fix RuntimeError: RMSNorm failed with error code an illegal memory access was encountered

[gongwei-130]

Motivation

issue: #14120

91e8dc3#diff-d7e3ce3bc8e85e0a99fbcde0049eedd948dcf84167802cf3342cc02d0beeab1aL1121

After fix, the test/srt/test_deepseek_v3_fp4_4gpu.py test could pass.

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
• Follow the SGLang code style guidance.
• Work with maintainers to merge your PR. See the PR Merge Process

[gemini-code-assist]

Summary of Changes

Hello @gongwei-130, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request addresses a critical RuntimeError that occurred during RMSNorm calculations, manifesting as an illegal memory access. The fix involves a targeted data type conversion, explicitly casting router_logits to torch.float32 within the fused_moe_triton layer. This change is essential for maintaining the stability and correctness of the model's operations, particularly for specific configurations like deepseek_v3_fp4, and resolves the underlying cause of the runtime exception.

Highlights

• Error Resolution: Fixed a RuntimeError related to RMSNorm operations, specifically an 'illegal memory access was encountered' error.
• Data Type Correction: Introduced an explicit cast of router_logits to torch.float32 to ensure correct data type handling and prevent runtime issues.
• Test Validation: Ensured that the test_deepseek_v3_fp4_4gpu.py test now passes successfully after applying the fix.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request addresses a RuntimeError related to an illegal memory access. The fix involves casting router_logits to torch.float32 before it's passed to the trtllm_fp4_block_scale_moe kernel. This change likely resolves a data type mismatch issue within the kernel that was manifesting as a memory access error. The fix is straightforward and appears correct. Good job on identifying and resolving this issue.

[Kangyan-Zhou]

@gongwei-130 thanks for the fix!