Skip to content

[CI] Pin final Paddle version due to VMM V2 incompatibility - #8108

Merged
EmmonsCurse merged 1 commit into
PaddlePaddle:developfrom
EmmonsCurse:freeze_paddle_version
Aug 10, 2026
Merged

[CI] Pin final Paddle version due to VMM V2 incompatibility#8108
EmmonsCurse merged 1 commit into
PaddlePaddle:developfrom
EmmonsCurse:freeze_paddle_version

Conversation

@EmmonsCurse

Copy link
Copy Markdown
Collaborator

Motivation

FastDeploy can no longer safely upgrade to newer Paddle versions.

After PaddlePaddle/Paddle#79605, the VMM V2 allocator is enabled by default. However, FastDeploy's KV Cache IPC mechanism still relies on cudaIpcGetMemHandle, which is incompatible with VMM-allocated memory.

As a result, upgrading to Paddle versions with VMM V2 enabled by default can cause FastDeploy runtime failures during KV Cache initialization.

To maintain FastDeploy runtime stability, the currently compatible Paddle version is pinned as the final supported version. FastDeploy CI will no longer follow subsequent Paddle version upgrades unless the KV Cache IPC mechanism is adapted to support the new allocator.

Modifications

  • Pinned the Paddle version used by FastDeploy CI to the current compatible version.
  • Prevented CI from automatically upgrading to newer Paddle versions.
  • Preserved compatibility with FastDeploy's KV Cache IPC mechanism based on cudaIpcGetMemHandle.
  • Pinned the current Paddle version as the final supported version for FastDeploy.
  • Stopped tracking subsequent Paddle version upgrades until the KV Cache IPC mechanism is adapted to support VMM-allocated memory.

Usage or Command

N/A

Accuracy Tests

N/A

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.

@EmmonsCurse

Copy link
Copy Markdown
Collaborator Author

/skip-ci ci_iluvatar
/skip-ci ci_hpu
/skip-ci build_xpu

@codecov-commenter

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (develop@b77324d). Learn more about missing BASE report.

Additional details and impacted files
@@            Coverage Diff             @@
##             develop    #8108   +/-   ##
==========================================
  Coverage           ?   77.46%           
==========================================
  Files              ?      408           
  Lines              ?    58186           
  Branches           ?     9134           
==========================================
  Hits               ?    45071           
  Misses             ?    10231           
  Partials           ?     2884           
Flag Coverage Δ
GPU 77.46% <ø> (?)

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

☔ View full report in Codecov by Harness.
📢 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.

@EmmonsCurse
EmmonsCurse merged commit ea003c0 into PaddlePaddle:develop Aug 10, 2026
42 checks passed
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