[draft](eplb): add per-layer expert-load statistics monitor for EP path#1210
Draft
JiaoliangYu wants to merge 1 commit into
Draft
[draft](eplb): add per-layer expert-load statistics monitor for EP path#1210JiaoliangYu wants to merge 1 commit into
JiaoliangYu wants to merge 1 commit into
Conversation
Collect per-layer, per-expert token counts from the MORI EP dispatch output (dispatch_recv_token_num) into a windowed ExpertLoadMonitor. Logs avg/max/balancedness and can emit a one-shot offline rebalance plan (hot/cold experts) via the offline_eplb_rebalance utility command. Scope is statistics only: per-rank, no cross-rank all-reduce, and no actual expert weight remap/transfer yet. All gated behind ATOM_ENABLE_EPLB_LOAD_STATS (default off). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Collect per-layer, per-expert token counts from the MORI EP dispatch output (dispatch_recv_token_num) into a windowed ExpertLoadMonitor. Logs avg/max/balancedness and can emit a one-shot offline rebalance plan (hot/cold experts) via the offline_eplb_rebalance utility command.
Scope is statistics only: per-rank, no cross-rank all-reduce, and no actual expert weight remap/transfer yet. All gated behind ATOM_ENABLE_EPLB_LOAD_STATS (default off).
Motivation
Technical Details
Test Plan
Test Result
Submission Checklist