pull_request_body = pull_request.body
-    # PR without description
-    if pull_request_body is None:
-        return
-
-    # get issue number from the PR body
-    if not re.search(r"#\d{1,6}", pull_request_body):
-        print("The pull request does not mention an issue.")
-        return
-    issue_number = int(re.findall(r"#(\d{1,6})", pull_request_body)[0])
-    issue = repo.get_issue(issue_number)
-    issue_labels = issue.labels
-    docathon_label_present = any(
-        label.name == "docathon-h1-2024" for label in issue_labels
-    )
-
-    # if the issue has a docathon label, add all labels from the issue to the PR.
-    if not docathon_label_present:
-        print("The 'docathon-h1-2024' label is not present in the issue.")
-        return
-    pull_request_labels = pull_request.get_labels()
-    pull_request_label_names = [label.name for label in pull_request_labels]
-    issue_label_names = [label.name for label in issue_labels]
-    labels_to_add = [
-        label for label in issue_label_names if label not in pull_request_label_names
-    ]
-    if not labels_to_add:
-        print("The pull request already has the same labels.")
-        return
-    pull_request.add_to_labels(*labels_to_add)
-    print("Labels added to the pull request!")
-
-
-if __name__ == "__main__":
-    main()

.github/scripts/generate_binary_build_matrix.py

@@ -13,16 +13,16 @@ architectures:
 import os
 from typing import Dict, List, Optional, Tuple
 
-CUDA_ARCHES = ["11.8", "12.1", "12.4"]
+CUDA_ARCHES = ["11.8", "12.1"]
 
 
-CUDA_ARCHES_FULL_VERSION = {"11.8": "11.8.0", "12.1": "12.1.1", "12.4": "12.4.0"}
+CUDA_ARCHES_FULL_VERSION = {"11.8": "11.8.0", "12.1": "12.1.1"}
 
 
-CUDA_ARCHES_CUDNN_VERSION = {"11.8": "9", "12.1": "9", "12.4": "9"}
+CUDA_ARCHES_CUDNN_VERSION = {"11.8": "8", "12.1": "8"}
 
 
-ROCM_ARCHES = ["6.0", "6.1"]
+ROCM_ARCHES = ["5.7", "6.0"]
 
 
 CPU_CXX11_ABI_ARCH = ["cpu-cxx11-abi"]
@@ -31,18 +31,12 @@ CPU_CXX11_ABI_ARCH = ["cpu-cxx11-abi"]
 CPU_AARCH64_ARCH = ["cpu-aarch64"]
 
 
-CPU_S390X_ARCH = ["cpu-s390x"]
-
-
-CUDA_AARCH64_ARCH = ["cuda-aarch64"]
-
-
 PYTORCH_EXTRA_INSTALL_REQUIREMENTS = {
     "11.8": (
         "nvidia-cuda-nvrtc-cu11==11.8.89; platform_system == 'Linux' and platform_machine == 'x86_64' | "  # noqa: B950
         "nvidia-cuda-runtime-cu11==11.8.89; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-cuda-cupti-cu11==11.8.87; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cudnn-cu11==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | "
+        "nvidia-cudnn-cu11==8.7.0.84; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-cublas-cu11==11.11.3.6; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-cufft-cu11==10.9.0.58; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-curand-cu11==10.3.0.86; platform_system == 'Linux' and platform_machine == 'x86_64' | "
@@ -55,7 +49,7 @@ PYTORCH_EXTRA_INSTALL_REQUIREMENTS = {
         "nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | "  # noqa: B950
         "nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | "
+        "nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | "
@@ -64,20 +58,6 @@ PYTORCH_EXTRA_INSTALL_REQUIREMENTS = {
         "nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | "
         "nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'"
     ),
-    "12.4": (
-        "nvidia-cuda-nvrtc-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cuda-runtime-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cuda-cupti-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cublas-cu12==12.4.2.65; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cufft-cu12==11.2.0.44; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-curand-cu12==10.3.5.119; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cusolver-cu12==11.6.0.99; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-cusparse-cu12==12.3.0.142; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-nvtx-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64' | "
-        "nvidia-nvjitlink-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64'"
-    ),
 }
 
 
@@ -136,10 +116,6 @@ def arch_type(arch_version: str) -> str:
         return "cpu-cxx11-abi"
     elif arch_version in CPU_AARCH64_ARCH:
         return "cpu-aarch64"
-    elif arch_version in CPU_S390X_ARCH:
-        return "cpu-s390x"
-    elif arch_version in CUDA_AARCH64_ARCH:
-        return "cuda-aarch64"
     else:  # arch_version should always be "cpu" in this case
         return "cpu"
 
@@ -159,8 +135,6 @@ WHEEL_CONTAINER_IMAGES = {
     "cpu": f"pytorch/manylinux-builder:cpu-{DEFAULT_TAG}",
     "cpu-cxx11-abi": f"pytorch/manylinuxcxx11-abi-builder:cpu-cxx11-abi-{DEFAULT_TAG}",
     "cpu-aarch64": f"pytorch/manylinuxaarch64-builder:cpu-aarch64-{DEFAULT_TAG}",
-    "cpu-s390x": f"pytorch/manylinuxs390x-builder:cpu-s390x-{DEFAULT_TAG}",
-    "cuda-aarch64": f"pytorch/manylinuxaarch64-builder:cuda12.4-{DEFAULT_TAG}",
 }
 
 CONDA_CONTAINER_IMAGES = {
@@ -217,9 +191,7 @@ def translate_desired_cuda(gpu_arch_type: str, gpu_arch_version: str) -> str:
         "cpu": "cpu",
         "cpu-aarch64": "cpu",
         "cpu-cxx11-abi": "cpu-cxx11-abi",
-        "cpu-s390x": "cpu",
         "cuda": f"cu{gpu_arch_version.replace('.', '')}",
-        "cuda-aarch64": "cu124",
         "rocm": f"rocm{gpu_arch_version}",
     }.get(gpu_arch_type, gpu_arch_version)
 
@@ -300,11 +272,11 @@ def generate_libtorch_matrix(
                     "libtorch_variant": libtorch_variant,
                     "libtorch_config": abi_version if os == "windows" else "",
                     "devtoolset": abi_version if os != "windows" else "",
-                    "container_image": (
-                        LIBTORCH_CONTAINER_IMAGES[(arch_version, abi_version)]
-                        if os != "windows"
-                        else ""
-                    ),
+                    "container_image": LIBTORCH_CONTAINER_IMAGES[
+                        (arch_version, abi_version)
+                    ]
+                    if os != "windows"
+                    else "",
                     "package_type": "libtorch",
                     "build_name": f"libtorch-{gpu_arch_type}{gpu_arch_version}-{libtorch_variant}-{abi_version}".replace(
                         ".", "_"
@@ -320,8 +292,8 @@ def generate_wheels_matrix(
     python_versions: Optional[List[str]] = None,
 ) -> List[Dict[str, str]]:
     package_type = "wheel"
-    if os == "linux" or os == "linux-aarch64" or os == "linux-s390x":
-        # NOTE: We only build manywheel packages for x86_64 and aarch64 and s390x linux
+    if os == "linux" or os == "linux-aarch64":
+        # NOTE: We only build manywheel packages for x86_64 and aarch64 linux
         package_type = "manywheel"
 
     if python_versions is None:
@@ -337,36 +309,22 @@ def generate_wheels_matrix(
         elif os == "linux-aarch64":
             # Only want the one arch as the CPU type is different and
             # uses different build/test scripts
-            arches = ["cpu-aarch64", "cuda-aarch64"]
-        elif os == "linux-s390x":
-            # Only want the one arch as the CPU type is different and
-            # uses different build/test scripts
-            arches = ["cpu-s390x"]
+            arches = ["cpu-aarch64"]
 
     ret: List[Dict[str, str]] = []
     for python_version in python_versions:
         for arch_version in arches:
             gpu_arch_type = arch_type(arch_version)
-            # Disable py3.12 builds for ROCm because of triton dependency
-            # on llnl-hatchet, which doesn't have py3.12 wheels available
-            if gpu_arch_type == "rocm" and python_version == "3.12":
-                continue
             gpu_arch_version = (
                 ""
                 if arch_version == "cpu"
                 or arch_version == "cpu-cxx11-abi"
                 or arch_version == "cpu-aarch64"
-                or arch_version == "cpu-s390x"
-                or arch_version == "cuda-aarch64"
                 else arch_version
             )
 
             # 12.1 linux wheels require PYTORCH_EXTRA_INSTALL_REQUIREMENTS to install
-            if (
-                arch_version in ["12.4", "12.1", "11.8"]
-                and os == "linux"
-                or arch_version == "cuda-aarch64"
-            ):
+            if arch_version in ["12.1", "11.8"] and os == "linux":
                 ret.append(
                     {
                         "python_version": python_version,
@@ -375,16 +333,10 @@ def generate_wheels_matrix(
                         "desired_cuda": translate_desired_cuda(
                             gpu_arch_type, gpu_arch_version
                         ),
-                        "devtoolset": (
-                            "cxx11-abi" if arch_version == "cuda-aarch64" else ""
-                        ),
+                        "devtoolset": "",
                         "container_image": WHEEL_CONTAINER_IMAGES[arch_version],
                         "package_type": package_type,
-                        "pytorch_extra_install_requirements": (
-                            PYTORCH_EXTRA_INSTALL_REQUIREMENTS[arch_version]  # fmt: skip
-                            if os != "linux-aarch64"
-                            else ""
-                        ),
+                        "pytorch_extra_install_requirements": PYTORCH_EXTRA_INSTALL_REQUIREMENTS[arch_version],  # fmt: skip
                         "build_name": f"{package_type}-py{python_version}-{gpu_arch_type}{gpu_arch_version}".replace(  # noqa: B950
                             ".", "_"
                         ),
@@ -399,24 +351,21 @@ def generate_wheels_matrix(
                         "desired_cuda": translate_desired_cuda(
                             gpu_arch_type, gpu_arch_version
                         ),
-                        "devtoolset": (
-                            "cxx11-abi" if arch_version == "cpu-cxx11-abi" else ""
-                        ),
+                        "devtoolset": "cxx11-abi"
+                        if arch_version == "cpu-cxx11-abi"
+                        else "",
                         "container_image": WHEEL_CONTAINER_IMAGES[arch_version],
                         "package_type": package_type,
                         "build_name": f"{package_type}-py{python_version}-{gpu_arch_type}{gpu_arch_version}".replace(
                             ".", "_"
                         ),
-                        "pytorch_extra_install_requirements": (
-                            PYTORCH_EXTRA_INSTALL_REQUIREMENTS["12.1"]  # fmt: skip
-                            if os != "linux"
-                            else ""
-                        ),
+                        "pytorch_extra_install_requirements":
+                        PYTORCH_EXTRA_INSTALL_REQUIREMENTS["12.1"]  # fmt: skip
+                        if os != "linux" else "",
                     }
                 )
     return ret
 
 
-validate_nccl_dep_consistency("12.4")
 validate_nccl_dep_consistency("12.1")
 validate_nccl_dep_consistency("11.8")

.github/scripts/generate_ci_workflows.py

@@ -5,11 +5,11 @@ import sys
 from dataclasses import asdict, dataclass, field
 from pathlib import Path
 from typing import Dict, Iterable, List, Literal, Set
-from typing_extensions import TypedDict  # Python 3.11+
 
 import generate_binary_build_matrix  # type: ignore[import]
 
 import jinja2
+from typing_extensions import TypedDict  # Python 3.11+
 
 Arch = Literal["windows", "linux", "macos"]
 
@@ -60,7 +60,7 @@ class BinaryBuildWorkflow:
     branches: str = "nightly"
     # Mainly for macos
     cross_compile_arm64: bool = False
-    macos_runner: str = "macos-14-xlarge"
+    macos_runner: str = "macos-12-xl"
 
     def __post_init__(self) -> None:
         if self.abi_version:
@@ -95,7 +95,6 @@ class OperatingSystem:
     MACOS = "macos"
     MACOS_ARM64 = "macos-arm64"
     LINUX_AARCH64 = "linux-aarch64"
-    LINUX_S390X = "linux-s390x"
 
 
 LINUX_BINARY_BUILD_WORFKLOWS = [
@@ -157,7 +156,7 @@ LINUX_BINARY_SMOKE_WORKFLOWS = [
         package_type="manywheel",
         build_configs=generate_binary_build_matrix.generate_wheels_matrix(
             OperatingSystem.LINUX,
-            arches=["11.8", "12.1", "12.4"],
+            arches=["11.8", "12.1"],
             python_versions=["3.8"],
         ),
         branches="main",
@@ -285,7 +284,7 @@ MACOS_BINARY_BUILD_WORKFLOWS = [
             libtorch_variants=["shared-with-deps"],
         ),
         cross_compile_arm64=False,
-        macos_runner="macos-14-xlarge",
+        macos_runner="macos-13-xlarge",
         ciflow_config=CIFlowConfig(
             labels={LABEL_CIFLOW_BINARIES, LABEL_CIFLOW_BINARIES_LIBTORCH},
             isolated_workflow=True,
@@ -298,7 +297,7 @@ MACOS_BINARY_BUILD_WORKFLOWS = [
             OperatingSystem.MACOS_ARM64
         ),
         cross_compile_arm64=False,
-        macos_runner="macos-14-xlarge",
+        macos_runner="macos-13-xlarge",
         ciflow_config=CIFlowConfig(
             labels={LABEL_CIFLOW_BINARIES, LABEL_CIFLOW_BINARIES_WHEEL},
             isolated_workflow=True,
@@ -308,7 +307,7 @@ MACOS_BINARY_BUILD_WORKFLOWS = [
         os=OperatingSystem.MACOS_ARM64,
         package_type="conda",
         cross_compile_arm64=False,
-        macos_runner="macos-14-xlarge",
+        macos_runner="macos-13-xlarge",
         build_configs=generate_binary_build_matrix.generate_conda_matrix(
             OperatingSystem.MACOS_ARM64
         ),
@@ -333,20 +332,6 @@ AARCH64_BINARY_BUILD_WORKFLOWS = [
     ),
 ]
 
-S390X_BINARY_BUILD_WORKFLOWS = [
-    BinaryBuildWorkflow(
-        os=OperatingSystem.LINUX_S390X,
-        package_type="manywheel",
-        build_configs=generate_binary_build_matrix.generate_wheels_matrix(
-            OperatingSystem.LINUX_S390X
-        ),
-        ciflow_config=CIFlowConfig(
-            labels={LABEL_CIFLOW_BINARIES, LABEL_CIFLOW_BINARIES_WHEEL},
-            isolated_workflow=True,
-        ),
-    ),
-]
-
 
 def main() -> None:
     jinja_env = jinja2.Environment(
@@ -365,10 +350,6 @@ def main() -> None:
             jinja_env.get_template("linux_binary_build_workflow.yml.j2"),
             AARCH64_BINARY_BUILD_WORKFLOWS,
         ),
-        (
-            jinja_env.get_template("linux_binary_build_workflow.yml.j2"),
-            S390X_BINARY_BUILD_WORKFLOWS,
-        ),
         (
             jinja_env.get_template("linux_binary_build_workflow.yml.j2"),
             LINUX_BINARY_SMOKE_WORKFLOWS,

.github/scripts/generate_docker_release_matrix.py

@@ -21,8 +21,6 @@ DOCKER_IMAGE_TYPES = ["runtime", "devel"]
 
 def generate_docker_matrix() -> Dict[str, List[Dict[str, str]]]:
     ret: List[Dict[str, str]] = []
-    # CUDA amd64 Docker images are available as both runtime and devel while
-    # CPU arm64 image is only available as runtime.
     for cuda, version in generate_binary_build_matrix.CUDA_ARCHES_FULL_VERSION.items():
         for image in DOCKER_IMAGE_TYPES:
             ret.append(
@@ -33,19 +31,9 @@ def generate_docker_matrix() -> Dict[str, List[Dict[str, str]]]:
                         cuda
                     ],
                     "image_type": image,
-                    "platform": "linux/amd64",
+                    "platform": "linux/arm64,linux/amd64",
                 }
             )
-    ret.append(
-        {
-            "cuda": "cpu",
-            "cuda_full_version": "",
-            "cudnn_version": "",
-            "image_type": "runtime",
-            "platform": "linux/arm64",
-        }
-    )
-
     return {"include": ret}
 
 

.github/scripts/get_workflow_type.py

@@ -1,99 +0,0 @@
-import json
-from argparse import ArgumentParser
-from typing import Any
-
-from github import Auth, Github
-from github.Issue import Issue
-
-
-WORKFLOW_TYPE_LABEL = "label"
-WORKFLOW_TYPE_RG = "rg"
-WORKFLOW_TYPE_BOTH = "both"
-
-
-def parse_args() -> Any:
-    parser = ArgumentParser("Get dynamic rollout settings")
-    parser.add_argument("--github-token", type=str, required=True, help="GitHub token")
-    parser.add_argument(
-        "--github-repo",
-        type=str,
-        required=False,
-        default="pytorch/test-infra",
-        help="GitHub repo to get the issue",
-    )
-    parser.add_argument(
-        "--github-issue", type=int, required=True, help="GitHub issue umber"
-    )
-    parser.add_argument(
-        "--github-user", type=str, required=True, help="GitHub username"
-    )
-    parser.add_argument(
-        "--github-branch", type=str, required=True, help="Current GitHub branch"
-    )
-
-    return parser.parse_args()
-
-
-def get_gh_client(github_token: str) -> Github:
-    auth = Auth.Token(github_token)
-    return Github(auth=auth)
-
-
-def get_issue(gh: Github, repo: str, issue_num: int) -> Issue:
-    repo = gh.get_repo(repo)
-    return repo.get_issue(number=issue_num)
-
-
-def is_exception_branch(branch: str) -> bool:
-    return branch.split("/")[0] in {"main", "nightly", "release", "landchecks"}
-
-
-def get_workflow_type(issue: Issue, username: str) -> str:
-    user_list = issue.get_comments()[0].body.split("\r\n")
-    try:
-        run_option = issue.get_comments()[1].body.split("\r\n")[0]
-    except Exception as e:
-        run_option = "single"
-
-    if user_list[0] == "!":
-        # Use old runners for everyone
-        return WORKFLOW_TYPE_LABEL
-    elif user_list[1] == "*":
-        if run_option == WORKFLOW_TYPE_BOTH:
-            # Use ARC runners and old runners for everyone
-            return WORKFLOW_TYPE_BOTH
-        else:
-            # Use only ARC runners for everyone
-            return WORKFLOW_TYPE_RG
-    elif username in user_list:
-        if run_option == WORKFLOW_TYPE_BOTH:
-            # Use ARC runners and old runners for a specific user
-            return WORKFLOW_TYPE_BOTH
-        else:
-            # Use only ARC runners for a specific user
-            return WORKFLOW_TYPE_RG
-    else:
-        # Use old runners by default
-        return WORKFLOW_TYPE_LABEL
-
-
-def main() -> None:
-    args = parse_args()
-
-    if is_exception_branch(args.github_branch):
-        output = {"workflow_type": WORKFLOW_TYPE_LABEL}
-    else:
-        try:
-            gh = get_gh_client(args.github_token)
-            issue = get_issue(gh, args.github_repo, args.github_issue)
-
-            output = {"workflow_type": get_workflow_type(issue, args.github_user)}
-        except Exception as e:
-            output = {"workflow_type": WORKFLOW_TYPE_LABEL}
-
-    json_output = json.dumps(output)
-    print(json_output)
-
-
-if __name__ == "__main__":
-    main()

.github/scripts/lintrunner.sh

@@ -7,7 +7,7 @@ eval "$(command conda 'shell.bash' 'hook' 2> /dev/null)"
 conda activate "${CONDA_ENV}"
 
 # Use uv to speed up lintrunner init
-python3 -m pip install uv==0.1.45
+python3 -m pip install uv
 
 CACHE_DIRECTORY="/tmp/.lintbin"
 # Try to recover the cached binaries

.github/scripts/pytest_caching_utils.py

@@ -18,7 +18,6 @@ PYTEST_CACHE_KEY_PREFIX = "pytest_cache"
 PYTEST_CACHE_DIR_NAME = ".pytest_cache"
 BUCKET = "gha-artifacts"
 LASTFAILED_FILE_PATH = Path("v/cache/lastfailed")
-TD_HEURISTIC_PREVIOUSLY_FAILED_ADDITIONAL = "previous_failures_additional.json"
 
 # Temp folders
 ZIP_UPLOAD = "zip-upload"
@@ -192,10 +191,6 @@ def _merge_pytest_caches(
         pytest_cache_dir_to_merge_from, pytest_cache_dir_to_merge_into
     )
 
-    _merge_additional_failures_files(
-        pytest_cache_dir_to_merge_from, pytest_cache_dir_to_merge_into
-    )
-
 
 def _merge_lastfailed_files(source_pytest_cache: Path, dest_pytest_cache: Path) -> None:
     # Simple cases where one of the files doesn't exist
@@ -237,27 +232,3 @@ def _merged_lastfailed_content(
             del to_lastfailed[""]
 
     return to_lastfailed
-
-
-def _merge_additional_failures_files(
-    source_pytest_cache: Path, dest_pytest_cache: Path
-) -> None:
-    # Simple cases where one of the files doesn't exist
-    source_lastfailed_file = (
-        source_pytest_cache / TD_HEURISTIC_PREVIOUSLY_FAILED_ADDITIONAL
-    )
-    dest_lastfailed_file = dest_pytest_cache / TD_HEURISTIC_PREVIOUSLY_FAILED_ADDITIONAL
-
-    if not source_lastfailed_file.exists():
-        return
-    if not dest_lastfailed_file.exists():
-        copy_file(source_lastfailed_file, dest_lastfailed_file)
-        return
-
-    # Both files exist, so we need to merge them
-    from_lastfailed = load_json_file(source_lastfailed_file)
-    to_lastfailed = load_json_file(dest_lastfailed_file)
-    merged_content = list(set(from_lastfailed + to_lastfailed))
-
-    # Save the results
-    write_json_file(dest_lastfailed_file, merged_content)

.github/scripts/sync_distributed_folder_prototype.sh

@@ -1,28 +0,0 @@
-#!/bin/bash
-
-set -eoux pipefail
-
-SYNC_BRANCH=fbcode/pytorch-stable-prototype
-
-git config user.email "fake@example.com"
-git config user.name  "PyTorch Stable Bot"
-
-git fetch origin main
-git fetch origin "$SYNC_BRANCH"
-git checkout "$SYNC_BRANCH"
-
-for SHA in $(git log 4333e122d4b74cdf84351ed2907045c6a767b4cd..origin/main --pretty="%h" --reverse -- torch/distributed torch/csrc/distributed test/distributed test/cpp/c10d benchmarks/distributed)
-do
-    # `git merge-base --is-ancestor` exits with code 0 if the given SHA is an ancestor, and non-0 otherwise
-    if git merge-base --is-ancestor $SHA HEAD || [[ $(git log --grep="(cherry picked from commit $SHA") ]]
-    then
-        echo "Skipping $SHA"
-        continue
-    fi
-    echo "Copying $SHA"
-    git cherry-pick -x "$SHA"
-done
-
-if [[ "${WITH_PUSH}" == true ]]; then
-  git push
-fi

.github/scripts/test_trymerge.py

@@ -773,13 +773,13 @@ class TestBypassFailures(TestCase):
                 # than the one on the base commit. This should still count as broken trunk
                 "pr_num": 104214,
                 "related_failure_count": 0,
-                "flaky_or_broken_trunk": 1,
+                "unrelated_failure_count": 1,
             },
             {
                 # This PR had one broken trunk failure and it used ghstack
                 "pr_num": 105145,
                 "related_failure_count": 0,
-                "flaky_or_broken_trunk": 1,
+                "unrelated_failure_count": 1,
             },
             {
                 # The failure on the merge base was retried successfully and
@@ -788,20 +788,20 @@ class TestBypassFailures(TestCase):
                 # be used to detect broken trunk
                 "pr_num": 107160,
                 "related_failure_count": 0,
-                "flaky_or_broken_trunk": 1,
+                "unrelated_failure_count": 4,
             },
             {
                 # This PR used Dr.CI broken trunk classification
                 "pr_num": 111253,
                 "related_failure_count": 1,
-                "flaky_or_broken_trunk": 1,
+                "unrelated_failure_count": 2,
             },
         ]
 
         for case in test_cases:
             pr_num = case["pr_num"]
             related_failure_count = case["related_failure_count"]
-            flaky_or_broken_trunk = case["flaky_or_broken_trunk"]
+            unrelated_failure_count = case["unrelated_failure_count"]
 
             pr = GitHubPR("pytorch", "pytorch", pr_num)
             checks = pr.get_checkrun_conclusions()
@@ -823,7 +823,7 @@ class TestBypassFailures(TestCase):
             )
             self.assertTrue(len(pending) == 0)
             self.assertTrue(
-                len(failed) == flaky_or_broken_trunk + related_failure_count
+                len(failed) == unrelated_failure_count + related_failure_count
             )
 
     def test_ignore_current(self, *args: Any) -> None:

.github/scripts/trymerge.py

@@ -2027,8 +2027,10 @@ def categorize_checks(
     pending_checks: List[Tuple[str, Optional[str], Optional[int]]] = []
     failed_checks: List[Tuple[str, Optional[str], Optional[int]]] = []
 
-    # failed_checks_categorization is used to keep track of all ignorable failures when saving the merge record on Rockset
-    failed_checks_categorization: Dict[str, List[Any]] = defaultdict(list)
+    # ok_failed_checks is used with ok_failed_checks_threshold while ignorable_failed_checks
+    # is used to keep track of all ignorable failures when saving the merge record on Rockset
+    ok_failed_checks: List[Tuple[str, Optional[str], Optional[int]]] = []
+    ignorable_failed_checks: Dict[str, List[Any]] = defaultdict(list)
 
     # If required_checks is not set or empty, consider all names are relevant
     relevant_checknames = [
@@ -2056,38 +2058,36 @@ def categorize_checks(
             continue
         elif not is_passing_status(check_runs[checkname].status):
             target = (
-                failed_checks_categorization[classification]
+                ignorable_failed_checks[classification]
                 if classification
                 in ("IGNORE_CURRENT_CHECK", "BROKEN_TRUNK", "FLAKY", "UNSTABLE")
                 else failed_checks
             )
             target.append((checkname, url, job_id))
 
-    flaky_or_broken_trunk = (
-        failed_checks_categorization["BROKEN_TRUNK"]
-        + failed_checks_categorization["FLAKY"]
-    )
+            if classification in ("BROKEN_TRUNK", "FLAKY", "UNSTABLE"):
+                ok_failed_checks.append((checkname, url, job_id))
 
-    if flaky_or_broken_trunk:
+    if ok_failed_checks:
         warn(
-            f"The following {len(flaky_or_broken_trunk)} checks failed but were likely due flakiness or broken trunk: "
-            + ", ".join([x[0] for x in flaky_or_broken_trunk])
+            f"The following {len(ok_failed_checks)} checks failed but were likely due flakiness or broken trunk: "
+            + ", ".join([x[0] for x in ok_failed_checks])
             + (
                 f" but this is greater than the threshold of {ok_failed_checks_threshold} so merge will fail"
                 if ok_failed_checks_threshold is not None
-                and len(flaky_or_broken_trunk) > ok_failed_checks_threshold
+                and len(ok_failed_checks) > ok_failed_checks_threshold
                 else ""
             )
         )
 
     if (
         ok_failed_checks_threshold is not None
-        and len(flaky_or_broken_trunk) > ok_failed_checks_threshold
+        and len(ok_failed_checks) > ok_failed_checks_threshold
     ):
-        failed_checks = failed_checks + flaky_or_broken_trunk
+        failed_checks = failed_checks + ok_failed_checks
 
-    # The list of failed_checks_categorization is returned so that it can be saved into the Rockset merge record
-    return (pending_checks, failed_checks, failed_checks_categorization)
+    # The list of ignorable_failed_checks is returned so that it can be saved into the Rockset merge record
+    return (pending_checks, failed_checks, ignorable_failed_checks)
 
 
 def merge(

.github/templates/linux_binary_build_workflow.yml.j2

@@ -33,8 +33,6 @@ env:
   # Needed for conda builds
   {%- if "aarch64" in build_environment %}
   ALPINE_IMAGE: "arm64v8/alpine"
-  {%- elif "s390x" in build_environment %}
-  ALPINE_IMAGE: "docker.io/s390x/alpine"
   {%- else %}
   ALPINE_IMAGE: "308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/alpine"
   {%- endif %}
@@ -48,7 +46,7 @@ env:
   PYTORCH_FINAL_PACKAGE_DIR: /artifacts
   PYTORCH_ROOT: /pytorch
   SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 !{{ common.concurrency(build_environment) }}
 
 jobs:
@@ -58,11 +56,8 @@ jobs:
     uses: ./.github/workflows/_binary-build-linux.yml
     with:!{{ upload.binary_env_as_input(config) }}
       {%- if "aarch64" in build_environment %}
-      runs_on: linux.arm64.m7g.4xlarge
+      runs_on: linux.arm64.2xlarge
       ALPINE_IMAGE: "arm64v8/alpine"
-      {%- elif "s390x" in build_environment %}
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
       {%- elif "conda" in build_environment and config["gpu_arch_type"] == "cuda" %}
       runs_on: linux.24xlarge
       {%- endif %}
@@ -71,17 +66,12 @@ jobs:
       {%- if config.pytorch_extra_install_requirements is defined and config.pytorch_extra_install_requirements|d('')|length > 0  %}
       PYTORCH_EXTRA_INSTALL_REQUIREMENTS: !{{ config.pytorch_extra_install_requirements }}
       {%- endif %}
-      {%- if config["gpu_arch_type"] == "cuda-aarch64" %}
-      timeout-minutes: 420
-      {%- endif %}
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
-
-  {%- if config["gpu_arch_type"] != "cuda-aarch64" %}
   !{{ config["build_name"] }}-test:  # Testing
     if: ${{ github.repository_owner == 'pytorch' }}
     needs: !{{ config["build_name"] }}-build
-    {%- if config["gpu_arch_type"] != "rocm" %}
+{%- if config["gpu_arch_type"] != "rocm" %}
     uses: ./.github/workflows/_binary-test-linux.yml
     with:!{{ upload.binary_env_as_input(config) }}
       build_name: !{{ config["build_name"] }}
@@ -89,9 +79,6 @@ jobs:
       {%- if "aarch64" in build_environment %}
       runs_on: linux.arm64.2xlarge
       ALPINE_IMAGE: "arm64v8/alpine"
-      {%- elif "s390x" in build_environment %}
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
       {%- elif config["gpu_arch_type"] == "rocm" %}
       runs_on: linux.rocm.gpu
       {%- elif config["gpu_arch_type"] == "cuda" %}
@@ -101,7 +88,7 @@ jobs:
       {%- endif %}
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
-    {%- else %}
+{%- else %}
     runs-on: linux.rocm.gpu
     timeout-minutes: !{{ common.timeout_minutes }}
     !{{ upload.binary_env(config) }}
@@ -126,8 +113,7 @@ jobs:
         uses: ./pytorch/.github/actions/test-pytorch-binary
       - name: Teardown ROCm
         uses: ./.github/actions/teardown-rocm
-    {%- endif %}
-  {%- endif %}
+{%- endif %}
 
 {%- if branches == "nightly" %}
   !{{ upload.upload_binaries(config) }}

.github/templates/macos_binary_build_workflow.yml.j2

@@ -48,7 +48,7 @@ env:
   BUILD_ENVIRONMENT: !{{ build_environment }}
   GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
   PR_NUMBER: ${{ github.event.pull_request.number }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 {%- if cross_compile_arm64 %}
   CROSS_COMPILE_ARM64: 1
 {% endif %}

.github/templates/upload.yml.j2

@@ -57,11 +57,7 @@
       id-token: write
       contents: read
 {%- if has_test %}
-    {%- if config["gpu_arch_type"] == "cuda-aarch64" %}
-    needs: !{{ config["build_name"] }}-build
-    {%- else %}
     needs: !{{ config["build_name"] }}-test
-    {%- endif %}
 {%- else %}
     needs: !{{ config["build_name"] }}-build
 {%- endif %}

.github/workflows/_binary-build-linux.yml

@@ -12,15 +12,10 @@ on:
         type: string
         description: The build environment
       runs_on:
-        required: false
-        default: linux.12xlarge
-        type: string
-        description: Hardware to run this "build"job on, linux.12xlarge or linux.arm64.2xlarge.
-      timeout-minutes:
-        required: false
-        default: 210
-        type: number
-        description: timeout for the job
+          required: false
+          default: linux.12xlarge
+          type: string
+          description: Hardware to run this "build"job on, linux.12xlarge or linux.arm64.2xlarge.
       ALPINE_IMAGE:
         required: false
         type: string
@@ -83,7 +78,7 @@ on:
 jobs:
   build:
     runs-on: ${{ inputs.runs_on }}
-    timeout-minutes: ${{ inputs.timeout-minutes }}
+    timeout-minutes: 210
     env:
       PYTORCH_ROOT: ${{ inputs.PYTORCH_ROOT }}
       BUILDER_ROOT: ${{ inputs.BUILDER_ROOT }}
@@ -144,7 +139,6 @@ jobs:
         run: env
 
       - name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
-        if: inputs.build_environment != 'linux-s390x-binary-manywheel'
         uses: pytorch/test-infra/.github/actions/setup-ssh@main
         continue-on-error: true
         with:
@@ -153,14 +147,12 @@ jobs:
       - name: Checkout PyTorch
         uses: pytorch/pytorch/.github/actions/checkout-pytorch@main
         with:
-          no-sudo: ${{ inputs.build_environment == 'linux-aarch64-binary-manywheel' || inputs.build_environment == 'linux-s390x-binary-manywheel' }}
+          no-sudo: ${{ inputs.build_environment == 'linux-aarch64-binary-manywheel' }}
 
       - name: Setup Linux
-        if: inputs.build_environment != 'linux-s390x-binary-manywheel'
         uses: ./.github/actions/setup-linux
 
       - name: Chown workspace
-        if: inputs.build_environment != 'linux-s390x-binary-manywheel'
         uses: ./.github/actions/chown-workspace
         with:
           ALPINE_IMAGE: ${{ inputs.ALPINE_IMAGE }}
@@ -173,7 +165,7 @@ jobs:
           rm -rf "${GITHUB_WORKSPACE}"
           mkdir "${GITHUB_WORKSPACE}"
 
-          if [[ ${{ inputs.build_environment }} == 'linux-aarch64-binary-manywheel' ]] || [[ ${{ inputs.build_environment }} == 'linux-s390x-binary-manywheel' ]] ; then
+          if [[ ${{ inputs.build_environment }} == 'linux-aarch64-binary-manywheel' ]]; then
             rm -rf "${RUNNER_TEMP}/artifacts"
             mkdir "${RUNNER_TEMP}/artifacts"
           fi
@@ -220,7 +212,7 @@ jobs:
             ]}
 
       - name: Pull Docker image
-        if: ${{ steps.filter.outputs.is-test-matrix-empty == 'False' && inputs.build_environment != 'linux-s390x-binary-manywheel' }}
+        if: ${{ steps.filter.outputs.is-test-matrix-empty == 'False' }}
         uses: pytorch/test-infra/.github/actions/pull-docker-image@main
         with:
           docker-image: ${{ inputs.DOCKER_IMAGE }}
@@ -262,7 +254,7 @@ jobs:
           fi
 
       - name: Chown artifacts
-        if: ${{ steps.filter.outputs.is-test-matrix-empty == 'False' && inputs.build_environment != 'linux-s390x-binary-manywheel' }}
+        if: ${{ steps.filter.outputs.is-test-matrix-empty == 'False' }}
         shell: bash
         run: |
           # Ensure the working directory gets chowned back to the current user
@@ -277,20 +269,11 @@ jobs:
             ${{ runner.temp }}/artifacts/*
 
       - name: Teardown Linux
-        if: always() && inputs.build_environment != 'linux-s390x-binary-manywheel'
+        if: always()
         uses: pytorch/test-infra/.github/actions/teardown-linux@main
 
       - name: Chown workspace
-        if: always() && inputs.build_environment != 'linux-s390x-binary-manywheel'
+        if: always()
         uses: ./pytorch/.github/actions/chown-workspace
         with:
           ALPINE_IMAGE: ${{ inputs.ALPINE_IMAGE }}
-
-      - name: Cleanup docker
-        if: always() && inputs.build_environment == 'linux-s390x-binary-manywheel'
-        shell: bash
-        run: |
-          # on s390x stop the container for clean worker stop
-          # ignore expansion of "docker ps -q" since it could be empty
-          # shellcheck disable=SC2046
-          docker stop $(docker ps -q) || true

.github/workflows/_binary-test-linux.yml

@@ -127,7 +127,6 @@ jobs:
           } >> "${GITHUB_ENV} }}"
 
       - name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
-        if: inputs.build_environment != 'linux-s390x-binary-manywheel'
         uses: pytorch/test-infra/.github/actions/setup-ssh@main
         continue-on-error: true
         with:
@@ -137,14 +136,12 @@ jobs:
       - name: Checkout PyTorch
         uses: pytorch/pytorch/.github/actions/checkout-pytorch@main
         with:
-          no-sudo: ${{ inputs.build_environment == 'linux-aarch64-binary-manywheel' || inputs.build_environment == 'linux-s390x-binary-manywheel' }}
+          no-sudo: ${{ inputs.build_environment == 'linux-aarch64-binary-manywheel' }}
 
       - name: Setup Linux
-        if: inputs.build_environment != 'linux-s390x-binary-manywheel'
         uses: ./.github/actions/setup-linux
 
       - name: Chown workspace
-        if: inputs.build_environment != 'linux-s390x-binary-manywheel'
         uses: ./.github/actions/chown-workspace
         with:
           ALPINE_IMAGE: ${{ inputs.ALPINE_IMAGE }}
@@ -206,7 +203,7 @@ jobs:
         if: ${{ inputs.GPU_ARCH_TYPE == 'cuda' && steps.filter.outputs.is-test-matrix-empty == 'False' }}
 
       - name: Pull Docker image
-        if: ${{ steps.filter.outputs.is-test-matrix-empty == 'False' && inputs.build_environment != 'linux-s390x-binary-manywheel' }}
+        if: ${{ steps.filter.outputs.is-test-matrix-empty == 'False' }}
         uses: pytorch/test-infra/.github/actions/pull-docker-image@main
         with:
           docker-image: ${{ inputs.DOCKER_IMAGE }}
@@ -216,11 +213,11 @@ jobs:
         uses: ./pytorch/.github/actions/test-pytorch-binary
 
       - name: Teardown Linux
-        if: always() && inputs.build_environment != 'linux-s390x-binary-manywheel'
+        if: always()
         uses: pytorch/test-infra/.github/actions/teardown-linux@main
 
       - name: Chown workspace
-        if: always() && inputs.build_environment != 'linux-s390x-binary-manywheel'
+        if: always()
         uses: ./pytorch/.github/actions/chown-workspace
         with:
           ALPINE_IMAGE: ${{ inputs.ALPINE_IMAGE }}

.github/workflows/_mac-test.yml

@@ -24,6 +24,11 @@ on:
         default: "3.8"
         description: |
           The python version to be used. Will be 3.8 by default
+      arch:
+        required: true
+        type: string
+        description: |
+          Contains the architecture to run the tests with
       timeout-minutes:
         required: false
         type: number
@@ -39,7 +44,7 @@ jobs:
     # Also ensure that we always run with the right architecture
     defaults:
       run:
-        shell: bash -e -l {0}
+        shell: arch -arch ${{ inputs.arch }} bash -e -l {0}
     strategy:
       matrix: ${{ fromJSON(inputs.test-matrix) }}
       fail-fast: false
@@ -128,6 +133,12 @@ jobs:
           test-matrix: ${{ inputs.test-matrix }}
           job-name: ${{ steps.get-job-id.outputs.job-name }}
 
+      - name: Pre-process arm64 wheels
+        if: inputs.build-environment == 'macos-12-py3-arm64'
+        run: |
+          # As wheels are cross-compiled they are reported as x86_64 ones
+          ORIG_WHLNAME=$(ls -1 dist/*.whl); ARM_WHLNAME=${ORIG_WHLNAME/x86_64/arm64}; mv "${ORIG_WHLNAME}" "${ARM_WHLNAME}"
+
       - name: Set Test step time
         id: test-timeout
         shell: bash

.github/workflows/_runner-determinator.yml

@@ -1,58 +0,0 @@
-name: Check whether the workflow owner can use ARC runners
-
-on:
-  workflow_call:
-    inputs:
-      user_name:
-        required: true
-        type: string
-        description: The name of the workflow owner.
-      curr_branch:
-        required: true
-        type: string
-        description: Current branch.
-      issue_number:
-        required: false
-        type: string
-        default: "5132"
-
-    outputs:
-      workflow-type:
-        description: Type of runners to use
-        value: ${{ jobs.runner-determinator.outputs.workflow-type }}
-
-jobs:
-  runner-determinator:
-    runs-on: linux.4xlarge
-    outputs:
-      workflow-type: ${{ steps.set-condition.outputs.workflow-type }}
-    env:
-      GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      ISSUE_NUMBER: ${{ inputs.issue_number }}
-      USERNAME: ${{ inputs.user_name }}
-    steps:
-      - name: Checkout PyTorch
-        uses: pytorch/pytorch/.github/actions/checkout-pytorch@main
-        with:
-          fetch-depth: 1
-          submodules: true
-
-      - name: Install dependencies
-        run: python3 -m pip install urllib3==1.26.18 PyGithub==2.3.0
-
-      - name: Get the workflow type for the current user
-        id: set-condition
-        run: |
-          curr_branch="${{ inputs.curr_branch }}"
-          echo "Current branch is '$curr_branch'"
-
-          output="$(python3 .github/scripts/get_workflow_type.py \
-            --github-token "$GITHUB_TOKEN" \
-            --github-issue "$ISSUE_NUMBER" \
-            --github-branch "$curr_branch" \
-            --github-user "$USERNAME")"
-
-          echo "Output: '${output}'"
-
-          WORKFLOW_TYPE=$(echo "${output}" | jq -r '.workflow_type')
-          echo "workflow-type=$WORKFLOW_TYPE" >> "$GITHUB_OUTPUT"

.github/workflows/assigntome-docathon.yml

@@ -8,8 +8,6 @@ on:
 jobs:
   assign:
     runs-on: ubuntu-latest
-    permissions:
-      issues: write
     steps:
       - name: Check for "/assigntome" in comment
         uses: actions/github-script@v6
@@ -28,14 +26,14 @@ jobs:
                   repo: context.repo.repo,
                   issue_number: issueNumber
                 });
-              const hasLabel = issue.labels.some(label => label.name === 'docathon-h1-2024');
+              const hasLabel = issue.labels.some(label => label.name === 'docathon-h2-2023');
               if (hasLabel) {
                 if (issue.assignee !== null) {
                   await github.rest.issues.createComment({
                     owner: context.repo.owner,
                     repo: context.repo.repo,
                     issue_number: issueNumber,
-                    body: "The issue is already assigned. Please pick an opened and unnasigned issue with the [docathon-h1-2024 label](https://github.com/pytorch/pytorch/issues?q=is%3Aopen+is%3Aissue+label%3Adocathon-h1-2024)."
+                    body: "The issue is already assigned. Please pick an opened and unnasigned issue with the [docathon-h2-2023 label](https://github.com/pytorch/pytorch/issues?q=is%3Aopen+is%3Aissue+label%3Adocathon-h2-2023)."
                   });
                 } else {
                   await github.rest.issues.addAssignees({
@@ -46,7 +44,7 @@ jobs:
                   });
                 }
               } else {
-                const commmentMessage = "This issue does not have the correct label. Please pick an opened and unnasigned issue with the [docathon-h1-2024 label](https://github.com/pytorch/pytorch/issues?q=is%3Aopen+is%3Aissue+label%3Adocathon-h1-2024)."
+                const commmentMessage = "This issue does not have the correct label. Please pick an opened and unnasigned issue with the [docathon-h2-2023 label](https://github.com/pytorch/pytorch/issues?q=is%3Aopen+is%3Aissue+label%3Adocathon-h2-2023)."
                 await github.rest.issues.createComment({
                   owner: context.repo.owner,
                   repo: context.repo.repo,

.github/workflows/build-ios-binaries.yml

@@ -49,7 +49,7 @@ jobs:
           { config: "default",
             shard: 1,
             num_shards: 1,
-            runner: "macos-14-xlarge",
+            runner: "macos-13-xlarge",
             ios_platform: "SIMULATOR",
             ios_arch: "arm64",
             use_lite_interpreter: ${{ inputs.use_lite_interpreter || 1 }},
@@ -60,7 +60,7 @@ jobs:
           { config: "default",
             shard: 1,
             num_shards: 1,
-            runner: "macos-14-xlarge",
+            runner: "macos-13-xlarge",
             ios_platform: "OS",
             ios_arch: "arm64",
             use_lite_interpreter: ${{ inputs.use_lite_interpreter || 1 }},

.github/workflows/build-triton-wheel.yml

@@ -37,7 +37,7 @@ jobs:
         device: ["cuda", "rocm"]
         include:
           - device: "rocm"
-            rocm_version: "6.1"
+            rocm_version: "6.0"
           - device: "cuda"
             rocm_version: ""
     timeout-minutes: 40

.github/workflows/close-nonexistent-disable-issues.yml

@@ -18,6 +18,6 @@ jobs:
           ROCKSET_API_KEY: ${{ secrets.ROCKSET_API_KEY }}
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
-          pip3 install requests==2.32.2
+          pip3 install requests==2.26
           pip3 install rockset==1.0.3
           python3 .github/scripts/close_nonexistent_disable_issues.py

.github/workflows/delete_old_branches.yml

@@ -29,7 +29,7 @@ jobs:
       - name: Setup Python
         uses: actions/setup-python@v4
         with:
-          python-version: '3.11'
+          python-version: '3.8'
           architecture: x64
           check-latest: false
 

.github/workflows/docathon-sync-label.yml

@@ -1,30 +0,0 @@
-name: Docathon Labels Sync
-
-on:
-  pull_request_target:
-    types: [opened, synchronize, edited]
-    branches: [main]
-
-jobs:
-  check-labels:
-    runs-on: ubuntu-latest
-    permissions:
-      issues: write
-      pull-requests: write
-    steps:
-      - name: Check out the repo
-        uses: actions/checkout@v2
-        with:
-          fetch-depth: 1
-      - name: Set up Python
-        uses: actions/setup-python@v2
-        with:
-          python-version: 3.x
-      - name: Install dependencies
-        run: |
-          pip install requests==2.32.3
-          pip install PyGithub==2.3.0
-      - name: Run Python script
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: python ./.github/scripts/docathon-label-sync.py ${{ github.event.pull_request.number }}

.github/workflows/docker-builds.yml

@@ -38,19 +38,15 @@ jobs:
       matrix:
         runner: [linux.12xlarge]
         docker-image-name: [
-          pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9,
-          pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks,
-          pytorch-linux-focal-cuda12.4-cudnn9-py3.12-gcc9-inductor-benchmarks,
-          pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9,
-          pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks,
-          pytorch-linux-focal-cuda12.1-cudnn9-py3.12-gcc9-inductor-benchmarks,
-          pytorch-linux-focal-cuda11.8-cudnn9-py3-gcc9,
+          pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9,
+          pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9-inductor-benchmarks,
+          pytorch-linux-focal-cuda11.8-cudnn8-py3-gcc9,
           pytorch-linux-focal-py3.8-clang10,
           pytorch-linux-focal-py3.11-clang10,
           pytorch-linux-focal-py3.12-clang10,
           pytorch-linux-focal-rocm-n-1-py3,
           pytorch-linux-focal-rocm-n-py3,
-          pytorch-linux-jammy-cuda11.8-cudnn9-py3.8-clang12,
+          pytorch-linux-jammy-cuda11.8-cudnn8-py3.8-clang12,
           pytorch-linux-focal-py3-clang9-android-ndk-r21e,
           pytorch-linux-jammy-py3.8-gcc11,
           pytorch-linux-jammy-py3.8-gcc11-inductor-benchmarks,
@@ -58,7 +54,7 @@ jobs:
           pytorch-linux-jammy-py3-clang15-asan,
           pytorch-linux-focal-py3-clang10-onnx,
           pytorch-linux-focal-linter,
-          pytorch-linux-jammy-cuda11.8-cudnn9-py3.9-linter,
+          pytorch-linux-jammy-cuda11.8-cudnn8-py3.9-linter,
           pytorch-linux-jammy-py3-clang12-executorch
           ]
         include:

.github/workflows/docker-release.yml

@@ -7,7 +7,6 @@ on:
       - Dockerfile
       - docker.Makefile
       - .github/workflows/docker-release.yml
-      - .github/scripts/generate_docker_release_matrix.py
   push:
     branches:
       - nightly
@@ -127,32 +126,20 @@ jobs:
         run: |
           make -f docker.Makefile "${BUILD_IMAGE_TYPE}-image"
       - name: Push nightly tags
-        if: ${{ github.event.ref == 'refs/heads/nightly' && matrix.image_type == 'runtime' && matrix.build_platforms == 'linux/amd4' }}
+        if: ${{ github.event.ref == 'refs/heads/nightly' && matrix.image_type == 'runtime' }}
         run: |
           PYTORCH_DOCKER_TAG="${PYTORCH_VERSION}-cuda${CUDA_VERSION_SHORT}-cudnn${CUDNN_VERSION}-runtime"
-          CUDA_SUFFIX="-cu${CUDA_VERSION}"
+
           PYTORCH_NIGHTLY_COMMIT=$(docker run ghcr.io/pytorch/pytorch-nightly:"${PYTORCH_DOCKER_TAG}" \
                                           python -c 'import torch; print(torch.version.git_version[:7],end="")')
 
           docker tag ghcr.io/pytorch/pytorch-nightly:"${PYTORCH_DOCKER_TAG}" \
-                 ghcr.io/pytorch/pytorch-nightly:"${PYTORCH_NIGHTLY_COMMIT}${CUDA_SUFFIX}"
-
-          docker push ghcr.io/pytorch/pytorch-nightly:"${PYTORCH_NIGHTLY_COMMIT}${CUDA_SUFFIX}"
-
-          # Please note, here we ned to pin specific verison of CUDA as with latest label
-          if [[ ${CUDA_VERSION_SHORT} == "12.1" ]]; then
-            docker tag ghcr.io/pytorch/pytorch-nightly:"${PYTORCH_NIGHTLY_COMMIT}${CUDA_SUFFIX}" \
-                    ghcr.io/pytorch/pytorch-nightly:latest
-            docker push ghcr.io/pytorch/pytorch-nightly:latest
-          fi
+                 ghcr.io/pytorch/pytorch-nightly:"${PYTORCH_NIGHTLY_COMMIT}-cu${CUDA_VERSION}"
+          docker push ghcr.io/pytorch/pytorch-nightly:"${PYTORCH_NIGHTLY_COMMIT}-cu${CUDA_VERSION}"
 
+          docker tag ghcr.io/pytorch/pytorch-nightly:"${PYTORCH_NIGHTLY_COMMIT}-cu${CUDA_VERSION}" \
+                 ghcr.io/pytorch/pytorch-nightly:latest
+          docker push ghcr.io/pytorch/pytorch-nightly:latest
       - name: Teardown Linux
         uses: pytorch/test-infra/.github/actions/teardown-linux@main
         if: always()
-
-  validate:
-    needs: build
-    uses: pytorch/builder/.github/workflows/validate-docker-images.yml@main
-    with:
-      channel: nightly
-      ref: main

.github/workflows/generated-linux-aarch64-binary-manywheel-nightly.yml

@@ -31,7 +31,7 @@ env:
   PYTORCH_FINAL_PACKAGE_DIR: /artifacts
   PYTORCH_ROOT: /pytorch
   SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: linux-aarch64-binary-manywheel-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true
@@ -50,11 +50,11 @@ jobs:
       GPU_ARCH_TYPE: cpu-aarch64
       DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cpu-aarch64-main
       DESIRED_PYTHON: "3.8"
-      runs_on: linux.arm64.m7g.4xlarge
+      runs_on: linux.arm64.2xlarge
       ALPINE_IMAGE: "arm64v8/alpine"
       build_name: manywheel-py3_8-cpu-aarch64
       build_environment: linux-aarch64-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
   manywheel-py3_8-cpu-aarch64-test:  # Testing
@@ -100,51 +100,6 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  manywheel-py3_8-cuda-aarch64-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.8"
-      runs_on: linux.arm64.m7g.4xlarge
-      ALPINE_IMAGE: "arm64v8/alpine"
-      build_name: manywheel-py3_8-cuda-aarch64
-      build_environment: linux-aarch64-binary-manywheel
-      timeout-minutes: 420
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_8-cuda-aarch64-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_8-cuda-aarch64-build
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.8"
-      build_name: manywheel-py3_8-cuda-aarch64
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
   manywheel-py3_9-cpu-aarch64-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
@@ -158,11 +113,11 @@ jobs:
       GPU_ARCH_TYPE: cpu-aarch64
       DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cpu-aarch64-main
       DESIRED_PYTHON: "3.9"
-      runs_on: linux.arm64.m7g.4xlarge
+      runs_on: linux.arm64.2xlarge
       ALPINE_IMAGE: "arm64v8/alpine"
       build_name: manywheel-py3_9-cpu-aarch64
       build_environment: linux-aarch64-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
   manywheel-py3_9-cpu-aarch64-test:  # Testing
@@ -208,51 +163,6 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  manywheel-py3_9-cuda-aarch64-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.9"
-      runs_on: linux.arm64.m7g.4xlarge
-      ALPINE_IMAGE: "arm64v8/alpine"
-      build_name: manywheel-py3_9-cuda-aarch64
-      build_environment: linux-aarch64-binary-manywheel
-      timeout-minutes: 420
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_9-cuda-aarch64-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_9-cuda-aarch64-build
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.9"
-      build_name: manywheel-py3_9-cuda-aarch64
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
   manywheel-py3_10-cpu-aarch64-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
@@ -266,11 +176,11 @@ jobs:
       GPU_ARCH_TYPE: cpu-aarch64
       DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cpu-aarch64-main
       DESIRED_PYTHON: "3.10"
-      runs_on: linux.arm64.m7g.4xlarge
+      runs_on: linux.arm64.2xlarge
       ALPINE_IMAGE: "arm64v8/alpine"
       build_name: manywheel-py3_10-cpu-aarch64
       build_environment: linux-aarch64-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
   manywheel-py3_10-cpu-aarch64-test:  # Testing
@@ -316,51 +226,6 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  manywheel-py3_10-cuda-aarch64-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.10"
-      runs_on: linux.arm64.m7g.4xlarge
-      ALPINE_IMAGE: "arm64v8/alpine"
-      build_name: manywheel-py3_10-cuda-aarch64
-      build_environment: linux-aarch64-binary-manywheel
-      timeout-minutes: 420
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_10-cuda-aarch64-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_10-cuda-aarch64-build
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.10"
-      build_name: manywheel-py3_10-cuda-aarch64
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
   manywheel-py3_11-cpu-aarch64-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
@@ -374,11 +239,11 @@ jobs:
       GPU_ARCH_TYPE: cpu-aarch64
       DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cpu-aarch64-main
       DESIRED_PYTHON: "3.11"
-      runs_on: linux.arm64.m7g.4xlarge
+      runs_on: linux.arm64.2xlarge
       ALPINE_IMAGE: "arm64v8/alpine"
       build_name: manywheel-py3_11-cpu-aarch64
       build_environment: linux-aarch64-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
   manywheel-py3_11-cpu-aarch64-test:  # Testing
@@ -424,51 +289,6 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  manywheel-py3_11-cuda-aarch64-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.11"
-      runs_on: linux.arm64.m7g.4xlarge
-      ALPINE_IMAGE: "arm64v8/alpine"
-      build_name: manywheel-py3_11-cuda-aarch64
-      build_environment: linux-aarch64-binary-manywheel
-      timeout-minutes: 420
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_11-cuda-aarch64-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_11-cuda-aarch64-build
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.11"
-      build_name: manywheel-py3_11-cuda-aarch64
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
   manywheel-py3_12-cpu-aarch64-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
@@ -482,11 +302,11 @@ jobs:
       GPU_ARCH_TYPE: cpu-aarch64
       DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cpu-aarch64-main
       DESIRED_PYTHON: "3.12"
-      runs_on: linux.arm64.m7g.4xlarge
+      runs_on: linux.arm64.2xlarge
       ALPINE_IMAGE: "arm64v8/alpine"
       build_name: manywheel-py3_12-cpu-aarch64
       build_environment: linux-aarch64-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
   manywheel-py3_12-cpu-aarch64-test:  # Testing
@@ -531,48 +351,3 @@ jobs:
       conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
-
-  manywheel-py3_12-cuda-aarch64-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.12"
-      runs_on: linux.arm64.m7g.4xlarge
-      ALPINE_IMAGE: "arm64v8/alpine"
-      build_name: manywheel-py3_12-cuda-aarch64
-      build_environment: linux-aarch64-binary-manywheel
-      timeout-minutes: 420
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_12-cuda-aarch64-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_12-cuda-aarch64-build
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_TYPE: cuda-aarch64
-      DOCKER_IMAGE: pytorch/manylinuxaarch64-builder:cuda12.4-main
-      DESIRED_DEVTOOLSET: cxx11-abi
-      DESIRED_PYTHON: "3.12"
-      build_name: manywheel-py3_12-cuda-aarch64
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml

.github/workflows/generated-linux-binary-conda-nightly.yml

@@ -31,7 +31,7 @@ env:
   PYTORCH_FINAL_PACKAGE_DIR: /artifacts
   PYTORCH_ROOT: /pytorch
   SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: linux-binary-conda-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true
@@ -222,69 +222,6 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  conda-py3_8-cuda12_4-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.8"
-      runs_on: linux.24xlarge
-      build_name: conda-py3_8-cuda12_4
-      build_environment: linux-binary-conda
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_8-cuda12_4-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: conda-py3_8-cuda12_4-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.8"
-      build_name: conda-py3_8-cuda12_4
-      build_environment: linux-binary-conda
-      runs_on: linux.4xlarge.nvidia.gpu
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_8-cuda12_4-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: conda-py3_8-cuda12_4-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.8"
-      build_name: conda-py3_8-cuda12_4
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
   conda-py3_9-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
@@ -470,69 +407,6 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  conda-py3_9-cuda12_4-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.9"
-      runs_on: linux.24xlarge
-      build_name: conda-py3_9-cuda12_4
-      build_environment: linux-binary-conda
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_9-cuda12_4-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: conda-py3_9-cuda12_4-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.9"
-      build_name: conda-py3_9-cuda12_4
-      build_environment: linux-binary-conda
-      runs_on: linux.4xlarge.nvidia.gpu
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_9-cuda12_4-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: conda-py3_9-cuda12_4-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.9"
-      build_name: conda-py3_9-cuda12_4
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
   conda-py3_10-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
@@ -718,69 +592,6 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  conda-py3_10-cuda12_4-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.10"
-      runs_on: linux.24xlarge
-      build_name: conda-py3_10-cuda12_4
-      build_environment: linux-binary-conda
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_10-cuda12_4-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: conda-py3_10-cuda12_4-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.10"
-      build_name: conda-py3_10-cuda12_4
-      build_environment: linux-binary-conda
-      runs_on: linux.4xlarge.nvidia.gpu
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_10-cuda12_4-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: conda-py3_10-cuda12_4-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.10"
-      build_name: conda-py3_10-cuda12_4
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
   conda-py3_11-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
@@ -966,69 +777,6 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  conda-py3_11-cuda12_4-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.11"
-      runs_on: linux.24xlarge
-      build_name: conda-py3_11-cuda12_4
-      build_environment: linux-binary-conda
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_11-cuda12_4-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: conda-py3_11-cuda12_4-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.11"
-      build_name: conda-py3_11-cuda12_4
-      build_environment: linux-binary-conda
-      runs_on: linux.4xlarge.nvidia.gpu
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_11-cuda12_4-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: conda-py3_11-cuda12_4-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.11"
-      build_name: conda-py3_11-cuda12_4
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
   conda-py3_12-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
@@ -1213,66 +961,3 @@ jobs:
       conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
-
-  conda-py3_12-cuda12_4-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.12"
-      runs_on: linux.24xlarge
-      build_name: conda-py3_12-cuda12_4
-      build_environment: linux-binary-conda
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_12-cuda12_4-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: conda-py3_12-cuda12_4-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.12"
-      build_name: conda-py3_12-cuda12_4
-      build_environment: linux-binary-conda
-      runs_on: linux.4xlarge.nvidia.gpu
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  conda-py3_12-cuda12_4-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: conda-py3_12-cuda12_4-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: conda
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/conda-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.12"
-      build_name: conda-py3_12-cuda12_4
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml

.github/workflows/generated-linux-binary-libtorch-cxx11-abi-main.yml

@@ -26,7 +26,7 @@ env:
   PYTORCH_FINAL_PACKAGE_DIR: /artifacts
   PYTORCH_ROOT: /pytorch
   SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: linux-binary-libtorch-cxx11-abi-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true

.github/workflows/generated-linux-binary-libtorch-cxx11-abi-nightly.yml

@@ -31,7 +31,7 @@ env:
   PYTORCH_FINAL_PACKAGE_DIR: /artifacts
   PYTORCH_ROOT: /pytorch
   SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: linux-binary-libtorch-cxx11-abi-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true
@@ -229,7 +229,7 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  libtorch-cuda12_4-shared-with-deps-cxx11-abi-build:
+  libtorch-rocm5_7-shared-with-deps-cxx11-abi-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
     with:
@@ -238,56 +238,97 @@ jobs:
       PACKAGE_TYPE: libtorch
       # TODO: This is a legacy variable that we eventually want to get rid of in
       #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:cuda12.4-main
+      DESIRED_CUDA: rocm5.7
+      GPU_ARCH_VERSION: 5.7
+      GPU_ARCH_TYPE: rocm
+      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:rocm5.7-main
       LIBTORCH_VARIANT: shared-with-deps
       DESIRED_DEVTOOLSET: cxx11-abi
-      build_name: libtorch-cuda12_4-shared-with-deps-cxx11-abi
+      build_name: libtorch-rocm5_7-shared-with-deps-cxx11-abi
       build_environment: linux-binary-libtorch-cxx11-abi
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
-  libtorch-cuda12_4-shared-with-deps-cxx11-abi-test:  # Testing
+  libtorch-rocm5_7-shared-with-deps-cxx11-abi-test:  # Testing
     if: ${{ github.repository_owner == 'pytorch' }}
-    needs: libtorch-cuda12_4-shared-with-deps-cxx11-abi-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
+    needs: libtorch-rocm5_7-shared-with-deps-cxx11-abi-build
+    runs-on: linux.rocm.gpu
+    timeout-minutes: 240
+    env:
       PYTORCH_ROOT: /pytorch
       BUILDER_ROOT: /builder
       PACKAGE_TYPE: libtorch
       # TODO: This is a legacy variable that we eventually want to get rid of in
       #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:cuda12.4-main
+      DESIRED_CUDA: rocm5.7
+      GPU_ARCH_VERSION: 5.7
+      GPU_ARCH_TYPE: rocm
+      SKIP_ALL_TESTS: 1
+      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:rocm5.7-main
       LIBTORCH_VARIANT: shared-with-deps
       DESIRED_DEVTOOLSET: cxx11-abi
-      build_name: libtorch-cuda12_4-shared-with-deps-cxx11-abi
-      build_environment: linux-binary-libtorch-cxx11-abi
-      runs_on: linux.4xlarge.nvidia.gpu
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  libtorch-cuda12_4-shared-with-deps-cxx11-abi-upload:  # Uploading
+    steps:
+      - name: Setup ROCm
+        uses: ./.github/actions/setup-rocm
+      - uses: actions/download-artifact@v3
+        name: Download Build Artifacts
+        with:
+          name: libtorch-rocm5_7-shared-with-deps-cxx11-abi
+          path: "${{ runner.temp }}/artifacts/"
+      - name: Checkout PyTorch
+        uses: malfet/checkout@silent-checkout
+        with:
+          ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
+          submodules: recursive
+          path: pytorch
+          quiet-checkout: true
+      - name: Clean PyTorch checkout
+        run: |
+          # Remove any artifacts from the previous checkouts
+          git clean -fxd
+        working-directory: pytorch
+      - name: Checkout pytorch/builder
+        uses: malfet/checkout@silent-checkout
+        with:
+          ref: main
+          submodules: recursive
+          repository: pytorch/builder
+          path: builder
+          quiet-checkout: true
+      - name: Clean pytorch/builder checkout
+        run: |
+          # Remove any artifacts from the previous checkouts
+          git clean -fxd
+        working-directory: builder
+      - name: ROCm set GPU_FLAG
+        run: |
+          echo "GPU_FLAG=--device=/dev/mem --device=/dev/kfd --device=/dev/dri --group-add video --group-add daemon" >> "${GITHUB_ENV}"
+      - name: Pull Docker image
+        uses: pytorch/test-infra/.github/actions/pull-docker-image@main
+        with:
+          docker-image: pytorch/libtorch-cxx11-builder:rocm5.7-main
+      - name: Test Pytorch binary
+        uses: ./pytorch/.github/actions/test-pytorch-binary
+      - name: Teardown ROCm
+        uses: ./.github/actions/teardown-rocm
+  libtorch-rocm5_7-shared-with-deps-cxx11-abi-upload:  # Uploading
     if: ${{ github.repository_owner == 'pytorch' }}
     permissions:
       id-token: write
       contents: read
-    needs: libtorch-cuda12_4-shared-with-deps-cxx11-abi-test
+    needs: libtorch-rocm5_7-shared-with-deps-cxx11-abi-test
     with:
       PYTORCH_ROOT: /pytorch
       BUILDER_ROOT: /builder
       PACKAGE_TYPE: libtorch
       # TODO: This is a legacy variable that we eventually want to get rid of in
       #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:cuda12.4-main
+      DESIRED_CUDA: rocm5.7
+      GPU_ARCH_VERSION: 5.7
+      GPU_ARCH_TYPE: rocm
+      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:rocm5.7-main
       LIBTORCH_VARIANT: shared-with-deps
       DESIRED_DEVTOOLSET: cxx11-abi
-      build_name: libtorch-cuda12_4-shared-with-deps-cxx11-abi
+      build_name: libtorch-rocm5_7-shared-with-deps-cxx11-abi
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
       conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
@@ -399,109 +440,3 @@ jobs:
       conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
-
-  libtorch-rocm6_1-shared-with-deps-cxx11-abi-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: rocm6.1
-      GPU_ARCH_VERSION: 6.1
-      GPU_ARCH_TYPE: rocm
-      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:rocm6.1-main
-      LIBTORCH_VARIANT: shared-with-deps
-      DESIRED_DEVTOOLSET: cxx11-abi
-      build_name: libtorch-rocm6_1-shared-with-deps-cxx11-abi
-      build_environment: linux-binary-libtorch-cxx11-abi
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  libtorch-rocm6_1-shared-with-deps-cxx11-abi-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: libtorch-rocm6_1-shared-with-deps-cxx11-abi-build
-    runs-on: linux.rocm.gpu
-    timeout-minutes: 240
-    env:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: rocm6.1
-      GPU_ARCH_VERSION: 6.1
-      GPU_ARCH_TYPE: rocm
-      SKIP_ALL_TESTS: 1
-      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:rocm6.1-main
-      LIBTORCH_VARIANT: shared-with-deps
-      DESIRED_DEVTOOLSET: cxx11-abi
-    steps:
-      - name: Setup ROCm
-        uses: ./.github/actions/setup-rocm
-      - uses: actions/download-artifact@v3
-        name: Download Build Artifacts
-        with:
-          name: libtorch-rocm6_1-shared-with-deps-cxx11-abi
-          path: "${{ runner.temp }}/artifacts/"
-      - name: Checkout PyTorch
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
-          submodules: recursive
-          path: pytorch
-          quiet-checkout: true
-      - name: Clean PyTorch checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: pytorch
-      - name: Checkout pytorch/builder
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: main
-          submodules: recursive
-          repository: pytorch/builder
-          path: builder
-          quiet-checkout: true
-      - name: Clean pytorch/builder checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: builder
-      - name: ROCm set GPU_FLAG
-        run: |
-          echo "GPU_FLAG=--device=/dev/mem --device=/dev/kfd --device=/dev/dri --group-add video --group-add daemon" >> "${GITHUB_ENV}"
-      - name: Pull Docker image
-        uses: pytorch/test-infra/.github/actions/pull-docker-image@main
-        with:
-          docker-image: pytorch/libtorch-cxx11-builder:rocm6.1-main
-      - name: Test Pytorch binary
-        uses: ./pytorch/.github/actions/test-pytorch-binary
-      - name: Teardown ROCm
-        uses: ./.github/actions/teardown-rocm
-  libtorch-rocm6_1-shared-with-deps-cxx11-abi-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: libtorch-rocm6_1-shared-with-deps-cxx11-abi-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: rocm6.1
-      GPU_ARCH_VERSION: 6.1
-      GPU_ARCH_TYPE: rocm
-      DOCKER_IMAGE: pytorch/libtorch-cxx11-builder:rocm6.1-main
-      LIBTORCH_VARIANT: shared-with-deps
-      DESIRED_DEVTOOLSET: cxx11-abi
-      build_name: libtorch-rocm6_1-shared-with-deps-cxx11-abi
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml

.github/workflows/generated-linux-binary-libtorch-pre-cxx11-main.yml

@@ -26,7 +26,7 @@ env:
   PYTORCH_FINAL_PACKAGE_DIR: /artifacts
   PYTORCH_ROOT: /pytorch
   SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: linux-binary-libtorch-pre-cxx11-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true

.github/workflows/generated-linux-binary-libtorch-pre-cxx11-nightly.yml

@@ -31,7 +31,7 @@ env:
   PYTORCH_FINAL_PACKAGE_DIR: /artifacts
   PYTORCH_ROOT: /pytorch
   SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: linux-binary-libtorch-pre-cxx11-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true
@@ -229,7 +229,7 @@ jobs:
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
 
-  libtorch-cuda12_4-shared-with-deps-pre-cxx11-build:
+  libtorch-rocm5_7-shared-with-deps-pre-cxx11-build:
     if: ${{ github.repository_owner == 'pytorch' }}
     uses: ./.github/workflows/_binary-build-linux.yml
     with:
@@ -238,56 +238,97 @@ jobs:
       PACKAGE_TYPE: libtorch
       # TODO: This is a legacy variable that we eventually want to get rid of in
       #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/manylinux-builder:cuda12.4-main
+      DESIRED_CUDA: rocm5.7
+      GPU_ARCH_VERSION: 5.7
+      GPU_ARCH_TYPE: rocm
+      DOCKER_IMAGE: pytorch/manylinux-builder:rocm5.7-main
       LIBTORCH_VARIANT: shared-with-deps
       DESIRED_DEVTOOLSET: pre-cxx11
-      build_name: libtorch-cuda12_4-shared-with-deps-pre-cxx11
+      build_name: libtorch-rocm5_7-shared-with-deps-pre-cxx11
       build_environment: linux-binary-libtorch-pre-cxx11
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
-  libtorch-cuda12_4-shared-with-deps-pre-cxx11-test:  # Testing
+  libtorch-rocm5_7-shared-with-deps-pre-cxx11-test:  # Testing
     if: ${{ github.repository_owner == 'pytorch' }}
-    needs: libtorch-cuda12_4-shared-with-deps-pre-cxx11-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
+    needs: libtorch-rocm5_7-shared-with-deps-pre-cxx11-build
+    runs-on: linux.rocm.gpu
+    timeout-minutes: 240
+    env:
       PYTORCH_ROOT: /pytorch
       BUILDER_ROOT: /builder
       PACKAGE_TYPE: libtorch
       # TODO: This is a legacy variable that we eventually want to get rid of in
       #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/manylinux-builder:cuda12.4-main
+      DESIRED_CUDA: rocm5.7
+      GPU_ARCH_VERSION: 5.7
+      GPU_ARCH_TYPE: rocm
+      SKIP_ALL_TESTS: 1
+      DOCKER_IMAGE: pytorch/manylinux-builder:rocm5.7-main
       LIBTORCH_VARIANT: shared-with-deps
       DESIRED_DEVTOOLSET: pre-cxx11
-      build_name: libtorch-cuda12_4-shared-with-deps-pre-cxx11
-      build_environment: linux-binary-libtorch-pre-cxx11
-      runs_on: linux.4xlarge.nvidia.gpu
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  libtorch-cuda12_4-shared-with-deps-pre-cxx11-upload:  # Uploading
+    steps:
+      - name: Setup ROCm
+        uses: ./.github/actions/setup-rocm
+      - uses: actions/download-artifact@v3
+        name: Download Build Artifacts
+        with:
+          name: libtorch-rocm5_7-shared-with-deps-pre-cxx11
+          path: "${{ runner.temp }}/artifacts/"
+      - name: Checkout PyTorch
+        uses: malfet/checkout@silent-checkout
+        with:
+          ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
+          submodules: recursive
+          path: pytorch
+          quiet-checkout: true
+      - name: Clean PyTorch checkout
+        run: |
+          # Remove any artifacts from the previous checkouts
+          git clean -fxd
+        working-directory: pytorch
+      - name: Checkout pytorch/builder
+        uses: malfet/checkout@silent-checkout
+        with:
+          ref: main
+          submodules: recursive
+          repository: pytorch/builder
+          path: builder
+          quiet-checkout: true
+      - name: Clean pytorch/builder checkout
+        run: |
+          # Remove any artifacts from the previous checkouts
+          git clean -fxd
+        working-directory: builder
+      - name: ROCm set GPU_FLAG
+        run: |
+          echo "GPU_FLAG=--device=/dev/mem --device=/dev/kfd --device=/dev/dri --group-add video --group-add daemon" >> "${GITHUB_ENV}"
+      - name: Pull Docker image
+        uses: pytorch/test-infra/.github/actions/pull-docker-image@main
+        with:
+          docker-image: pytorch/manylinux-builder:rocm5.7-main
+      - name: Test Pytorch binary
+        uses: ./pytorch/.github/actions/test-pytorch-binary
+      - name: Teardown ROCm
+        uses: ./.github/actions/teardown-rocm
+  libtorch-rocm5_7-shared-with-deps-pre-cxx11-upload:  # Uploading
     if: ${{ github.repository_owner == 'pytorch' }}
     permissions:
       id-token: write
       contents: read
-    needs: libtorch-cuda12_4-shared-with-deps-pre-cxx11-test
+    needs: libtorch-rocm5_7-shared-with-deps-pre-cxx11-test
     with:
       PYTORCH_ROOT: /pytorch
       BUILDER_ROOT: /builder
       PACKAGE_TYPE: libtorch
       # TODO: This is a legacy variable that we eventually want to get rid of in
       #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/manylinux-builder:cuda12.4-main
+      DESIRED_CUDA: rocm5.7
+      GPU_ARCH_VERSION: 5.7
+      GPU_ARCH_TYPE: rocm
+      DOCKER_IMAGE: pytorch/manylinux-builder:rocm5.7-main
       LIBTORCH_VARIANT: shared-with-deps
       DESIRED_DEVTOOLSET: pre-cxx11
-      build_name: libtorch-cuda12_4-shared-with-deps-pre-cxx11
+      build_name: libtorch-rocm5_7-shared-with-deps-pre-cxx11
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
       conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
@@ -399,109 +440,3 @@ jobs:
       conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
-
-  libtorch-rocm6_1-shared-with-deps-pre-cxx11-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: rocm6.1
-      GPU_ARCH_VERSION: 6.1
-      GPU_ARCH_TYPE: rocm
-      DOCKER_IMAGE: pytorch/manylinux-builder:rocm6.1-main
-      LIBTORCH_VARIANT: shared-with-deps
-      DESIRED_DEVTOOLSET: pre-cxx11
-      build_name: libtorch-rocm6_1-shared-with-deps-pre-cxx11
-      build_environment: linux-binary-libtorch-pre-cxx11
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  libtorch-rocm6_1-shared-with-deps-pre-cxx11-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: libtorch-rocm6_1-shared-with-deps-pre-cxx11-build
-    runs-on: linux.rocm.gpu
-    timeout-minutes: 240
-    env:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: rocm6.1
-      GPU_ARCH_VERSION: 6.1
-      GPU_ARCH_TYPE: rocm
-      SKIP_ALL_TESTS: 1
-      DOCKER_IMAGE: pytorch/manylinux-builder:rocm6.1-main
-      LIBTORCH_VARIANT: shared-with-deps
-      DESIRED_DEVTOOLSET: pre-cxx11
-    steps:
-      - name: Setup ROCm
-        uses: ./.github/actions/setup-rocm
-      - uses: actions/download-artifact@v3
-        name: Download Build Artifacts
-        with:
-          name: libtorch-rocm6_1-shared-with-deps-pre-cxx11
-          path: "${{ runner.temp }}/artifacts/"
-      - name: Checkout PyTorch
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
-          submodules: recursive
-          path: pytorch
-          quiet-checkout: true
-      - name: Clean PyTorch checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: pytorch
-      - name: Checkout pytorch/builder
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: main
-          submodules: recursive
-          repository: pytorch/builder
-          path: builder
-          quiet-checkout: true
-      - name: Clean pytorch/builder checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: builder
-      - name: ROCm set GPU_FLAG
-        run: |
-          echo "GPU_FLAG=--device=/dev/mem --device=/dev/kfd --device=/dev/dri --group-add video --group-add daemon" >> "${GITHUB_ENV}"
-      - name: Pull Docker image
-        uses: pytorch/test-infra/.github/actions/pull-docker-image@main
-        with:
-          docker-image: pytorch/manylinux-builder:rocm6.1-main
-      - name: Test Pytorch binary
-        uses: ./pytorch/.github/actions/test-pytorch-binary
-      - name: Teardown ROCm
-        uses: ./.github/actions/teardown-rocm
-  libtorch-rocm6_1-shared-with-deps-pre-cxx11-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: libtorch-rocm6_1-shared-with-deps-pre-cxx11-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: rocm6.1
-      GPU_ARCH_VERSION: 6.1
-      GPU_ARCH_TYPE: rocm
-      DOCKER_IMAGE: pytorch/manylinux-builder:rocm6.1-main
-      LIBTORCH_VARIANT: shared-with-deps
-      DESIRED_DEVTOOLSET: pre-cxx11
-      build_name: libtorch-rocm6_1-shared-with-deps-pre-cxx11
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml

.github/workflows/generated-linux-binary-manywheel-main.yml

@@ -26,7 +26,7 @@ env:
   PYTORCH_FINAL_PACKAGE_DIR: /artifacts
   PYTORCH_ROOT: /pytorch
   SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: linux-binary-manywheel-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true
@@ -48,7 +48,7 @@ jobs:
       DESIRED_PYTHON: "3.8"
       build_name: manywheel-py3_8-cuda11_8
       build_environment: linux-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu11==11.8.89; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu11==11.8.89; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu11==11.8.87; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu11==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu11==11.11.3.6; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu11==10.9.0.58; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu11==10.3.0.86; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu11==11.4.1.48; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu11==11.7.5.86; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu11==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu11==11.8.86; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu11==11.8.89; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu11==11.8.89; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu11==11.8.87; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu11==8.7.0.84; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu11==11.11.3.6; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu11==10.9.0.58; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu11==10.3.0.86; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu11==11.4.1.48; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu11==11.7.5.86; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu11==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu11==11.8.86; platform_system == 'Linux' and platform_machine == 'x86_64'
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
   manywheel-py3_8-cuda11_8-test:  # Testing
@@ -88,7 +88,7 @@ jobs:
       DESIRED_PYTHON: "3.8"
       build_name: manywheel-py3_8-cuda12_1
       build_environment: linux-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
   manywheel-py3_8-cuda12_1-test:  # Testing
@@ -111,43 +111,3 @@ jobs:
       runs_on: linux.4xlarge.nvidia.gpu
     secrets:
       github-token: ${{ secrets.GITHUB_TOKEN }}
-
-  manywheel-py3_8-cuda12_4-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/manylinux-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.8"
-      build_name: manywheel-py3_8-cuda12_4
-      build_environment: linux-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.4.2.65; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.2.0.44; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.5.119; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.6.0.99; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.3.0.142; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvjitlink-cu12==12.4.99; platform_system == 'Linux' and platform_machine == 'x86_64'
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_8-cuda12_4-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: manywheel-py3_8-cuda12_4-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      DOCKER_IMAGE: pytorch/manylinux-builder:cuda12.4-main
-      DESIRED_PYTHON: "3.8"
-      build_name: manywheel-py3_8-cuda12_4
-      build_environment: linux-binary-manywheel
-      runs_on: linux.4xlarge.nvidia.gpu
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/generated-linux-s390x-binary-manywheel-nightly.yml

@@ -1,353 +0,0 @@
-# @generated DO NOT EDIT MANUALLY
-
-# Template is at:    .github/templates/linux_binary_build_workflow.yml.j2
-# Generation script: .github/scripts/generate_ci_workflows.py
-name: linux-s390x-binary-manywheel
-
-
-on:
-  push:
-    # NOTE: Meta Employees can trigger new nightlies using: https://fburl.com/trigger_pytorch_nightly_build
-    branches:
-      - nightly
-    tags:
-      # NOTE: Binary build pipelines should only get triggered on release candidate builds
-      # Release candidate tags look like: v1.11.0-rc1
-      - v[0-9]+.[0-9]+.[0-9]+-rc[0-9]+
-      - 'ciflow/binaries/*'
-      - 'ciflow/binaries_wheel/*'
-  workflow_dispatch:
-
-env:
-  # Needed for conda builds
-  ALPINE_IMAGE: "docker.io/s390x/alpine"
-  ANACONDA_USER: pytorch
-  AWS_DEFAULT_REGION: us-east-1
-  BINARY_ENV_FILE: /tmp/env
-  BUILD_ENVIRONMENT: linux-s390x-binary-manywheel
-  BUILDER_ROOT: /builder
-  GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-  PR_NUMBER: ${{ github.event.pull_request.number }}
-  PYTORCH_FINAL_PACKAGE_DIR: /artifacts
-  PYTORCH_ROOT: /pytorch
-  SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
-  SKIP_ALL_TESTS: 0
-concurrency:
-  group: linux-s390x-binary-manywheel-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
-  cancel-in-progress: true
-
-jobs:
-  manywheel-py3_8-cpu-s390x-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.8"
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-      build_name: manywheel-py3_8-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_8-cpu-s390x-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: manywheel-py3_8-cpu-s390x-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.8"
-      build_name: manywheel-py3_8-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_8-cpu-s390x-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_8-cpu-s390x-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.8"
-      build_name: manywheel-py3_8-cpu-s390x
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
-  manywheel-py3_9-cpu-s390x-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.9"
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-      build_name: manywheel-py3_9-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_9-cpu-s390x-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: manywheel-py3_9-cpu-s390x-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.9"
-      build_name: manywheel-py3_9-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_9-cpu-s390x-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_9-cpu-s390x-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.9"
-      build_name: manywheel-py3_9-cpu-s390x
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
-  manywheel-py3_10-cpu-s390x-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.10"
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-      build_name: manywheel-py3_10-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_10-cpu-s390x-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: manywheel-py3_10-cpu-s390x-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.10"
-      build_name: manywheel-py3_10-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_10-cpu-s390x-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_10-cpu-s390x-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.10"
-      build_name: manywheel-py3_10-cpu-s390x
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
-  manywheel-py3_11-cpu-s390x-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.11"
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-      build_name: manywheel-py3_11-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_11-cpu-s390x-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: manywheel-py3_11-cpu-s390x-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.11"
-      build_name: manywheel-py3_11-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_11-cpu-s390x-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_11-cpu-s390x-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.11"
-      build_name: manywheel-py3_11-cpu-s390x
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml
-
-  manywheel-py3_12-cpu-s390x-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    uses: ./.github/workflows/_binary-build-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.12"
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-      build_name: manywheel-py3_12-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_12-cpu-s390x-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: manywheel-py3_12-cpu-s390x-build
-    uses: ./.github/workflows/_binary-test-linux.yml
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.12"
-      build_name: manywheel-py3_12-cpu-s390x
-      build_environment: linux-s390x-binary-manywheel
-      runs_on: linux.s390x
-      ALPINE_IMAGE: "docker.io/s390x/alpine"
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-  manywheel-py3_12-cpu-s390x-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: manywheel-py3_12-cpu-s390x-test
-    with:
-      PYTORCH_ROOT: /pytorch
-      BUILDER_ROOT: /builder
-      PACKAGE_TYPE: manywheel
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cpu
-      GPU_ARCH_TYPE: cpu-s390x
-      DOCKER_IMAGE: pytorch/manylinuxs390x-builder:cpu-s390x-main
-      DESIRED_PYTHON: "3.12"
-      build_name: manywheel-py3_12-cpu-s390x
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml

.github/workflows/generated-macos-arm64-binary-conda-nightly.yml

@@ -26,7 +26,7 @@ env:
   BUILD_ENVIRONMENT: macos-arm64-binary-conda
   GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
   PR_NUMBER: ${{ github.event.pull_request.number }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: macos-arm64-binary-conda-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true
@@ -34,7 +34,7 @@ concurrency:
 jobs:
   conda-py3_8-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -152,7 +152,7 @@ jobs:
     uses: ./.github/workflows/_binary-upload.yml
   conda-py3_9-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -270,7 +270,7 @@ jobs:
     uses: ./.github/workflows/_binary-upload.yml
   conda-py3_10-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -388,7 +388,7 @@ jobs:
     uses: ./.github/workflows/_binary-upload.yml
   conda-py3_11-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -506,7 +506,7 @@ jobs:
     uses: ./.github/workflows/_binary-upload.yml
   conda-py3_12-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch

.github/workflows/generated-macos-arm64-binary-libtorch-cxx11-abi-nightly.yml

@@ -26,7 +26,7 @@ env:
   BUILD_ENVIRONMENT: macos-arm64-binary-libtorch-cxx11-abi
   GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
   PR_NUMBER: ${{ github.event.pull_request.number }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: macos-arm64-binary-libtorch-cxx11-abi-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true
@@ -34,7 +34,7 @@ concurrency:
 jobs:
   libtorch-cpu-shared-with-deps-cxx11-abi-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch

.github/workflows/generated-macos-arm64-binary-wheel-nightly.yml

@@ -26,7 +26,7 @@ env:
   BUILD_ENVIRONMENT: macos-arm64-binary-wheel
   GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
   PR_NUMBER: ${{ github.event.pull_request.number }}
-  SKIP_ALL_TESTS: 0
+  SKIP_ALL_TESTS: 1
 concurrency:
   group: macos-arm64-binary-wheel-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
   cancel-in-progress: true
@@ -34,7 +34,7 @@ concurrency:
 jobs:
   wheel-py3_8-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -46,7 +46,7 @@ jobs:
       GPU_ARCH_TYPE: cpu
       SKIP_ALL_TESTS: 1
       DESIRED_PYTHON: "3.8"
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
       # For sccache access (only on non-forked PRs)
       AWS_ACCESS_KEY_ID: ${{ secrets.MACOS_SCCACHE_S3_ACCESS_KEY_ID }}
       AWS_SECRET_ACCESS_KEY: ${{ secrets.MACOS_SCCACHE_S3_SECRET_ACCESS_KEY }}
@@ -153,7 +153,7 @@ jobs:
     uses: ./.github/workflows/_binary-upload.yml
   wheel-py3_9-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -165,7 +165,7 @@ jobs:
       GPU_ARCH_TYPE: cpu
       SKIP_ALL_TESTS: 1
       DESIRED_PYTHON: "3.9"
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
       # For sccache access (only on non-forked PRs)
       AWS_ACCESS_KEY_ID: ${{ secrets.MACOS_SCCACHE_S3_ACCESS_KEY_ID }}
       AWS_SECRET_ACCESS_KEY: ${{ secrets.MACOS_SCCACHE_S3_SECRET_ACCESS_KEY }}
@@ -272,7 +272,7 @@ jobs:
     uses: ./.github/workflows/_binary-upload.yml
   wheel-py3_10-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -284,7 +284,7 @@ jobs:
       GPU_ARCH_TYPE: cpu
       SKIP_ALL_TESTS: 1
       DESIRED_PYTHON: "3.10"
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
       # For sccache access (only on non-forked PRs)
       AWS_ACCESS_KEY_ID: ${{ secrets.MACOS_SCCACHE_S3_ACCESS_KEY_ID }}
       AWS_SECRET_ACCESS_KEY: ${{ secrets.MACOS_SCCACHE_S3_SECRET_ACCESS_KEY }}
@@ -391,7 +391,7 @@ jobs:
     uses: ./.github/workflows/_binary-upload.yml
   wheel-py3_11-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -403,7 +403,7 @@ jobs:
       GPU_ARCH_TYPE: cpu
       SKIP_ALL_TESTS: 1
       DESIRED_PYTHON: "3.11"
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
       # For sccache access (only on non-forked PRs)
       AWS_ACCESS_KEY_ID: ${{ secrets.MACOS_SCCACHE_S3_ACCESS_KEY_ID }}
       AWS_SECRET_ACCESS_KEY: ${{ secrets.MACOS_SCCACHE_S3_SECRET_ACCESS_KEY }}
@@ -510,7 +510,7 @@ jobs:
     uses: ./.github/workflows/_binary-upload.yml
   wheel-py3_12-cpu-build:
     if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: macos-14-xlarge
+    runs-on: macos-13-xlarge
     timeout-minutes: 240
     env:
       PYTORCH_ROOT: ${{ github.workspace }}/pytorch
@@ -522,7 +522,7 @@ jobs:
       GPU_ARCH_TYPE: cpu
       SKIP_ALL_TESTS: 1
       DESIRED_PYTHON: "3.12"
-      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==9.1.0.70; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
+      PYTORCH_EXTRA_INSTALL_REQUIREMENTS: nvidia-cuda-nvrtc-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-runtime-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cuda-cupti-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cudnn-cu12==8.9.2.26; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cublas-cu12==12.1.3.1; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cufft-cu12==11.0.2.54; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-curand-cu12==10.3.2.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusolver-cu12==11.4.5.107; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-cusparse-cu12==12.1.0.106; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nccl-cu12==2.20.5; platform_system == 'Linux' and platform_machine == 'x86_64' | nvidia-nvtx-cu12==12.1.105; platform_system == 'Linux' and platform_machine == 'x86_64'
       # For sccache access (only on non-forked PRs)
       AWS_ACCESS_KEY_ID: ${{ secrets.MACOS_SCCACHE_S3_ACCESS_KEY_ID }}
       AWS_SECRET_ACCESS_KEY: ${{ secrets.MACOS_SCCACHE_S3_SECRET_ACCESS_KEY }}

.github/workflows/generated-windows-binary-libtorch-debug-nightly.yml

@@ -800,260 +800,3 @@ jobs:
       conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
-  libtorch-cuda12_4-shared-with-deps-debug-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: windows.4xlarge.nonephemeral
-    timeout-minutes: 240
-    env:
-      PYTORCH_ROOT: ${{ github.workspace }}/pytorch
-      BUILDER_ROOT: ${{ github.workspace }}/builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      SKIP_ALL_TESTS: 1
-      LIBTORCH_CONFIG: debug
-      LIBTORCH_VARIANT: shared-with-deps
-      # This is a dummy value for libtorch to work correctly with our batch scripts
-      # without this value pip does not get installed for some reason
-      DESIRED_PYTHON: "3.8"
-    steps:
-      - name: Display EC2 information
-        shell: bash
-        run: |
-          set -euo pipefail
-          function get_ec2_metadata() {
-            # Pulled from instance metadata endpoint for EC2
-            # see https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/instancedata-data-retrieval.html
-            category=$1
-            curl -fsSL "http://169.254.169.254/latest/meta-data/${category}"
-          }
-          echo "ami-id: $(get_ec2_metadata ami-id)"
-          echo "instance-id: $(get_ec2_metadata instance-id)"
-          echo "instance-type: $(get_ec2_metadata instance-type)"
-          echo "system info $(uname -a)"
-      - name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
-        uses: pytorch/test-infra/.github/actions/setup-ssh@main
-        continue-on-error: true
-        with:
-          github-secret: ${{ secrets.GITHUB_TOKEN }}
-      # Needed for binary builds, see: https://github.com/pytorch/pytorch/issues/73339#issuecomment-1058981560
-      - name: Enable long paths on Windows
-        shell: powershell
-        run: |
-          Set-ItemProperty -Path "HKLM:\\SYSTEM\CurrentControlSet\Control\FileSystem" -Name "LongPathsEnabled" -Value 1
-      # Since it's just a defensive command, the workflow should continue even the command fails. This step can be
-      # removed once Windows Defender is removed from the AMI
-      - name: Disables Windows Defender scheduled and real-time scanning for files in directories used by PyTorch
-        continue-on-error: true
-        shell: powershell
-        run: |
-          Add-MpPreference -ExclusionPath $(Get-Location).tostring(),$Env:TEMP -ErrorAction Ignore
-          # Let's both exclude the path and disable Windows Defender completely just to be sure
-          # that it doesn't interfere
-          Set-MpPreference -DisableRealtimeMonitoring $True -ErrorAction Ignore
-      # NOTE: These environment variables are put here so that they can be applied on every job equally
-      #       They are also here because setting them at a workflow level doesn't give us access to the
-      #       runner.temp variable, which we need.
-      - name: Populate binary env
-        shell: bash
-        run: |
-          echo "BINARY_ENV_FILE=${RUNNER_TEMP}/env" >> "${GITHUB_ENV}"
-          echo "PYTORCH_FINAL_PACKAGE_DIR=${RUNNER_TEMP}/artifacts" >> "${GITHUB_ENV}"
-          echo "WIN_PACKAGE_WORK_DIR=${RUNNER_TEMP}"
-      - name: Checkout PyTorch
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
-          submodules: recursive
-          path: pytorch
-          quiet-checkout: true
-      - name: Clean PyTorch checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: pytorch
-      - name: Checkout pytorch/builder
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: main
-          submodules: recursive
-          repository: pytorch/builder
-          path: builder
-          quiet-checkout: true
-      - name: Clean pytorch/builder checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: builder
-      - name: Populate binary env
-        shell: bash
-        run: |
-          "${PYTORCH_ROOT}/.circleci/scripts/binary_populate_env.sh"
-      - name: Build PyTorch binary
-        shell: bash
-        run: |
-          "${PYTORCH_ROOT}/.circleci/scripts/binary_windows_build.sh"
-      - uses: actions/upload-artifact@v3
-        if: always()
-        with:
-          name: libtorch-cuda12_4-shared-with-deps-debug
-          retention-days: 14
-          if-no-files-found: error
-          path: "${{ env.PYTORCH_FINAL_PACKAGE_DIR }}"
-      - name: Wait until all sessions have drained
-        shell: powershell
-        working-directory: pytorch
-        if: always()
-        timeout-minutes: 120
-        run: |
-          .github\scripts\wait_for_ssh_to_drain.ps1
-      - name: Kill active ssh sessions if still around (Useful if workflow was cancelled)
-        shell: powershell
-        working-directory: pytorch
-        if: always()
-        run: |
-          .github\scripts\kill_active_ssh_sessions.ps1
-  libtorch-cuda12_4-shared-with-deps-debug-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: libtorch-cuda12_4-shared-with-deps-debug-build
-    runs-on: windows.8xlarge.nvidia.gpu
-    timeout-minutes: 240
-    env:
-      PYTORCH_ROOT: ${{ github.workspace }}/pytorch
-      BUILDER_ROOT: ${{ github.workspace }}/builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      SKIP_ALL_TESTS: 1
-      LIBTORCH_CONFIG: debug
-      LIBTORCH_VARIANT: shared-with-deps
-      # This is a dummy value for libtorch to work correctly with our batch scripts
-      # without this value pip does not get installed for some reason
-      DESIRED_PYTHON: "3.8"
-    steps:
-      - name: Display EC2 information
-        shell: bash
-        run: |
-          set -euo pipefail
-          function get_ec2_metadata() {
-            # Pulled from instance metadata endpoint for EC2
-            # see https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/instancedata-data-retrieval.html
-            category=$1
-            curl -fsSL "http://169.254.169.254/latest/meta-data/${category}"
-          }
-          echo "ami-id: $(get_ec2_metadata ami-id)"
-          echo "instance-id: $(get_ec2_metadata instance-id)"
-          echo "instance-type: $(get_ec2_metadata instance-type)"
-          echo "system info $(uname -a)"
-      - name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
-        uses: pytorch/test-infra/.github/actions/setup-ssh@main
-        continue-on-error: true
-        with:
-          github-secret: ${{ secrets.GITHUB_TOKEN }}
-      # Needed for binary builds, see: https://github.com/pytorch/pytorch/issues/73339#issuecomment-1058981560
-      - name: Enable long paths on Windows
-        shell: powershell
-        run: |
-          Set-ItemProperty -Path "HKLM:\\SYSTEM\CurrentControlSet\Control\FileSystem" -Name "LongPathsEnabled" -Value 1
-      # Since it's just a defensive command, the workflow should continue even the command fails. This step can be
-      # removed once Windows Defender is removed from the AMI
-      - name: Disables Windows Defender scheduled and real-time scanning for files in directories used by PyTorch
-        continue-on-error: true
-        shell: powershell
-        run: |
-          Add-MpPreference -ExclusionPath $(Get-Location).tostring(),$Env:TEMP -ErrorAction Ignore
-          # Let's both exclude the path and disable Windows Defender completely just to be sure
-          # that it doesn't interfere
-          Set-MpPreference -DisableRealtimeMonitoring $True -ErrorAction Ignore
-      # NOTE: These environment variables are put here so that they can be applied on every job equally
-      #       They are also here because setting them at a workflow level doesn't give us access to the
-      #       runner.temp variable, which we need.
-      - name: Populate binary env
-        shell: bash
-        run: |
-          echo "BINARY_ENV_FILE=${RUNNER_TEMP}/env" >> "${GITHUB_ENV}"
-          echo "PYTORCH_FINAL_PACKAGE_DIR=${RUNNER_TEMP}/artifacts" >> "${GITHUB_ENV}"
-          echo "WIN_PACKAGE_WORK_DIR=${RUNNER_TEMP}"
-      - uses: actions/download-artifact@v3
-        name: Download Build Artifacts
-        with:
-          name: libtorch-cuda12_4-shared-with-deps-debug
-          path: "${{ env.PYTORCH_FINAL_PACKAGE_DIR }}"
-      - name: Checkout PyTorch
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
-          submodules: recursive
-          path: pytorch
-          quiet-checkout: true
-      - name: Clean PyTorch checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: pytorch
-      - name: Checkout pytorch/builder
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: main
-          submodules: recursive
-          repository: pytorch/builder
-          path: builder
-          quiet-checkout: true
-      - name: Clean pytorch/builder checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: builder
-      - name: Populate binary env
-        shell: bash
-        run: |
-          "${PYTORCH_ROOT}/.circleci/scripts/binary_populate_env.sh"
-      - name: Test PyTorch binary
-        shell: bash
-        run: |
-          "${PYTORCH_ROOT}/.circleci/scripts/binary_windows_test.sh"
-      - name: Wait until all sessions have drained
-        shell: powershell
-        working-directory: pytorch
-        if: always()
-        timeout-minutes: 120
-        run: |
-          .github\scripts\wait_for_ssh_to_drain.ps1
-      - name: Kill active ssh sessions if still around (Useful if workflow was cancelled)
-        shell: powershell
-        working-directory: pytorch
-        if: always()
-        run: |
-          .github\scripts\kill_active_ssh_sessions.ps1
-  libtorch-cuda12_4-shared-with-deps-debug-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: libtorch-cuda12_4-shared-with-deps-debug-test
-    with:
-      PYTORCH_ROOT: ${{ github.workspace }}/pytorch
-      BUILDER_ROOT: ${{ github.workspace }}/builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      LIBTORCH_CONFIG: debug
-      LIBTORCH_VARIANT: shared-with-deps
-      # This is a dummy value for libtorch to work correctly with our batch scripts
-      # without this value pip does not get installed for some reason
-      DESIRED_PYTHON: "3.8"
-      build_name: libtorch-cuda12_4-shared-with-deps-debug
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml

.github/workflows/generated-windows-binary-libtorch-release-nightly.yml

@@ -800,260 +800,3 @@ jobs:
       conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
       conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
     uses: ./.github/workflows/_binary-upload.yml
-  libtorch-cuda12_4-shared-with-deps-release-build:
-    if: ${{ github.repository_owner == 'pytorch' }}
-    runs-on: windows.4xlarge.nonephemeral
-    timeout-minutes: 240
-    env:
-      PYTORCH_ROOT: ${{ github.workspace }}/pytorch
-      BUILDER_ROOT: ${{ github.workspace }}/builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      SKIP_ALL_TESTS: 1
-      LIBTORCH_CONFIG: release
-      LIBTORCH_VARIANT: shared-with-deps
-      # This is a dummy value for libtorch to work correctly with our batch scripts
-      # without this value pip does not get installed for some reason
-      DESIRED_PYTHON: "3.8"
-    steps:
-      - name: Display EC2 information
-        shell: bash
-        run: |
-          set -euo pipefail
-          function get_ec2_metadata() {
-            # Pulled from instance metadata endpoint for EC2
-            # see https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/instancedata-data-retrieval.html
-            category=$1
-            curl -fsSL "http://169.254.169.254/latest/meta-data/${category}"
-          }
-          echo "ami-id: $(get_ec2_metadata ami-id)"
-          echo "instance-id: $(get_ec2_metadata instance-id)"
-          echo "instance-type: $(get_ec2_metadata instance-type)"
-          echo "system info $(uname -a)"
-      - name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
-        uses: pytorch/test-infra/.github/actions/setup-ssh@main
-        continue-on-error: true
-        with:
-          github-secret: ${{ secrets.GITHUB_TOKEN }}
-      # Needed for binary builds, see: https://github.com/pytorch/pytorch/issues/73339#issuecomment-1058981560
-      - name: Enable long paths on Windows
-        shell: powershell
-        run: |
-          Set-ItemProperty -Path "HKLM:\\SYSTEM\CurrentControlSet\Control\FileSystem" -Name "LongPathsEnabled" -Value 1
-      # Since it's just a defensive command, the workflow should continue even the command fails. This step can be
-      # removed once Windows Defender is removed from the AMI
-      - name: Disables Windows Defender scheduled and real-time scanning for files in directories used by PyTorch
-        continue-on-error: true
-        shell: powershell
-        run: |
-          Add-MpPreference -ExclusionPath $(Get-Location).tostring(),$Env:TEMP -ErrorAction Ignore
-          # Let's both exclude the path and disable Windows Defender completely just to be sure
-          # that it doesn't interfere
-          Set-MpPreference -DisableRealtimeMonitoring $True -ErrorAction Ignore
-      # NOTE: These environment variables are put here so that they can be applied on every job equally
-      #       They are also here because setting them at a workflow level doesn't give us access to the
-      #       runner.temp variable, which we need.
-      - name: Populate binary env
-        shell: bash
-        run: |
-          echo "BINARY_ENV_FILE=${RUNNER_TEMP}/env" >> "${GITHUB_ENV}"
-          echo "PYTORCH_FINAL_PACKAGE_DIR=${RUNNER_TEMP}/artifacts" >> "${GITHUB_ENV}"
-          echo "WIN_PACKAGE_WORK_DIR=${RUNNER_TEMP}"
-      - name: Checkout PyTorch
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
-          submodules: recursive
-          path: pytorch
-          quiet-checkout: true
-      - name: Clean PyTorch checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: pytorch
-      - name: Checkout pytorch/builder
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: main
-          submodules: recursive
-          repository: pytorch/builder
-          path: builder
-          quiet-checkout: true
-      - name: Clean pytorch/builder checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: builder
-      - name: Populate binary env
-        shell: bash
-        run: |
-          "${PYTORCH_ROOT}/.circleci/scripts/binary_populate_env.sh"
-      - name: Build PyTorch binary
-        shell: bash
-        run: |
-          "${PYTORCH_ROOT}/.circleci/scripts/binary_windows_build.sh"
-      - uses: actions/upload-artifact@v3
-        if: always()
-        with:
-          name: libtorch-cuda12_4-shared-with-deps-release
-          retention-days: 14
-          if-no-files-found: error
-          path: "${{ env.PYTORCH_FINAL_PACKAGE_DIR }}"
-      - name: Wait until all sessions have drained
-        shell: powershell
-        working-directory: pytorch
-        if: always()
-        timeout-minutes: 120
-        run: |
-          .github\scripts\wait_for_ssh_to_drain.ps1
-      - name: Kill active ssh sessions if still around (Useful if workflow was cancelled)
-        shell: powershell
-        working-directory: pytorch
-        if: always()
-        run: |
-          .github\scripts\kill_active_ssh_sessions.ps1
-  libtorch-cuda12_4-shared-with-deps-release-test:  # Testing
-    if: ${{ github.repository_owner == 'pytorch' }}
-    needs: libtorch-cuda12_4-shared-with-deps-release-build
-    runs-on: windows.8xlarge.nvidia.gpu
-    timeout-minutes: 240
-    env:
-      PYTORCH_ROOT: ${{ github.workspace }}/pytorch
-      BUILDER_ROOT: ${{ github.workspace }}/builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      SKIP_ALL_TESTS: 1
-      LIBTORCH_CONFIG: release
-      LIBTORCH_VARIANT: shared-with-deps
-      # This is a dummy value for libtorch to work correctly with our batch scripts
-      # without this value pip does not get installed for some reason
-      DESIRED_PYTHON: "3.8"
-    steps:
-      - name: Display EC2 information
-        shell: bash
-        run: |
-          set -euo pipefail
-          function get_ec2_metadata() {
-            # Pulled from instance metadata endpoint for EC2
-            # see https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/instancedata-data-retrieval.html
-            category=$1
-            curl -fsSL "http://169.254.169.254/latest/meta-data/${category}"
-          }
-          echo "ami-id: $(get_ec2_metadata ami-id)"
-          echo "instance-id: $(get_ec2_metadata instance-id)"
-          echo "instance-type: $(get_ec2_metadata instance-type)"
-          echo "system info $(uname -a)"
-      - name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
-        uses: pytorch/test-infra/.github/actions/setup-ssh@main
-        continue-on-error: true
-        with:
-          github-secret: ${{ secrets.GITHUB_TOKEN }}
-      # Needed for binary builds, see: https://github.com/pytorch/pytorch/issues/73339#issuecomment-1058981560
-      - name: Enable long paths on Windows
-        shell: powershell
-        run: |
-          Set-ItemProperty -Path "HKLM:\\SYSTEM\CurrentControlSet\Control\FileSystem" -Name "LongPathsEnabled" -Value 1
-      # Since it's just a defensive command, the workflow should continue even the command fails. This step can be
-      # removed once Windows Defender is removed from the AMI
-      - name: Disables Windows Defender scheduled and real-time scanning for files in directories used by PyTorch
-        continue-on-error: true
-        shell: powershell
-        run: |
-          Add-MpPreference -ExclusionPath $(Get-Location).tostring(),$Env:TEMP -ErrorAction Ignore
-          # Let's both exclude the path and disable Windows Defender completely just to be sure
-          # that it doesn't interfere
-          Set-MpPreference -DisableRealtimeMonitoring $True -ErrorAction Ignore
-      # NOTE: These environment variables are put here so that they can be applied on every job equally
-      #       They are also here because setting them at a workflow level doesn't give us access to the
-      #       runner.temp variable, which we need.
-      - name: Populate binary env
-        shell: bash
-        run: |
-          echo "BINARY_ENV_FILE=${RUNNER_TEMP}/env" >> "${GITHUB_ENV}"
-          echo "PYTORCH_FINAL_PACKAGE_DIR=${RUNNER_TEMP}/artifacts" >> "${GITHUB_ENV}"
-          echo "WIN_PACKAGE_WORK_DIR=${RUNNER_TEMP}"
-      - uses: actions/download-artifact@v3
-        name: Download Build Artifacts
-        with:
-          name: libtorch-cuda12_4-shared-with-deps-release
-          path: "${{ env.PYTORCH_FINAL_PACKAGE_DIR }}"
-      - name: Checkout PyTorch
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
-          submodules: recursive
-          path: pytorch
-          quiet-checkout: true
-      - name: Clean PyTorch checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: pytorch
-      - name: Checkout pytorch/builder
-        uses: malfet/checkout@silent-checkout
-        with:
-          ref: main
-          submodules: recursive
-          repository: pytorch/builder
-          path: builder
-          quiet-checkout: true
-      - name: Clean pytorch/builder checkout
-        run: |
-          # Remove any artifacts from the previous checkouts
-          git clean -fxd
-        working-directory: builder
-      - name: Populate binary env
-        shell: bash
-        run: |
-          "${PYTORCH_ROOT}/.circleci/scripts/binary_populate_env.sh"
-      - name: Test PyTorch binary
-        shell: bash
-        run: |
-          "${PYTORCH_ROOT}/.circleci/scripts/binary_windows_test.sh"
-      - name: Wait until all sessions have drained
-        shell: powershell
-        working-directory: pytorch
-        if: always()
-        timeout-minutes: 120
-        run: |
-          .github\scripts\wait_for_ssh_to_drain.ps1
-      - name: Kill active ssh sessions if still around (Useful if workflow was cancelled)
-        shell: powershell
-        working-directory: pytorch
-        if: always()
-        run: |
-          .github\scripts\kill_active_ssh_sessions.ps1
-  libtorch-cuda12_4-shared-with-deps-release-upload:  # Uploading
-    if: ${{ github.repository_owner == 'pytorch' }}
-    permissions:
-      id-token: write
-      contents: read
-    needs: libtorch-cuda12_4-shared-with-deps-release-test
-    with:
-      PYTORCH_ROOT: ${{ github.workspace }}/pytorch
-      BUILDER_ROOT: ${{ github.workspace }}/builder
-      PACKAGE_TYPE: libtorch
-      # TODO: This is a legacy variable that we eventually want to get rid of in
-      #       favor of GPU_ARCH_VERSION
-      DESIRED_CUDA: cu124
-      GPU_ARCH_VERSION: 12.4
-      GPU_ARCH_TYPE: cuda
-      LIBTORCH_CONFIG: release
-      LIBTORCH_VARIANT: shared-with-deps
-      # This is a dummy value for libtorch to work correctly with our batch scripts
-      # without this value pip does not get installed for some reason
-      DESIRED_PYTHON: "3.8"
-      build_name: libtorch-cuda12_4-shared-with-deps-release
-    secrets:
-      github-token: ${{ secrets.GITHUB_TOKEN }}
-      conda-pytorchbot-token: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
-      conda-pytorchbot-token-test: ${{ secrets.CONDA_PYTORCHBOT_TOKEN_TEST }}
-    uses: ./.github/workflows/_binary-upload.yml

.github/workflows/inductor-cu124.yml

@@ -1,108 +0,0 @@
-name: inductor-cu124
-
-on:
-  push:
-    tags:
-      - ciflow/inductor-cu124/*
-  workflow_dispatch:
-  schedule:
-    # Run every 4 hours during the week and every 12 hours on the weekend
-    - cron: 45 0,4,8,12,16,20 * * 1-5
-    - cron: 45 4,12 * * 0,6
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
-  cancel-in-progress: true
-
-permissions: read-all
-
-jobs:
-  linux-focal-cuda12_4-py3_10-gcc9-inductor-build:
-    # Should be synced with the one in inductor.yml, but this doesn't run inductor_timm
-    name: cuda12.4-py3.10-gcc9-sm86
-    uses: ./.github/workflows/_linux-build.yml
-    with:
-      sync-tag: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
-      build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86
-      docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
-      cuda-arch-list: '8.6'
-      test-matrix: |
-        { include: [
-          { config: "inductor", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "inductor_distributed", shard: 1, num_shards: 1, runner: "linux.g5.12xlarge.nvidia.gpu" },
-          { config: "inductor_huggingface", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "inductor_torchbench", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "inductor_torchbench", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "dynamic_inductor_huggingface", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "dynamic_inductor_timm", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "dynamic_inductor_timm", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "dynamic_inductor_torchbench", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "dynamic_inductor_torchbench", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "aot_inductor_huggingface", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "aot_inductor_timm", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "aot_inductor_timm", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "aot_inductor_torchbench", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "aot_inductor_torchbench", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "inductor_cpp_wrapper_abi_compatible", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
-        ]}
-    secrets:
-      HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
-
-  linux-focal-cuda12_4-py3_10-gcc9-inductor-test:
-    name: cuda12.4-py3.10-gcc9-sm86
-    uses: ./.github/workflows/_linux-test.yml
-    needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
-    with:
-      sync-tag: linux-focal-cuda12_4-py3_10-gcc9-inductor-test
-      build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86
-      docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.docker-image }}
-      test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.test-matrix }}
-    secrets:
-      HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
-
-  linux-focal-cuda12_4-py3_10-gcc9-inductor-build-gcp:
-    name: cuda12.4-py3.10-gcc9-sm80
-    uses: ./.github/workflows/_linux-build.yml
-    with:
-      build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
-      docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
-      cuda-arch-list: '8.0'
-      test-matrix: |
-        { include: [
-          { config: "inductor_torchbench_smoketest_perf", shard: 1, num_shards: 1, runner: "linux.gcp.a100" },
-        ]}
-    secrets:
-      HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
-
-  linux-focal-cuda12_4-py3_10-gcc9-inductor-test-gcp:
-    name: cuda12.4-py3.10-gcc9-sm80
-    uses: ./.github/workflows/_linux-test.yml
-    needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build-gcp
-    with:
-      build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
-      docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build-gcp.outputs.docker-image }}
-      test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build-gcp.outputs.test-matrix }}
-      use-gha: anything-non-empty-to-use-gha
-    secrets:
-      HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
-
-  linux-focal-cuda12_4-py3_12-gcc9-inductor-build:
-    name: cuda12.4-py3.12-gcc9-sm86
-    uses: ./.github/workflows/_linux-build.yml
-    with:
-      build-environment: linux-focal-cuda12.4-py3.12-gcc9-sm86
-      docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3.12-gcc9-inductor-benchmarks
-      cuda-arch-list: '8.6'
-      test-matrix: |
-        { include: [
-          { config: "inductor", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
-        ]}
-
-  linux-focal-cuda12_4-py3_12-gcc9-inductor-test:
-    name: cuda12.4-py3.12-gcc9-sm86
-    uses: ./.github/workflows/_linux-test.yml
-    needs: linux-focal-cuda12_4-py3_12-gcc9-inductor-build
-    with:
-      build-environment: linux-focal-cuda12.4-py3.12-gcc9-sm86
-      docker-image: ${{ needs.linux-focal-cuda12_4-py3_12-gcc9-inductor-build.outputs.docker-image }}
-      test-matrix: ${{ needs.linux-focal-cuda12_4-py3_12-gcc9-inductor-build.outputs.test-matrix }}

.github/workflows/inductor-micro-benchmark.yml

@@ -1,40 +0,0 @@
-name: inductor-micro-benchmark
-
-on:
-  schedule:
-    - cron: 0 7 * * *
-  push:
-    tags:
-      - ciflow/inductor-micro-benchmark/*
-  workflow_dispatch:
-
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}-${{ github.event_name == 'schedule' }}
-  cancel-in-progress: true
-
-permissions: read-all
-
-jobs:
-  linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-build:
-    name: cuda12.1-py3.10-gcc9-sm80
-    uses: ./.github/workflows/_linux-build.yml
-    with:
-      build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80
-      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks
-      cuda-arch-list: '8.0'
-      test-matrix: |
-        { include: [
-          { config: "inductor-micro-benchmark", shard: 1, num_shards: 1, runner: "linux.gcp.a100" },
-        ]}
-
-  linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-test:
-    name: cuda12.1-py3.10-gcc9-sm80
-    uses: ./.github/workflows/_linux-test.yml
-    needs: linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-build
-    with:
-      build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80
-      docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-build.outputs.docker-image }}
-      test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-build.outputs.test-matrix }}
-      use-gha: anything-non-empty-to-use-gha
-      timeout-minutes: 720

.github/workflows/inductor-perf-compare.yml

@@ -18,7 +18,7 @@ jobs:
     uses: ./.github/workflows/_linux-build.yml
     with:
       build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80
-      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks
+      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9-inductor-benchmarks
       cuda-arch-list: '8.0'
       test-matrix: |
         { include: [

.github/workflows/inductor-perf-test-nightly.yml

@@ -71,7 +71,7 @@ jobs:
     uses: ./.github/workflows/_linux-build.yml
     with:
       build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80
-      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks
+      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9-inductor-benchmarks
       cuda-arch-list: '8.0'
       test-matrix: |
         { include: [

.github/workflows/inductor-periodic.yml

@@ -23,7 +23,7 @@ jobs:
     uses: ./.github/workflows/_linux-build.yml
     with:
       build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm86
-      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks
+      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9-inductor-benchmarks
       cuda-arch-list: '8.6'
       test-matrix: |
         { include: [

.github/workflows/inductor.yml

@@ -16,35 +16,35 @@ concurrency:
 permissions: read-all
 
 jobs:
-  linux-focal-rocm6_1-py3_8-inductor-build:
-    name: rocm6.1-py3.8-inductor
+  linux-focal-rocm6_0-py3_8-inductor-build:
+    name: rocm6.0-py3.8-inductor
     uses: ./.github/workflows/_linux-build.yml
     with:
-      build-environment: linux-focal-rocm6.1-py3.8
+      build-environment: linux-focal-rocm6.0-py3.8
       docker-image-name: pytorch-linux-focal-rocm-n-py3
       test-matrix: |
         { include: [
           { config: "inductor", shard: 1, num_shards: 1, runner: "linux.rocm.gpu.2" },
         ]}
 
-  linux-focal-rocm6_1-py3_8-inductor-test:
+  linux-focal-rocm6_0-py3_8-inductor-test:
     permissions:
       id-token: write
       contents: read
-    name: rocm6.1-py3.8-inductor
+    name: rocm6.0-py3.8-inductor
     uses: ./.github/workflows/_rocm-test.yml
-    needs: linux-focal-rocm6_1-py3_8-inductor-build
+    needs: linux-focal-rocm6_0-py3_8-inductor-build
     with:
-      build-environment: linux-focal-rocm6.1-py3.8
-      docker-image: ${{ needs.linux-focal-rocm6_1-py3_8-inductor-build.outputs.docker-image }}
-      test-matrix:  ${{ needs.linux-focal-rocm6_1-py3_8-inductor-build.outputs.test-matrix }}
+      build-environment: linux-focal-rocm6.0-py3.8
+      docker-image: ${{ needs.linux-focal-rocm6_0-py3_8-inductor-build.outputs.docker-image }}
+      test-matrix:  ${{ needs.linux-focal-rocm6_0-py3_8-inductor-build.outputs.test-matrix }}
 
   linux-focal-cuda12_1-py3_10-gcc9-inductor-build:
     name: cuda12.1-py3.10-gcc9-sm86
     uses: ./.github/workflows/_linux-build.yml
     with:
       build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm86
-      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks
+      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9-inductor-benchmarks
       cuda-arch-list: '8.6'
       test-matrix: |
         { include: [
@@ -86,7 +86,7 @@ jobs:
     uses: ./.github/workflows/_linux-build.yml
     with:
       build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80
-      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks
+      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn8-py3-gcc9-inductor-benchmarks
       cuda-arch-list: '8.0'
       test-matrix: |
         { include: [
@@ -107,56 +107,6 @@ jobs:
     secrets:
       HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
 
-  linux-focal-cuda12_1-py3_12-gcc9-inductor-build:
-    name: cuda12.1-py3.12-gcc9-sm86
-    uses: ./.github/workflows/_linux-build.yml
-    with:
-      build-environment: linux-focal-cuda12.1-py3.12-gcc9-sm86
-      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3.12-gcc9-inductor-benchmarks
-      cuda-arch-list: '8.6'
-      test-matrix: |
-        { include: [
-          { config: "inductor", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
-        ]}
-
-  linux-focal-cuda12_1-py3_12-gcc9-inductor-test:
-    name: cuda12.1-py3.12-gcc9-sm86
-    uses: ./.github/workflows/_linux-test.yml
-    needs: linux-focal-cuda12_1-py3_12-gcc9-inductor-build
-    with:
-      build-environment: linux-focal-cuda12.1-py3.12-gcc9-sm86
-      docker-image: ${{ needs.linux-focal-cuda12_1-py3_12-gcc9-inductor-build.outputs.docker-image }}
-      test-matrix: ${{ needs.linux-focal-cuda12_1-py3_12-gcc9-inductor-build.outputs.test-matrix }}
-
-  linux-focal-cuda12_4-py3_10-gcc9-inductor-build:
-    # Should be synced with the one in inductor-periodic.yml but this only runs inductor_timm
-    name: cuda12.4-py3.10-gcc9-sm86
-    uses: ./.github/workflows/_linux-build.yml
-    with:
-      sync-tag: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
-      build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86
-      docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
-      cuda-arch-list: '8.6'
-      test-matrix: |
-        { include: [
-          { config: "inductor_timm", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-          { config: "inductor_timm", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
-        ]}
-    secrets:
-      HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
-
-  linux-focal-cuda12_4-py3_10-gcc9-inductor-test:
-    name: cuda12.4-py3.10-gcc9-sm86
-    uses: ./.github/workflows/_linux-test.yml
-    needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
-    with:
-      sync-tag: linux-focal-cuda12_4-py3_10-gcc9-inductor-test
-      build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86
-      docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.docker-image }}
-      test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.test-matrix }}
-    secrets:
-      HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
-
   linux-jammy-cpu-py3_8-gcc11-inductor-build:
     name: linux-jammy-cpu-py3.8-gcc11-inductor
     uses: ./.github/workflows/_linux-build.yml
@@ -170,17 +120,12 @@ jobs:
           { config: "cpu_inductor_timm", shard: 2, num_shards: 2, runner: "linux.12xlarge" },
           { config: "cpu_inductor_torchbench", shard: 1, num_shards: 2, runner: "linux.12xlarge" },
           { config: "cpu_inductor_torchbench", shard: 2, num_shards: 2, runner: "linux.12xlarge" },
-          { config: "cpu_inductor_huggingface_freezing", shard: 1, num_shards: 1, runner: "linux.12xlarge" },
-          { config: "cpu_inductor_timm_freezing", shard: 1, num_shards: 2, runner: "linux.12xlarge" },
-          { config: "cpu_inductor_timm_freezing", shard: 2, num_shards: 2, runner: "linux.12xlarge" },
-          { config: "cpu_inductor_torchbench_freezing", shard: 1, num_shards: 2, runner: "linux.12xlarge" },
-          { config: "cpu_inductor_torchbench_freezing", shard: 2, num_shards: 2, runner: "linux.12xlarge" },
           { config: "dynamic_cpu_inductor_huggingface", shard: 1, num_shards: 1, runner: "linux.12xlarge" },
           { config: "dynamic_cpu_inductor_timm", shard: 1, num_shards: 2, runner: "linux.12xlarge" },
           { config: "dynamic_cpu_inductor_timm", shard: 2, num_shards: 2, runner: "linux.12xlarge" },
           { config: "dynamic_cpu_inductor_torchbench", shard: 1, num_shards: 2, runner: "linux.12xlarge" },
           { config: "dynamic_cpu_inductor_torchbench", shard: 2, num_shards: 2, runner: "linux.12xlarge" },
-          { config: "inductor_torchbench_cpu_smoketest_perf", shard: 1, num_shards: 1, runner: "linux.24xl.spr-metal" },
+          { config: "inductor_torchbench_cpu_smoketest_perf", shard: 1, num_shards: 1, runner: "linux.12xlarge" },
         ]}
     secrets:
       HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}

.github/workflows/lint.yml

@@ -20,7 +20,7 @@ jobs:
     with:
       timeout: 120
       runner: linux.2xlarge
-      docker-image: pytorch-linux-jammy-cuda11.8-cudnn9-py3.9-linter
+      docker-image: pytorch-linux-jammy-cuda11.8-cudnn8-py3.9-linter
       # NB: A shallow checkout won't work here because calculate-docker-image requires a full checkout
       # to run git rev-parse HEAD~:.ci/docker when a new image is needed
       fetch-depth: 0
@@ -36,7 +36,7 @@ jobs:
     with:
       timeout: 120
       runner: linux.2xlarge
-      docker-image: pytorch-linux-jammy-cuda11.8-cudnn9-py3.9-linter
+      docker-image: pytorch-linux-jammy-cuda11.8-cudnn8-py3.9-linter
       # NB: A shallow checkout won't work here because calculate-docker-image requires a full checkout
       # to run git rev-parse HEAD~:.ci/docker when a new image is needed
       fetch-depth: 0
@@ -230,11 +230,11 @@ jobs:
         with:
           submodules: false
           fetch-depth: 1
-      - name: Setup Python 3.6
+      - name: Setup Python 3.5
         if: matrix.test_type == 'older_python_version'
         uses: actions/setup-python@v4
         with:
-          python-version: '3.6'
+          python-version: '3.5'
           architecture: x64
           check-latest: false
           cache: pip

.github/workflows/linux-aarch64.yml

@@ -1,39 +0,0 @@
-name: linux-aarch64
-
-on:
-  push:
-    tags:
-      - ciflow/linux-aarch64/*
-  workflow_dispatch:
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.sha }}-${{ github.event_name == 'workflow_dispatch' }} but found ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}-${{ github.event_name == 'schedule' }}
-  cancel-in-progress: true
-
-jobs:
-  linux-jammy-aarch64-py3_10-build:
-    name: linux-jammy-aarch64-py3.10
-    uses: ./.github/workflows/_linux-build.yml
-    with:
-      build-environment: linux-jammy-aarch64-py3.10
-      docker-image-name: pytorch-linux-jammy-aarch64-py3.10-gcc11
-      runner: linux.arm64.2xlarge
-      test-matrix: |
-        { include: [
-          { config: "default", shard: 1, num_shards: 4, runner: "linux.arm64.2xlarge" },
-          { config: "default", shard: 2, num_shards: 4, runner: "linux.arm64.2xlarge" },
-          { config: "default", shard: 3, num_shards: 4, runner: "linux.arm64.2xlarge" },
-          { config: "default", shard: 4, num_shards: 4, runner: "linux.arm64.2xlarge" },
-        ]}
-
-  linux-jammy-aarch64-py3_10-test:
-    name: linux-jammy-aarch64-py3.10
-    uses: ./.github/workflows/_linux-test.yml
-    needs: linux-jammy-aarch64-py3_10-build
-    permissions:
-      id-token: write
-      contents: read
-    with:
-      build-environment: linux-jammy-aarch64-py3.10
-      docker-image: ${{ needs.linux-jammy-aarch64-py3_10-build.outputs.docker-image }}
-      test-matrix: ${{ needs.linux-jammy-aarch64-py3_10-build.outputs.test-matrix }}

.github/workflows/mac-mps.yml

@@ -13,31 +13,33 @@ concurrency:
 permissions: read-all
 
 jobs:
-  macos-py3-arm64-build:
-    name: macos-py3-arm64
+  macos-12-py3-arm64-build:
+    name: macos-12-py3-arm64
     uses: ./.github/workflows/_mac-build.yml
     with:
-      sync-tag: macos-py3-arm64-build
-      build-environment: macos-py3-arm64
+      sync-tag: macos-12-py3-arm64-build
+      build-environment: macos-12-py3-arm64
       runner-type: macos-m1-stable
       build-generates-artifacts: true
       # To match the one pre-installed in the m1 runners
       python-version: 3.9.12
-      # The runner macos-m2-14 is not a typo, it's a custom runner that is different
-      # than our AWS macos-m1-14 runners
+      # We need to set the environment file here instead of trying to detect it automatically because
+      # MacOS arm64 is cross-compiled from x86-64. Specifically, it means that arm64 conda environment
+      # is needed when building PyTorch MacOS arm64 from x86-64
+      environment-file: .github/requirements/conda-env-macOS-ARM64
       test-matrix: |
         { include: [
-          { config: "mps", shard: 1, num_shards: 1, runner: "macos-m1-13" },
+          { config: "mps", shard: 1, num_shards: 1, runner: "macos-m1-stable" },
           { config: "mps", shard: 1, num_shards: 1, runner: "macos-m2-14" },
         ]}
 
-  macos-py3-arm64-mps-test:
-    name: macos-py3-arm64-mps
+  macos-12-py3-arm64-mps-test:
+    name: macos-12-py3-arm64-mps
     uses: ./.github/workflows/_mac-test-mps.yml
-    needs: macos-py3-arm64-build
+    needs: macos-12-py3-arm64-build
     with:
-      sync-tag: macos-py3-arm64-mps-test
-      build-environment: macos-py3-arm64
+      sync-tag: macos-12-py3-arm64-mps-test
+      build-environment: macos-12-py3-arm64
       # Same as the build job
       python-version: 3.9.12
-      test-matrix: ${{ needs.macos-py3-arm64-build.outputs.test-matrix }}
+      test-matrix: ${{ needs.macos-12-py3-arm64-build.outputs.test-matrix }}