Skip to content

CI: split per-arch NPU/sim workflows and merge OS setup steps - #1687

Merged
ChaoWao merged 1 commit into
hw-native-sys:mainfrom
ChaoWao:support/split-per-arch-ci-workflows
Aug 5, 2026
Merged

CI: split per-arch NPU/sim workflows and merge OS setup steps#1687
ChaoWao merged 1 commit into
hw-native-sys:mainfrom
ChaoWao:support/split-per-arch-ci-workflows

Conversation

@ChaoWao

@ChaoWao ChaoWao commented Aug 4, 2026

Copy link
Copy Markdown
Collaborator

Summary

Follow-up to #1640, which consolidated the a2a3/a5 NPU and sim jobs into shared reusable workflows branching on platform / runner.os via if:. Behavior was preserved, but each job rendered the other variant's steps as skipped rows in the Actions UI — st-onboard-a2a3 showed 7 skipped a5/legacy rows, and packaging/ut showed skipped Linux/macOS compiler sibling rows.

  • Split the arch-specific workflows one file per platform so each job lists only its own steps:
    • _st-npu.yml_st-npu-a2a3.yml + _st-npu-a5.yml
    • _st-sim.yml_st-sim-a2a3.yml + _st-sim-a5.yml
    • _ut-npu.yml_ut-npu-a2a3.yml + _ut-npu-a5.yml
    • The platform input is dropped (now implied by the file) and the platform value hardcoded. Lane-correlated gates (a2a3_sdma_mode, include_dfx_smokes, include_cann_examples) remain as the accepted residual skip.
  • Merge the OS-conditional setup pairs into single uname-branching steps: _packaging.yml (C++ compiler) and _ut-no-hardware.yml (GoogleTest).
  • Doc fix: PTO2_*_TIMEOUT_*SIMPLER_*_TIMEOUT_* in docs/ci.md and docs/troubleshooting/device-error-codes/stall.md — the PTO2_* timeout env vars do not exist; the runtime getenv-reads SIMPLER_*.

Effect (github lane — what every PR sees)

  • st-onboard-a2a3: 7 → 2 skipped (legacy SDMA mode only)
  • st-onboard-a5 / st-sim-* / ut-*: → 0 skipped
  • packaging / ut: no more Linux/macOS sibling skip

Testing

  • yaml.safe_load over all workflows + composite actions; actionlint clean (only pre-existing shellcheck style notes + the known self-hosted cpu label)
  • task-submit invocation counts unchanged (_st-npu 14→14, _ut-npu 4→4); no test steps lost
  • Zero residual inputs.platform in split files; zero runner.os == Linux/macOS step pairs; callers pass no dangling platform:
  • Observe on a PR that touches a2a3 code: st-onboard-a2a3 shows ~12 steps with only 2 skipped (legacy), and st-onboard-a5 shows 0 skipped

@coderabbitai

coderabbitai Bot commented Aug 4, 2026

Copy link
Copy Markdown

Review Change Stack

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: d18ae911-e85c-4570-839e-af8b9b7bbaa0

📥 Commits

Reviewing files that changed from the base of the PR and between f8106e2 and b5c87dd.

📒 Files selected for processing (12)
  • .github/workflows/_packaging.yml
  • .github/workflows/_st-npu-a2a3.yml
  • .github/workflows/_st-npu-a5.yml
  • .github/workflows/_st-sim-a2a3.yml
  • .github/workflows/_st-sim-a5.yml
  • .github/workflows/_ut-no-hardware.yml
  • .github/workflows/_ut-npu-a2a3.yml
  • .github/workflows/_ut-npu-a5.yml
  • .github/workflows/ci-self-cpu.yml
  • .github/workflows/ci.yml
  • docs/ci.md
  • docs/troubleshooting/device-error-codes/stall.md

📝 Walkthrough

Walkthrough

The PR splits simulator and NPU CI workflows by a2a3 and a5 architecture. It adds dedicated A5 workflows, specializes existing a2a3 workflows, updates CI dispatch, consolidates setup steps, and updates CI timeout documentation.

Changes

Architecture-specific CI workflows

Layer / File(s) Summary
Dedicated A5 workflows
.github/workflows/_st-sim-a5.yml, .github/workflows/_st-npu-a5.yml
Adds A5 simulator scene-test and NPU unit-test workflows with environment setup, test execution, and optional DFX smoke tests.
Specialized a2a3 workflows
.github/workflows/_st-sim-a2a3.yml, .github/workflows/_st-npu-a2a3.yml, .github/workflows/_ut-npu-a2a3.yml
Removes platform inputs and fixes each workflow to its a2a3 target. Existing test and smoke conditions use architecture-specific inputs.
CI workflow dispatch
.github/workflows/ci.yml, .github/workflows/ci-self-cpu.yml
Routes simulator and NPU jobs to dedicated architecture-specific reusable workflows.
Shared setup and documentation
.github/workflows/_packaging.yml, .github/workflows/_ut-no-hardware.yml, docs/ci.md, docs/troubleshooting/device-error-codes/stall.md
Consolidates OS-specific setup steps and replaces PTO2_* timeout names with SIMPLER_* names in documentation.

Estimated code review effort: 4 (Complex) | ~45 minutes

Sequence Diagram(s)

sequenceDiagram
  participant CI
  participant ReusableWorkflow
  participant TestRunner
  CI->>ReusableWorkflow: call architecture-specific workflow
  ReusableWorkflow->>TestRunner: configure environment and tests
  TestRunner->>ReusableWorkflow: return execution status
Loading

Possibly related PRs

Poem

A rabbit hops through CI lanes,
Splitting paths for chips and brains.
A2A3 here, A5 beside,
Smoke tests run with flags applied.
Timeout names now clearly gleam—
Carrots cheer the workflow stream!

🚥 Pre-merge checks | ✅ 5
✅ Passed checks (5 passed)
Check name Status Explanation
Title check ✅ Passed The title clearly summarizes the two main changes: splitting architecture-specific workflows and merging operating-system setup steps.
Description check ✅ Passed The description directly explains the workflow split, setup-step consolidation, documentation fixes, and validation results.
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

workflows branching on platform / runner.os via if:. Behavior was
preserved, but each job rendered the other variant's steps as skipped
rows in the Actions UI — st-onboard-a2a3 showed 7 skipped a5/legacy rows,
and packaging/ut showed skipped Linux/macOS compiler sibling rows.

Split the arch-specific workflows one file per platform so each job lists
only its own steps:
  _st-npu.yml -> _st-npu-a2a3.yml + _st-npu-a5.yml
  _st-sim.yml -> _st-sim-a2a3.yml + _st-sim-a5.yml
  _ut-npu.yml -> _ut-npu-a2a3.yml + _ut-npu-a5.yml
The platform input is dropped (now implied by the file) and the platform
value hardcoded. Lane-correlated gates (a2a3_sdma_mode, include_dfx_smokes,
include_cann_examples) remain as the accepted residual skip.

Merge the OS-conditional setup pairs into single uname-branching steps:
_packaging.yml (C++ compiler) and _ut-no-hardware.yml (GoogleTest).

Also correct a stale env-var name in docs (PTO2_*_TIMEOUT_* ->
SIMPLER_*_TIMEOUT_*); the PTO2_* timeout env vars do not exist, the
runtime reads SIMPLER_*.

Effect on the github lane: st-onboard-a2a3 7->2 skipped, st-onboard-a5 /
st-sim / ut -> 0 skipped; packaging/ut drop the OS sibling skip.

Verified: YAML parse + actionlint clean; task-submit invocation counts
unchanged (_st-npu 14, _ut-npu 4); zero residual inputs.platform and
runner.os Linux/macOS step pairs.
@ChaoWao
ChaoWao force-pushed the support/split-per-arch-ci-workflows branch from 3d93dc3 to b5c87dd Compare August 5, 2026 00:51
@ChaoWao
ChaoWao merged commit 06fbd42 into hw-native-sys:main Aug 5, 2026
18 checks passed
@ChaoWao
ChaoWao deleted the support/split-per-arch-ci-workflows branch August 5, 2026 01:12
yanghaoran29 pushed a commit to yanghaoran29/simpler that referenced this pull request Aug 6, 2026
…ive-sys#1687)

workflows branching on platform / runner.os via if:. Behavior was
preserved, but each job rendered the other variant's steps as skipped
rows in the Actions UI — st-onboard-a2a3 showed 7 skipped a5/legacy rows,
and packaging/ut showed skipped Linux/macOS compiler sibling rows.

Split the arch-specific workflows one file per platform so each job lists
only its own steps:
  _st-npu.yml -> _st-npu-a2a3.yml + _st-npu-a5.yml
  _st-sim.yml -> _st-sim-a2a3.yml + _st-sim-a5.yml
  _ut-npu.yml -> _ut-npu-a2a3.yml + _ut-npu-a5.yml
The platform input is dropped (now implied by the file) and the platform
value hardcoded. Lane-correlated gates (a2a3_sdma_mode, include_dfx_smokes,
include_cann_examples) remain as the accepted residual skip.

Merge the OS-conditional setup pairs into single uname-branching steps:
_packaging.yml (C++ compiler) and _ut-no-hardware.yml (GoogleTest).

Also correct a stale env-var name in docs (PTO2_*_TIMEOUT_* ->
SIMPLER_*_TIMEOUT_*); the PTO2_* timeout env vars do not exist, the
runtime reads SIMPLER_*.

Effect on the github lane: st-onboard-a2a3 7->2 skipped, st-onboard-a5 /
st-sim / ut -> 0 skipped; packaging/ut drop the OS sibling skip.

Verified: YAML parse + actionlint clean; task-submit invocation counts
unchanged (_st-npu 14, _ut-npu 4); zero residual inputs.platform and
runner.os Linux/macOS step pairs.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant