vllm-project/vllm

[Feature]: Integrate RMS+fp4 fused kernel from FlashInfer

Open

#32,612 创建于 2026年1月19日

在 GitHub 查看
 (14 评论) (0 反应) (1 负责人)Python (80,034 star) (16,816 fork)batch import
feature requestgood first issuehelp wantedstaletorch.compile

描述

🚀 The feature, motivation and pitch

Kernel: https://github.com/flashinfer-ai/flashinfer/blob/main/flashinfer/norm.py#L406-L409

This should be integrated in the existing rms + quant fusion pass using a similar approach as the silu+fp4 fusion in the act quant fusion pass.

Alternatives

No response

Additional context

No response

Before submitting a new issue...

  • Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

贡献者指南

技术栈
pythoncpp
领域
backendperformance
议题类型
feature
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
4
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
over 1 week
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
blocked
清晰度议题是否清楚说明期望改动、验收标准和下一步。
clear
前置要求
PythonC++/CUDAFused kernel conceptsQuantization
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
15
研究方向
Study the existing RMS+quant fusion pass in the vllm codebase and the analogous silu+fp4 fusion in the act quant fusion pass. Review the FlashInfer RMS+fp4 kernel implementation at flashinfer/norm.py#L406 L409. Identify integration points and potential modifications. Engage with the assignee in the issue comments to clarify scope and avoid redundant work.
[Feature]: Integrate RMS+fp4 fused kernel from FlashInfer · vllm-project/vllm#32612 | Good First Issue