Ascend NPU Test Suite #4046
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
name: "Ascend NPU Test Suite" | ||
on: | ||
push: | ||
branches: | ||
- "main" | ||
paths: | ||
- ".github/workflows/ascend_npu_test.yml" | ||
- ".github/workflows/_ascend_npu_build_torch.yml" | ||
- ".github/workflows/_ascend_npu_build_torch_npu.yml" | ||
- ".github/workflows/_ascend_npu_ut.yml" | ||
- ".github/workflows/_ascend_npu_benchmark.yml" | ||
- ".github/workflows/_ascend_npu_torchtitan.yml" | ||
- ".ci/**" | ||
- "ascend_npu/**" | ||
- "src/**" | ||
- "!**/*.md" | ||
pull_request: | ||
branches: | ||
- "main" | ||
paths: | ||
- ".github/workflows/ascend_npu_test.yml" | ||
- ".github/workflows/_ascend_npu_build_torch.yml" | ||
- ".github/workflows/_ascend_npu_build_torch_npu.yml" | ||
- ".github/workflows/_ascend_npu_ut.yml" | ||
- ".github/workflows/_ascend_npu_benchmark.yml" | ||
- ".github/workflows/_ascend_npu_torchtitan.yml" | ||
- ".ci/**" | ||
- "ascend_npu/**" | ||
- "src/**" | ||
- "!**/*.md" | ||
release: | ||
types: | ||
- "published" | ||
schedule: | ||
- cron: "0 12 * * *" | ||
repository_dispatch: | ||
types: [pytorch-pr-event-redispatch] | ||
workflow_dispatch: | ||
inputs: | ||
runner: | ||
required: true | ||
type: choice | ||
options: | ||
- npu-arm64 | ||
- linux-arm64-npu-0 | ||
- linux-arm64-npu-1 | ||
- linux-arm64-npu-2 | ||
- linux-arm64-npu-4 | ||
default: "linux-arm64-npu-1" | ||
description: "The runner selected to run on" | ||
image: | ||
required: true | ||
type: choice | ||
options: | ||
- ascendai/cann:7.0.1-910b-ubuntu22.04-py3.8 | ||
- ascendai/cann:8.0.0-910b-ubuntu22.04-py3.10 | ||
- ascendai/cann:latest | ||
default: "ascendai/cann:latest" | ||
description: "The docker image which will be loaded" | ||
# Only cancel the previous runs when triggered by a pull_request event | ||
# | ||
# TODO: As the doc says, there can be at most one running and one pending job | ||
# in a concurrency group at any time. But we want all workflows to be | ||
# queued, not cancelled. This is a shortcomings of GitHub Actions. | ||
# | ||
# Doc: https://docs.github.com/en/actions/writing-workflows/workflow-syntax-for-github-actions#concurrency | ||
# Discussion: https://github.com/orgs/community/discussions/5435 | ||
concurrency: | ||
group: ${{ github.workflow }}-${{ github.event_name }}-${{ github.event.client_payload.pull_request.number || github.ref }} | ||
cancel-in-progress: ${{ github.event_name == 'pull_request' }} | ||
jobs: | ||
prepare: | ||
name: Prepare | ||
runs-on: ubuntu-latest | ||
outputs: | ||
runner: ${{ steps.set-env.outputs.runner }} | ||
image: ${{ steps.set-env.outputs.image }} | ||
device: ${{ steps.set-env.outputs.device }} | ||
ref: ${{ steps.list-ref.outputs.ref }} | ||
pr-number: ${{ steps.list-ref.outputs.pr_number }} | ||
steps: | ||
- name: Set environment params | ||
id: set-env | ||
run: | | ||
echo "runner=${{ github.event.inputs.runner || 'linux-arm64-npu-1' }}" >> $GITHUB_OUTPUT | ||
echo "image=${{ github.event.inputs.image || 'ascendai/cann:latest' }}" >> $GITHUB_OUTPUT | ||
# TODO(shink): List ghstack PR's ref | ||
- name: List ref to the PyTorch branch | ||
id: list-ref | ||
if: ${{ github.event_name == 'repository_dispatch' }} | ||
run: | | ||
echo "pr_number=${{ github.event.client_payload.pull_request.number }}" >> $GITHUB_OUTPUT | ||
echo "ref=refs/pull/${{ github.event.client_payload.pull_request.number }}/merge" >> $GITHUB_OUTPUT | ||
build-torch: | ||
name: Build torch | ||
needs: | ||
- prepare | ||
if: ${{ needs.prepare.outputs.ref }} | ||
uses: ./.github/workflows/_ascend_npu_build_torch.yml | ||
with: | ||
runner: ${{ needs.prepare.outputs.runner }} | ||
image: ${{ needs.prepare.outputs.image }} | ||
ref: ${{ needs.prepare.outputs.ref }} | ||
pr-number: ${{ needs.prepare.outputs.pr-number }} | ||
build: | ||
name: Build torch_npu | ||
needs: | ||
- prepare | ||
- build-torch | ||
if: ${{ !cancelled() && (success() || needs.build-torch.result == 'skipped') }} | ||
uses: ./.github/workflows/_ascend_npu_build_torch_npu.yml | ||
with: | ||
runner: ${{ needs.prepare.outputs.runner }} | ||
image: ${{ needs.prepare.outputs.image }} | ||
torch-artifact: ${{ needs.build-torch.outputs.torch-artifact }} | ||
torchtitan: | ||
name: Run torchtitan | ||
needs: | ||
- prepare | ||
- build-torch | ||
- build | ||
if: | | ||
!cancelled() && github.event_name != 'repository_dispatch' && | ||
(success() || (needs.build-torch.result == 'skipped' && needs.build.result == 'success')) | ||
uses: ./.github/workflows/_ascend_npu_torchtitan.yml | ||
Check failure on line 132 in .github/workflows/ascend_npu_test.yml
|
||
with: | ||
runner: ${{ needs.prepare.outputs.runner }} | ||
image: ${{ needs.prepare.outputs.image }} | ||
torch-artifact: ${{ needs.build-torch.outputs.torch-artifact }} | ||
torch-npu-artifact: ${{ needs.build.outputs.torch-npu-artifact }} | ||