[Feature] Staggered batch scheduling for DP+EP#6359
Open
ApplEOFDiscord wants to merge 22 commits intoPaddlePaddle:developfrom
Open
[Feature] Staggered batch scheduling for DP+EP#6359ApplEOFDiscord wants to merge 22 commits intoPaddlePaddle:developfrom
ApplEOFDiscord wants to merge 22 commits intoPaddlePaddle:developfrom
Conversation
This reverts commit 0528baf.
|
Thanks for your contribution! |
|
gzz seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account. You have signed the CLA already but the status is still pending? Let us recheck it. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
In MoE inference system, requests arrive DP instances at different timepoints, which causes some DP instances make an idle forward, thereby reducing resource utilization and increasing TTFT
Modifications
To mitigate the above issue, we adopt a batching strategy on both scheduler side and engine side, keeping all DP instances as fully utilized as possible
Usage or Command
None
Accuracy Tests
None
Checklist
[FDConfig],[APIServer],[Engine],[Scheduler],[PD Disaggregation],[Executor],[Graph Optimization],[Speculative Decoding],[RL],[Models],[Quantization],[Loader],[OP],[KVCache],[DataProcessor],[BugFix],[Docs],[CI],[Optimization],[Feature],[Benchmark],[Others],[XPU],[HPU],[GCU],[DCU],[Iluvatar],[Metax]]pre-commitbefore commit.releasebranch, make sure the PR has been submitted to thedevelopbranch, then cherry-pick it to thereleasebranch with the[Cherry-Pick]PR tag.