pytorch/.github/workflows/generated-linux-binary-conda.yml
Eli Uriegas 9c8fb2ee2d .github: Consolidate binary checkout logic
Consolidates binary checkout logic to use the standard common logic we
have in our common templates. Also fixes issues related to
pytorch/builder trying to checkout the head commit for pytorch/pytorch
instead of checking out the builder commit we actually want

Signed-off-by: Eli Uriegas <eliuriegasfb.com>

Pull Request resolved: https://github.com/pytorch/pytorch/pull/73092

Signed-off-by: Eli Uriegas <eliuriegas@fb.com>
2022-02-18 19:03:28 +00:00

8287 lines
338 KiB
YAML
Generated

# @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-binary-conda
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_conda/*'
workflow_dispatch:
env:
# Needed for conda builds
ALPINE_IMAGE: "308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/alpine"
ANACONDA_USER: pytorch
AWS_DEFAULT_REGION: us-east-1
BINARY_ENV_FILE: /tmp/env
BUILD_ENVIRONMENT: linux-binary-conda
BUILDER_ROOT: /builder
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
IN_CI: 1
IS_GHA: 1
PR_LABELS: ${{ toJson(github.event.pull_request.labels.*.name) }}
PR_NUMBER: ${{ github.event.pull_request.number }}
PYTORCH_FINAL_PACKAGE_DIR: /artifacts
PYTORCH_RETRY_TEST_CASES: 1
PYTORCH_ROOT: /pytorch
SHA1: ${{ github.event.pull_request.head.sha || github.sha }}
SKIP_ALL_TESTS: 1
concurrency:
group: linux-binary-conda-${{ github.event.pull_request.number || github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
cancel-in-progress: true
jobs:
conda-py3_7-cpu-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_7-cpu
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cpu-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cpu-build
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cpu
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cpu-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cpu-test
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cpu
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda10_2-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_7-cuda10_2
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda10_2-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cuda10_2-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cuda10_2
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda10_2-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cuda10_2-test
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cuda10_2
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_1-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_7-cuda11_1
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_1-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cuda11_1-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cuda11_1
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_1-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cuda11_1-test
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cuda11_1
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_3-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_7-cuda11_3
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_3-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cuda11_3-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cuda11_3
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_3-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cuda11_3-test
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cuda11_3
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_5-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_7-cuda11_5
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_5-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cuda11_5-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cuda11_5
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_7-cuda11_5-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_7-cuda11_5-test
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.7"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_7-cuda11_5
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cpu-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_8-cpu
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cpu-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cpu-build
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cpu
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cpu-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cpu-test
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cpu
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda10_2-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_8-cuda10_2
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda10_2-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cuda10_2-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cuda10_2
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda10_2-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cuda10_2-test
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cuda10_2
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_1-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_8-cuda11_1
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_1-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cuda11_1-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cuda11_1
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_1-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cuda11_1-test
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cuda11_1
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_3-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_8-cuda11_3
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_3-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cuda11_3-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cuda11_3
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_3-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cuda11_3-test
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cuda11_3
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_5-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_8-cuda11_5
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_5-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cuda11_5-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cuda11_5
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_8-cuda11_5-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_8-cuda11_5-test
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_8-cuda11_5
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cpu-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_9-cpu
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cpu-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cpu-build
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cpu
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cpu-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cpu-test
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cpu
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda10_2-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_9-cuda10_2
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda10_2-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cuda10_2-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cuda10_2
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda10_2-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cuda10_2-test
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cuda10_2
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_1-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_9-cuda11_1
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_1-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cuda11_1-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cuda11_1
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_1-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cuda11_1-test
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cuda11_1
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_3-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_9-cuda11_3
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_3-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cuda11_3-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cuda11_3
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_3-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cuda11_3-test
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cuda11_3
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_5-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_9-cuda11_5
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_5-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cuda11_5-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cuda11_5
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_9-cuda11_5-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_9-cuda11_5-test
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.9"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_9-cuda11_5
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cpu-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_10-cpu
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cpu-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cpu-build
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cpu
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cpu-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cpu-test
env:
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: cpu
GPU_ARCH_TYPE: cpu
DOCKER_IMAGE: pytorch/conda-builder:cpu
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cpu
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda10_2-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_10-cuda10_2
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda10_2-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cuda10_2-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cuda10_2
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda10_2-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cuda10_2-test
env:
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: cu102
GPU_ARCH_VERSION: 10.2
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda10.2
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cuda10_2
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_1-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_10-cuda11_1
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_1-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cuda11_1-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cuda11_1
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_1-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cuda11_1-test
env:
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: cu111
GPU_ARCH_VERSION: 11.1
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.1
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cuda11_1
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_3-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_10-cuda11_3
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_3-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cuda11_3-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cuda11_3
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_3-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cuda11_3-test
env:
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: cu113
GPU_ARCH_VERSION: 11.3
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.3
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cuda11_3
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_5-build:
if: ${{ github.repository_owner == 'pytorch' }}
runs-on: linux.4xlarge
timeout-minutes: 240
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Set BUILD_SPLIT_CUDA
run: |
echo "BUILD_SPLIT_CUDA='ON'" >> "$GITHUB_ENV"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Build PyTorch binary
run: |
set -x
mkdir -p artifacts/
container_name=$(docker run \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
docker exec -t "${container_name}" bash -c "source ${BINARY_ENV_FILE} && bash /builder/conda/build.sh"
- name: Chown artifacts
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "${RUNNER_TEMP}/artifacts:/v" -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- uses: seemethere/upload-artifact-s3@v3
with:
name: conda-py3_10-cuda11_5
retention-days: 14
if-no-files-found: error
path:
${{ runner.temp }}/artifacts/*
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_5-test: # Testing
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cuda11_5-build
runs-on: linux.4xlarge.nvidia.gpu
timeout-minutes: 240
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cuda11_5
path: "${{ runner.temp }}/artifacts/"
- name: Checkout PyTorch
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
submodules: recursive
path: pytorch
- name: Clean PyTorch checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: pytorch
- name: Checkout pytorch/builder
uses: zhouzhuojie/checkout@05b13c9a0d21f08f6d5e64a1d5042246d13619d9
with:
ref: main
submodules: recursive
repository: pytorch/builder
path: builder
- name: Clean pytorch/builder checkout
run: |
# Remove any artifacts from the previous checkouts
git clean -fxd
working-directory: builder
- name: Install nvidia driver, nvidia-docker runtime, set GPU_FLAG
working-directory: pytorch/
run: |
bash .github/scripts/install_nvidia_utils_linux.sh
echo "GPU_FLAG=--gpus all" >> "${GITHUB_ENV}"
- name: Pull Docker image
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${DOCKER_IMAGE}"
- name: Test PyTorch binary
run: |
set -x
# shellcheck disable=SC2086,SC2090
container_name=$(docker run \
${GPU_FLAG:-} \
-e BINARY_ENV_FILE \
-e BUILDER_ROOT \
-e BUILD_ENVIRONMENT \
-e BUILD_SPLIT_CUDA \
-e DESIRED_CUDA \
-e DESIRED_DEVTOOLSET \
-e DESIRED_PYTHON \
-e GPU_ARCH_TYPE \
-e GPU_ARCH_VERSION \
-e IS_GHA \
-e LIBTORCH_VARIANT \
-e PACKAGE_TYPE \
-e PYTORCH_FINAL_PACKAGE_DIR \
-e PYTORCH_ROOT \
-e SKIP_ALL_TESTS \
--tty \
--detach \
-v "${GITHUB_WORKSPACE}/pytorch:/pytorch" \
-v "${GITHUB_WORKSPACE}/builder:/builder" \
-v "${RUNNER_TEMP}/artifacts:/final_pkgs" \
-w / \
"${DOCKER_IMAGE}"
)
docker exec -t -w "${PYTORCH_ROOT}" "${container_name}" bash -c "bash .circleci/scripts/binary_populate_env.sh"
# Generate test script
docker exec -t -w "${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 ${BINARY_ENV_FILE} && bash -x /run.sh"
- name: Hold runner for 2 hours or until ssh sessions have drained
working-directory: pytorch/
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af
conda-py3_10-cuda11_5-upload: # Uploading
runs-on: linux.2xlarge # self hosted runner to download ec2 artifacts
if: ${{ github.repository_owner == 'pytorch' }}
needs: conda-py3_10-cuda11_5-test
env:
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: cu115
GPU_ARCH_VERSION: 11.5
GPU_ARCH_TYPE: cuda
DOCKER_IMAGE: pytorch/conda-builder:cuda11.5
SKIP_ALL_TESTS: 1
DESIRED_PYTHON: "3.10"
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: $(get_ec2_metadata instance-type)"
- name: Log in to ECR
env:
AWS_RETRY_MODE: standard
AWS_MAX_ATTEMPTS: 5
run: |
AWS_ACCOUNT_ID=$(aws sts get-caller-identity|grep Account|cut -f4 -d\")
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry aws ecr get-login-password --region "$AWS_DEFAULT_REGION" | docker login --username AWS \
--password-stdin "$AWS_ACCOUNT_ID.dkr.ecr.$AWS_DEFAULT_REGION.amazonaws.com"
- name: Chown workspace
run: |
retry () {
"$@" || (sleep 1 && "$@") || (sleep 2 && "$@")
}
retry docker pull "${ALPINE_IMAGE}"
# Ensure the working directory gets chowned back to the current user
docker run --pull=never --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Clean workspace
run: |
rm -rf "${GITHUB_WORKSPACE}"
mkdir "${GITHUB_WORKSPACE}"
- name: "[FB EMPLOYEES] Enable SSH (Click me for login details)"
uses: seemethere/add-github-ssh-key@v1
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Preserve github env variables for use in docker
run: |
env | grep '^GITHUB' > "/tmp/github_env_${GITHUB_RUN_ID}"
- name: Clone pytorch/pytorch
uses: actions/checkout@v2
- uses: seemethere/download-artifact-s3@0504774707cbc8603d7dca922e8026eb8bf3b47b
name: Download Build Artifacts
with:
name: conda-py3_10-cuda11_5
path: "${{ runner.temp }}/artifacts/"
- name: Set DRY_RUN (only for tagged pushes)
if: ${{ github.event_name == 'push' && (github.event.ref == 'refs/heads/nightly' || (startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/'))) }}
run: |
echo "DRY_RUN=disabled" >> "$GITHUB_ENV"
- name: Set UPLOAD_CHANNEL (only for tagged pushes)
if: ${{ github.event_name == 'push' && startsWith(github.event.ref, 'refs/tags/') && !startsWith(github.event.ref, 'refs/tags/ciflow/') }}
run: |
# reference ends with an RC suffix
if [[ ${GITHUB_REF_NAME} = *-rc[0-9]* ]]; then
echo "UPLOAD_CHANNEL=test" >> "$GITHUB_ENV"
fi
- name: Upload binaries
env:
PKG_DIR: "${{ runner.temp }}/artifacts"
UPLOAD_SUBFOLDER: "${{ env.DESIRED_CUDA }}"
# When running these on pull_request events these should be blank
AWS_ACCESS_KEY_ID: ${{ secrets.AWS_PYTORCH_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_PYTORCH_SECRET_KEY }}
ANACONDA_API_TOKEN: ${{ secrets.CONDA_PYTORCHBOT_TOKEN }}
run: |
docker run --rm -i \
-e ANACONDA_API_TOKEN \
-e AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY \
-e DRY_RUN \
-e PACKAGE_TYPE \
-e PKG_DIR=/artifacts \
-e UPLOAD_CHANNEL \
-e UPLOAD_SUBFOLDER \
-v "${RUNNER_TEMP}/artifacts:/artifacts" \
-v "${GITHUB_WORKSPACE}:/v" \
-w /v \
308535385114.dkr.ecr.us-east-1.amazonaws.com/tool/miniconda3:4.10.3 \
bash -c '.circleci/scripts/binary_upload.sh'
- name: Hold runner for 2 hours or until ssh sessions have drained
# Always hold for active ssh sessions
if: always()
run: .github/scripts/wait_for_ssh_to_drain.sh
- name: Chown workspace
if: always()
run: |
# Ensure the working directory gets chowned back to the current user
docker run --rm -v "$(pwd)":/v -w /v "${ALPINE_IMAGE}" chown -R "$(id -u):$(id -g)" .
- name: Kill containers, clean up images
if: always()
run: |
# ignore expansion of "docker ps -q" since it could be empty
# shellcheck disable=SC2046
docker stop $(docker ps -q) || true
# Prune all of the docker images
docker system prune -af