Feat/low bit ep#1336
Draft
JiaoliangYu wants to merge 13 commits into
Draft
Conversation
Pin each worker to a contiguous core range keyed on its rank at the very top of AsyncIOProc.__init__, before any large allocation, so Linux first-touch also places memory on the local NUMA node. Gated by ATOM_CPU_AFFINITY so baseline vs pinned A/B needs no code change. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Feat/cpu affinity per rank
Add amd-smi performance-determinism lock before the benchmark/server runs and an always() unlock to AUTO before container teardown, in both the main benchmark job and the regression-rerun job. The lock is driver-level and persists across jobs on the bare-metal runner, so the unlock must run even on failure. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…rnel, cast back to bf16 for GEMM
JiaoliangYu
force-pushed
the
feat/mori-fp8-dispatch-clean
branch
from
e4c971c to
2a57892
Compare
JiaoliangYu
force-pushed
the
feat/mori-fp8-dispatch-clean
branch
from
2a57892 to
11a717b
Compare
zufayu
requested review from
ZhangLirong-amd
and removed request for
ZhangLirong-amd
| Returns the (possibly padded) tensor and the original row count so the | ||
| caller can unpad after reduce-scatter. | ||
| """ | ||
| def pad_for_all_gather(x: torch.Tensor): |
Collaborator
There was a problem hiding this comment.
Why we need change the old logic in all_gather and reduce scatter, I think quant dis/com may not influence such old logic
zufayu
requested review from
ZhangLirong-amd
and removed request for
ZhangLirong-amd
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Motivation
Technical Details
Test Plan
Test Result
Submission Checklist