Skip to content

chore: add perf-changelog for MI325X --exclusive flag impact#931

Merged
cquil11 merged 3 commits intomainfrom
fix/mi325x-exclusive
Mar 23, 2026
Merged

chore: add perf-changelog for MI325X --exclusive flag impact#931
cquil11 merged 3 commits intomainfrom
fix/mi325x-exclusive

Conversation

@cquil11
Copy link
Collaborator

@cquil11 cquil11 commented Mar 23, 2026

Summary

  • MI325X runner already has --exclusive in its salloc command
  • Add perf-changelog entry to document impact on non-TP8 configs (gptoss-fp4-mi325x-vllm, minimaxm2.5-fp8-mi325x-vllm)
  • TP8 configs are unaffected since they already use all GPUs on the node

Test plan

  • No code changes to runners — perf-changelog documentation only

🤖 Generated with Claude Code

MI325X already has --exclusive in its salloc command. This documents
the performance impact on non-TP8 configs where node sharing could
cause contention.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@cquil11 cquil11 requested a review from a team March 23, 2026 14:56
@github-actions
Copy link
Contributor

Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook

If it is not, please create a PR first before we can merge your PR into the master branch. Let's ensure that the documentation is first class such that the entire ML community can benefit from your hard work! Thank you

1 similar comment
@github-actions
Copy link
Contributor

Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook

If it is not, please create a PR first before we can merge your PR into the master branch. Let's ensure that the documentation is first class such that the entire ML community can benefit from your hard work! Thank you

@cquil11 cquil11 merged commit 7468fb3 into main Mar 23, 2026
13 checks passed
@cquil11 cquil11 deleted the fix/mi325x-exclusive branch March 23, 2026 15:02
cquil11 added a commit that referenced this pull request Mar 23, 2026
MI325X now has its own PR (#931). Update perf-changelog to list
specific non-TP8 MI355X configs instead of wildcards.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
cquil11 added a commit that referenced this pull request Mar 23, 2026
* slurm command fix

Signed-off-by: seungrokj <seungrok.jung@amd.com>

* update perf changelog

* add --exclusive to mi355x multinode

* fix: scope PR to MI355X only, remove MI325X changes

MI325X now has its own PR (#931). Update perf-changelog to list
specific non-TP8 MI355X configs instead of wildcards.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* chore: clarify perf-changelog description for MI355X --exclusive

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* Update perf-changelog.yaml

---------

Signed-off-by: seungrokj <seungrok.jung@amd.com>
Co-authored-by: Cam Quilici <cjquilici@gmail.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Development

Successfully merging this pull request may close these issues.

1 participant