Skip to content

Conversation

@hubertlu-tw
Copy link
Collaborator

Motivation

Apply AITER_MXFP4_MOE_SF=1 only to gfx950 in aiter build which resolves the current aiter build error on MI300X.
https://github.com/sgl-project/sglang/actions/runs/19269898923/job/55095147727#step:5:854

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

CC: @sogalin

@gemini-code-assist
Copy link
Contributor

Summary of Changes

Hello @hubertlu-tw, 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 introduces targeted modifications to the Dockerfile responsible for building aiter on ROCm-enabled systems. The primary goal is to resolve a build error on MI300X GPUs by ensuring that the AITER_MXFP4_MOE_SF flag and corresponding environment variable are only activated when building for or running on the gfx950 architecture. This prevents unintended side effects on other GPU architectures while maintaining the intended performance optimizations for gfx950.

Highlights

  • Conditional Build Flag Application: The AITER_MXFP4_MOE_SF=1 build flag for aiter is now applied conditionally, specifically targeting the gfx950 GPU architecture during the build process.
  • Conditional Environment Variable Setting: The AITER_MXFP4_MOE_SF environment variable is now set to 1 at runtime only for gfx950 architectures, preventing its application to other GPU types.
  • MI300X Build Error Resolution: These changes are implemented to resolve a reported build error that occurred on MI300X systems, which was caused by the unconditional application of the AITER_MXFP4_MOE_SF flag.
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.

Copy link
Contributor

@gemini-code-assist gemini-code-assist bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request conditionally applies the AITER_MXFP4_MOE_SF=1 flag only for the gfx950 architecture during the aiter build in the ROCm Dockerfile. This is intended to fix a build error on MI300X. The changes correctly implement the conditional logic. My review includes suggestions to remove some dead code and to refactor the shell commands for better readability and maintainability.

@HaiShaw HaiShaw merged commit 36d1471 into sgl-project:main Nov 11, 2025
27 checks passed
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.

3 participants