Reuse _linux-build.yml for ppc64le build and updated other files accordingly

Author: sandeepgupta12

.ci/docker/manywheel/Dockerfile_ppc64le

@@ -1,31 +1,92 @@
-# Use UBI 9 as base image
-FROM registry.access.redhat.com/ubi9/ubi:9.5 AS base
+# Use the manylinux_2_28 base image for ppc64le
+FROM quay.io/pypa/manylinux_2_28_ppc64le as base
 
-# Install necessary dependencies
-RUN dnf install -y \
-    https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm && \
-    dnf install -y git cmake ninja-build gcc-toolset-13 rust cargo zip \
-    python3 python3-devel && \
-    dnf clean all
+# Language variables
+ENV LC_ALL=C.UTF-8
+ENV LANG=C.UTF-8
+ENV LANGUAGE=C.UTF-8
 
-ENV PATH="/opt/rh/gcc-toolset-13/root/usr/bin:$PATH"
-ENV MANPATH="/opt/rh/gcc-toolset-13/root/usr/share/man"
-ENV INFOPATH="/opt/rh/gcc-toolset-13/root/usr/share/info"
-ENV PCP_DIR="/opt/rh/gcc-toolset-13/root"
-ENV LD_LIBRARY_PATH="/opt/rh/gcc-toolset-13/root/usr/lib64:/opt/rh/gcc-toolset-13/root/usr/lib"
+ARG DEVTOOLSET_VERSION=13
 
-# Set Python and pip aliases to use Python 3.9
-RUN ln -sf /usr/bin/python3 /usr/bin/python && \
-    ln -sf /usr/bin/pip3 /usr/bin/pip
+# Create symbolic links for Python 3.12
+RUN ln -sf /opt/python/cp312-cp312/bin/python3.12 /usr/bin/python3 && \
+    ln -sf /opt/python/cp312-cp312/bin/python3.12 /usr/bin/python
 
-COPY requirements.txt .
-# Install Python packages via pip
-RUN pip install wheel 
-RUN pip install -r requirements.txt
+# Install required system dependencies
+RUN yum -y install epel-release && \
+    yum -y update && \
+    yum install -y \
+      sudo \
+      autoconf \
+      automake \
+      bison \
+      bzip2 \
+      curl \
+      diffutils \
+      file \
+      git \
+      make \
+      patch \
+      perl \
+      unzip \
+      util-linux \
+      wget \
+      which \
+      xz \
+      yasm \
+      less \
+      zstd \
+      libgomp \
+      gcc-toolset-${DEVTOOLSET_VERSION}-gcc \
+      gcc-toolset-${DEVTOOLSET_VERSION}-gcc-c++ \
+      gcc-toolset-${DEVTOOLSET_VERSION}-binutils \
+      gcc-toolset-${DEVTOOLSET_VERSION}-gcc-gfortran \
+      cmake \
+      ninja-build \
+      rust \
+      cargo \
+      llvm-devel \
+      libzstd-devel \
+      python3.12-devel \
+      python3.12-setuptools \
+      python3.12-pip \
+      python3-virtualenv \
+      python3.12-pyyaml \
+      python3.12-numpy \
+      python3.12-wheel \
+      python3.12-cryptography \
+      blas-devel \
+      openblas-devel \
+      lapack-devel \
+      atlas-devel \
+      libjpeg-devel \
+      libxslt-devel \
+      libxml2-devel \
+      openssl-devel \
+      valgrind
+  
 
-RUN mkdir -p /workspace/pytorch
+# Ensure the correct Python version is used
+ENV PATH=/opt/python/cp312-cp312/bin:$PATH
+# Add gcc-toolset to the path
+ENV PATH=/opt/rh/gcc-toolset-${DEVTOOLSET_VERSION}/root/usr/bin:$PATH
+ENV LD_LIBRARY_PATH=/opt/rh/gcc-toolset-${DEVTOOLSET_VERSION}/root/usr/lib64:/opt/rh/gcc-toolset-${DEVTOOLSET_VERSION}/root/usr/lib:$LD_LIBRARY_PATH
 
-ENTRYPOINT []
-CMD ["/bin/bash"]
+# Configure git to avoid safe directory issues
+RUN git config --global --add safe.directory "*"
 
+# Install required Python packages
+RUN pip install --upgrade pip
+RUN pip install typing_extensions pyyaml setuptools
+ 
+# Install test dependencies
+RUN dnf install -y \
+      protobuf-devel \
+      protobuf-c-devel \
+      protobuf-lite-devel \
+      wget \
+      patch
 
+# Set default entrypoint
+ENTRYPOINT []
+CMD ["/bin/bash"]

.ci/docker/manywheel/build_scripts/build.sh

@@ -20,7 +20,7 @@ AUTOCONF_HASH=954bd69b391edc12d6a4a51a2dd1476543da5c6bbf05a95b59dc0dd6fd4c2969
 # the final image after compiling Python
 PYTHON_COMPILE_DEPS="zlib-devel bzip2-devel ncurses-devel sqlite-devel readline-devel tk-devel gdbm-devel libpcap-devel xz-devel libffi-devel"
 
-if [ "$(uname -m)" != "s390x" ] ; then
+if [ "$(uname -m)" != "s390x" && "$(uname -m)" != "ppc64le" ] ; then
     PYTHON_COMPILE_DEPS="${PYTHON_COMPILE_DEPS} db4-devel"
 else
     PYTHON_COMPILE_DEPS="${PYTHON_COMPILE_DEPS} libdb-devel"

.ci/pytorch/build.sh

@@ -230,7 +230,7 @@ fi
 
 # Do not change workspace permissions for ROCm and s390x CI jobs
 # as it can leave workspace with bad permissions for cancelled jobs
-if [[ "$BUILD_ENVIRONMENT" != *rocm* && "$BUILD_ENVIRONMENT" != *s390x* && -d /var/lib/jenkins/workspace ]]; then
+if [[ "$BUILD_ENVIRONMENT" != *rocm* && "$BUILD_ENVIRONMENT" != *s390x* && "$BUILD_ENVIRONMENT" != *ppc64le* && -d /var/lib/jenkins/workspace ]]; then
   # Workaround for dind-rootless userid mapping (https://github.com/pytorch/ci-infra/issues/96)
   WORKSPACE_ORIGINAL_OWNER_ID=$(stat -c '%u' "/var/lib/jenkins/workspace")
   cleanup_workspace() {
@@ -274,8 +274,10 @@ else
     # XLA test build fails when WERROR=1
     # set only when building other architectures
     # or building non-XLA tests.
+    # ppc64le builds fail when WERROR=1
     if [[ "$BUILD_ENVIRONMENT" != *rocm*  &&
-          "$BUILD_ENVIRONMENT" != *xla* ]]; then
+          "$BUILD_ENVIRONMENT" != *xla* && 
+          "$BUILD_ENVIRONMENT" != *ppc64le* ]]; then
       if [[ "$BUILD_ENVIRONMENT" != *py3.8* ]]; then
         # Install numpy-2.0.2 for builds which are backward compatible with 1.X
         python -mpip install numpy==2.0.2
@@ -396,6 +398,6 @@ if [[ "$BUILD_ENVIRONMENT" != *libtorch* && "$BUILD_ENVIRONMENT" != *bazel* ]];
   python tools/stats/export_test_times.py
 fi
 # don't do this for bazel or s390x as they don't use sccache
-if [[ "$BUILD_ENVIRONMENT" != *s390x* && "$BUILD_ENVIRONMENT" != *-bazel-* ]]; then
+if [[ "$BUILD_ENVIRONMENT" != *s390x* && "$BUILD_ENVIRONMENT" != *ppc64le* && "$BUILD_ENVIRONMENT" != *-bazel-* ]]; then
   print_sccache_stats
 fi

.github/scripts/ppc64le-build.sh

@@ -3,7 +3,7 @@
 ## Install prerequisites.
 
 ```
-$ sudo dnf install podman podman-docker jq
+$ sudo apt install podman podman-docker jq
 ```
 ## Add services.
 
@@ -14,17 +14,17 @@ $ sudo systemctl daemon-reload
 
 ## Rebuild the image
 
-First build ppc64le builder image `docker.io/pytorch/ubippc64le-builder`,
+First build ppc64le builder image `docker.io/pytorch/manylinuxppc64le-builder`,
 using following commands:
 
 ```
 $ cd ~
 $ git clone https://github.com/pytorch/pytorch
 $ cd pytorch
 $ git submodule update --init --recursive
-$ GPU_ARCH_TYPE=cpu-ppc64le "$(pwd)/.ci/docker/manywheel/build.sh" ubippc64le-builder
-$ docker image tag localhost/pytorch/ubippc64le-builder docker.io/pytorch/ubippc64le-builder:cpu-ppc64le
-$ docker image save -o ~/pytorch-ubi-ppc64le.tar docker.io/pytorch/ubippc64le-builder:cpu-ppc64le
+$ GPU_ARCH_TYPE=cpu-ppc64le "$(pwd)/.ci/docker/manywheel/build.sh" manylinuxppc64le-builder
+$ docker image tag localhost/pytorch/manylinuxppc64le-builder docker.io/pytorch/manylinuxppc64le-builder:cpu-ppc64le
+$ docker image save -o ~/manywheel-ppc64le.tar docker.io/pytorch/manylinuxppc64le-builder:cpu-ppc64le
 ```
 
 Next step is to build `actions-runner` image using:
@@ -60,4 +60,4 @@ $ sudo chmod 755 /usr/local/bin/app_token.sh /usr/local/bin/gh_token_generator.s
 
 ```
 $ sudo systemctl enable --now actions-runner@$NAME
-```
+```

.github/scripts/ppc64le-ci/README.md

@@ -96,7 +96,7 @@ USER runner
 # Set working directory
 WORKDIR /opt/runner
 
-COPY --chown=runner:runner pytorch-ubi-ppc64le.tar /opt/runner/pytorch-ubi-ppc64le.tar
+COPY --chown=runner:runner manywheel-ppc64le.tar /opt/runner/manywheel-ppc64le.tar
 
 # Define entry point and command
 ENTRYPOINT ["/usr/bin/entrypoint"]

.github/scripts/ppc64le-ci/self-hosted-builder/actions-runner.Dockerfile

@@ -3,10 +3,10 @@
 set -e -u
 
 # first import docker image
-if [ -f ./pytorch-ubi-ppc64le.tar ] ; then
-        docker image load --input pytorch-ubi-ppc64le.tar
-        docker image tag docker.io/pytorch/ubippc64le-builder:cpu-ppc64le docker.io/pytorch/ubippc64le-builder:cpu-ppc64le-main
-        rm -f ubi-ppc64le.tar
+if [ -f ./manywheel-ppc64le.tar ] ; then
+        docker image load --input manywheel-ppc64le.tar
+        docker image tag docker.io/pytorch/manylinuxppc64le-builder:cpu-ppc64le docker.io/pytorch/manylinuxppc64le-builder:cpu-ppc64le-main
+        rm -f manywheel-ppc64le.tar
 fi
 
 token_file=registration-token.json