Skip to content

[RL]Perf: Optimize batch delete prefix and fused put in R3#6604

Open
gongshaotian wants to merge 4 commits intoPaddlePaddle:release/2.4from
gongshaotian:r3_opt_delete_2.4
Open

[RL]Perf: Optimize batch delete prefix and fused put in R3#6604
gongshaotian wants to merge 4 commits intoPaddlePaddle:release/2.4from
gongshaotian:r3_opt_delete_2.4

Conversation

@gongshaotian
Copy link
Collaborator

@gongshaotian gongshaotian commented Mar 2, 2026

Motivation

TODO:

  • Cherry-pick to develop, release/2.5
  1. The transmission performance under the Fused put mode has been optimized, and now bytes are used for transmission
  2. Optimize the prefix removal performance of P2PStore in the trajectory splitting scenario. Before optimization, the prefix removal overhead was in the hour-level when dealing with a million-level total number of keys

Modifications

  1. RoutingReplayManager: refine submit task code
    2.RDMAStore: opt the performance of delete_prefix_batch and put()

Usage or Command

Pass

Accuracy Tests

Pass

Checklist

  • Add at least a tag in the PR title.
    • Tag list: [[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]]
    • You can add new tags based on the PR content, but the semantics must be clear.
  • Format your code, run pre-commit before commit.
  • Add unit tests. Please write the reason in this PR if no unit tests.
  • Provide accuracy results.
  • If the current PR is submitting to the release branch, make sure the PR has been submitted to the develop branch, then cherry-pick it to the release branch with the [Cherry-Pick] PR tag.

@paddle-bot
Copy link

paddle-bot bot commented Mar 2, 2026

Thanks for your contribution!

@gongshaotian gongshaotian changed the title [RL]Optimize the performance of delete prefix batch and fused put [RL]Optimize the performance of R3 delete prefix batch and fused put Mar 2, 2026
@gongshaotian gongshaotian changed the title [RL]Optimize the performance of R3 delete prefix batch and fused put [RL]Optimize R3: batch delete prefix and fused put performance Mar 2, 2026
@gongshaotian gongshaotian changed the title [RL]Optimize R3: batch delete prefix and fused put performance [RL]Perf: Optimize batch delete prefix and fused put in R3 Mar 2, 2026
@codecov-commenter
Copy link

codecov-commenter commented Mar 2, 2026

Codecov Report

❌ Patch coverage is 22.72727% with 17 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (release/2.4@7d76fd4). Learn more about missing BASE report.

Files with missing lines Patch % Lines
...model_executor/layers/moe/routing_indices_cache.py 5.55% 17 Missing ⚠️
Additional details and impacted files
@@              Coverage Diff               @@
##             release/2.4    #6604   +/-   ##
==============================================
  Coverage               ?   56.30%           
==============================================
  Files                  ?      333           
  Lines                  ?    42378           
  Branches               ?     6451           
==============================================
  Hits                   ?    23861           
  Misses                 ?    16652           
  Partials               ?     1865           
Flag Coverage Δ
GPU 56.30% <22.72%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants