[Cherry-Pick][Optimization] Use triton qk_norm both in Prefill and Decode.(#7213)

[EmmonsCurse]

Cherry-pick of #7213 (authored by @K11OntheBoat) to release/2.5.

devPR:#7213

---

Motivation

Prefill 阶段使用QKRMSNorm融合算子. 部分模型 单Kernel部分加速2～7倍. Prefill 空泡较大的模型单次Forward可加速2倍左右.

Modifications

使用QKRMSNorm 替代paddle 散Op

Usage or Command

Accuracy Tests

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]

Thanks for your contribution!

[fastdeploy-bot]

🤖 AI Code Review | 2026-04-10 16:01 CST

📋 Review 摘要

PR 概述：在 Prefill 和 Decode 阶段都使用 Triton QK RMSNorm 融合算子，移除 CUDA Graph 模式限制

变更范围：fastdeploy/model_executor/layers/normalization.py、tests/e2e/test_Qwen3VL_serving.py

影响面 Tag：[OP]

PR 规范检查

✅ 标题包含有效 Tag [Cherry-Pick] 和 [Optimization]
✅ Motivation 和 Modifications 填写完整

问题

级别	文件	概述
🟡 建议	test_Qwen3VL_serving.py:176	测试预期输出变更未在 PR 描述中说明原因

总体评价

代码变更合理，移除 step_use_cudagraph 条件后 triton kernel 在 Prefill 阶段也能使用，符合优化目的。triton kernel 实现支持动态 batch size，且有独立的数值精度测试验证正确性。

[fastdeploy-bot]

🟡 建议 测试预期输出从 "视频中手机支架的颜色是黑色的。" 修改为 "视频中手机支架的颜色是黑色。"，这个变更应该在 PR 描述中说明原因。

建议在 PR 描述中补充说明：由于改用 triton qk_rmsnorm_fused kernel 替代 paddle 算子，导致模型输出在端到端测试中存在微小差异（属于正常范围内的浮点计算差异）。

[codecov-commenter]

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (release/2.5@1e0ab31). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/model_executor/layers/normalization.py	0.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@              Coverage Diff               @@
##             release/2.5    #7305   +/-   ##
==============================================
  Coverage               ?   69.93%           
==============================================
  Files                  ?      390           
  Lines                  ?    54417           
  Branches               ?     8581           
==============================================
  Hits                   ?    38056           
  Misses                 ?    13619           
  Partials               ?     2742           

Flag	Coverage Δ
GPU	69.93% <0.00%> (?)

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.