[CI] Fix b200 flashinfer installation

[ShangmingCai]

Motivation

A supplement PR of #11902.
This can fix https://github.com/sgl-project/sglang/actions/runs/18681539949/job/53280297485:

+ pip install flashinfer-python==0.4.0 --prerelease=allow --break-system-packages

Usage:   
  pip install [options] <requirement specifier> [package-index-options] ...
  pip install [options] -r <requirements file> [package-index-options] ...
  pip install [options] [-e] <vcs project url> ...
  pip install [options] [-e] <local project path> ...
  pip install [options] <archive url/path> ...

no such option: --prerelease

TODO: restore the original logic when flashinfer-ai/flashinfer#1962 and flashinfer-ai/flashinfer#1960 are addressed.

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.

[gemini-code-assist]

Summary of Changes

Hello @ShangmingCai, 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 and fixes issues within the B200 CI environment by refining the dependency installation script. The changes ensure that the correct versions of the main project package and its associated dependencies are installed based on whether the CI is running on a Blackwell system or a standard environment, thereby improving the reliability and consistency of the continuous integration process.

Highlights

• Blackwell Environment Installation: For Blackwell (B200) CI environments, the main python[dev] package is now explicitly installed using a PyTorch extra index URL and with a forced reinstallation flag, ensuring the correct version and dependencies are in place.
• Conditional Installation Logic Refinement: The dependency installation script has been refactored to clearly separate installation steps for Blackwell and non-Blackwell environments. Specific installations, such as flashinfer-python==0.4.0 and the main package with --no-deps and --upgrade, are now exclusively applied to non-Blackwell setups.

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 correctly fixes the CI environment for Blackwell (b200) GPUs by separating the package installation logic from the standard environment. The previous implementation incorrectly applied installation steps meant for the normal CI, which uses uv, to the Blackwell environment, which uses pip. The changes resolve this by moving the uv-specific steps into an else block and adding a dedicated pip installation command for the Blackwell case. The fix appears sound. I have added one suggestion to refactor the installation script for the non-Blackwell case to improve clarity and reduce redundancy.