Skip to content

[Optimization] Optimization for gather_logprob by 10GB#5817

Merged
ckl117 merged 3 commits into
PaddlePaddle:developfrom
ckl117:dev_opt_logprob_mem
Dec 30, 2025
Merged

[Optimization] Optimization for gather_logprob by 10GB#5817
ckl117 merged 3 commits into
PaddlePaddle:developfrom
ckl117:dev_opt_logprob_mem

Conversation

@ckl117
Copy link
Copy Markdown
Collaborator

@ckl117 ckl117 commented Dec 29, 2025

Motivation

Reduced peak GPU memory usage of (x>=y).sum(-1) by 10GB.

💡 If this PR is a Cherry Pick, the PR title needs to follow the format by adding the [Cherry-Pick] label at the very beginning and appending the original PR ID at the end. For example, [Cherry-Pick][CI] Add check trigger and logic(#5191)

💡 如若此PR是Cherry Pick,PR标题需遵循格式,在最开始加上[Cherry-Pick]标签,以及最后面加上原PR ID,例如[Cherry-Pick][CI] Add check trigger and logic(#5191)

Modifications

Implement (x>=y).sum(-1) using Triton

Usage or Command

No

Accuracy Tests

add test_batched_count_greater_than.py

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.

@codecov-commenter
Copy link
Copy Markdown

codecov-commenter commented Dec 29, 2025

Codecov Report

❌ Patch coverage is 65.62500% with 11 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@ffb3ccf). Learn more about missing BASE report.

Files with missing lines Patch % Lines
...astdeploy/model_executor/layers/sample/logprobs.py 60.71% 11 Missing ⚠️
Additional details and impacted files
@@            Coverage Diff             @@
##             develop    #5817   +/-   ##
==========================================
  Coverage           ?   66.83%           
==========================================
  Files              ?      347           
  Lines              ?    44245           
  Branches           ?     6798           
==========================================
  Hits               ?    29571           
  Misses             ?    12493           
  Partials           ?     2181           
Flag Coverage Δ
GPU 66.83% <65.62%> (?)

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.

@ckl117 ckl117 merged commit 0bcf924 into PaddlePaddle:develop Dec 30, 2025
27 of 30 checks passed
ckl117 added a commit to ckl117/FastDeploy that referenced this pull request Dec 30, 2025
…5817)

* opt logprobs gather_logprob,reduce device memory usage by 10GB when token_num=8k
qingqing01 pushed a commit that referenced this pull request Dec 31, 2025
…5817)(#5846) (#5834)

* [Optimization] Optimization for gather_logprob by 10GB (#5817)

* opt logprobs gather_logprob,reduce device memory usage by 10GB when token_num=8k

* only cuda run triton op (#5846)
chang-wenbin pushed a commit to chang-wenbin/FastDeploy that referenced this pull request Mar 2, 2026
…5817)

* opt logprobs gather_logprob,reduce device memory usage by 10GB when token_num=8k
xiaoguoguo626807 pushed a commit to xiaoguoguo626807/FastDeploy that referenced this pull request May 7, 2026
…5817)

* opt logprobs gather_logprob,reduce device memory usage by 10GB when token_num=8k
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants