Skip to content

[AMD][CI] Consolidate AMD workflows and retire ROCm 7.0 CI - #38632

Merged
yctseng0211 merged 18 commits into
mainfrom
codex/rocm10-ci-default
Sep 15, 2026
Merged

yctseng0211 merged 18 commits into
mainfrom
codex/rocm10-ci-default

Conversation

@bingxche

@bingxche bingxche commented Sep 9, 2026 •

Copy link
Copy Markdown
Collaborator

Purpose

Consolidate AMD workflows to one pair of yml files pr-test-amd.yml and nightly-test-amd.yml

Removes the duplicate ROCm 7.0 workflows and uses one standard PR/Nightly workflow for every supported ROCm version.

After this PR:

  • Normal PR CI and manually started Nightly tests default to ROCm 10.
  • Scheduled Nightly tests ROCm 10, ROCm 7.2.4, and ROCm 7.2.0.
  • Weekly AITER Scout tests ROCm 10, ROCm 7.2.4 and ROCm 7.2.0
  • Standard ROCm 7.0 CI and Docker image builds are removed.
  • ROCm 10, ROCm 7.2.4, and ROCm 7.2.0 remain supported through the same workflow files.

Lint and workflow validation pass.


CI States

Latest PR Test (Base): ✅ Run #34988389638
Latest PR Test (Extra): ❌ Run #34988389125
Latest PR Test (AMD ROCm 10): ➖ No AMD PR run found for this commit.

@github-actions github-actions Bot added the amd label Sep 9, 2026
@bingxche bingxche closed this Sep 9, 2026
@bingxche bingxche changed the title [AMD][CI] Make ROCm 10 the Default for AMD PR and Nightly Tests [AMD][CI] Consolidate AMD CI and Docker Images on ROCm 10 Sep 9, 2026
@bingxche bingxche reopened this Sep 9, 2026
@github-actions github-actions Bot added the documentation Improvements or additions to documentation label Sep 9, 2026
@bingxche bingxche changed the title [AMD][CI] Consolidate AMD CI and Docker Images on ROCm 10 [AMD][CI] Consolidate AMD workflows and retire ROCm 7.0 CI Sep 9, 2026
@bingxche
bingxche marked this pull request as ready for review September 10, 2026 13:53
cursor Bot pushed a commit that referenced this pull request Sep 24, 2026
Resolve the modify/delete conflict on
.github/workflows/nightly-test-amd-rocm720.yml, which #38632 folded into
nightly-test-amd.yml. Port the two Qwen3.8-Flash-Next-FP8 jobs there with
the same renames #38632 applied to every other job: drop the -rocm720
job-ID suffix, use the shared rocm10/rocm724/rocm720 matrix, and say
"ROCm" instead of "ROCm 7.2" in step names.

Co-authored-by: quitenode <quitenode@users.noreply.github.com>
michaelzhang-ai added a commit that referenced this pull request Sep 27, 2026
main retired nightly-test-amd-rocm720.yml in #38632, so move the GLM-5.2-MXFP4
agentic MTP job into nightly-test-amd.yml as
nightly-8-gpu-mi35x-glm52-fp4-agentic-mtp, following the consolidated
workflow's naming. The job body is unchanged.

Co-Authored-By: Claude <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

amd documentation Improvements or additions to documentation

Projects

None yet

Development

Successfully merging this pull request may close these issues.

7 participants