refactor: unify autotuner for bmm_fp8

[ttyio]

add "auto" backend and autotune tactics among cutlass, cudnn and cublas.

📌 Description

🔍 Related Issues

🚀 Pull Request Checklist

Thank you for contributing to FlashInfer! Before we review your pull request, please make sure the following items are complete.

✅ Pre-commit Checks

• I have installed pre-commit by running pip install pre-commit (or used your preferred method).
• I have installed the hooks with pre-commit install.
• I have run the hooks manually with pre-commit run --all-files and fixed any reported issues.

If you are unsure about how to set up pre-commit, see the pre-commit documentation.

🧪 Tests

• Tests have been added or updated as needed.
• All tests are passing (unittest, etc.).

Reviewer Notes

[gemini-code-assist]

Summary of Changes

Hello @ttyio, 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 a unified autotuning mechanism for bmm_fp8 (batch matrix multiplication with FP8 precision) operations. It adds an "auto" backend option that intelligently selects the most performant implementation among cuBLAS, cuDNN, and CUTLASS, thereby optimizing performance for various hardware and input configurations.

Highlights

• New "auto" backend for bmm_fp8: The bmm_fp8 function now supports an "auto" backend option, which automatically selects the most performant GEMM (General Matrix Multiply) kernel among available implementations (cuBLAS, cuDNN, CUTLASS) for FP8 batch matrix multiplication.
• Refactored FP8 GEMM implementations: The cuBLAS and CUTLASS FP8 GEMM implementations have been refactored into TunableRunner classes, making them compatible with the autotuning framework for dynamic performance optimization.
• Added cuDNN FP8 GEMM TunableRunner: A new TunableRunner has been introduced for cuDNN FP8 GEMM operations, expanding the range of backends available for autotuning.
• Unified FP8 GEMM dispatch logic: The dispatch logic for bmm_fp8 has been consolidated into a single autotuning entry point (fp8_gemm_sm100), streamlining the process of selecting the optimal backend based on runtime characteristics.
• Updated test suite for "auto" backend: The test suite for bmm_fp8 has been updated to include test cases for the new "auto" backend, ensuring its correctness and performance across various configurations.

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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

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 issue 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 introduces an "auto" backend for bmm_fp8 to enable autotuning across cutlass, cudnn, and cublas backends, which is a great enhancement. The code is well-structured and the new autotuning logic is centralized. My feedback focuses on improving code clarity and maintainability by adding missing type hints and simplifying some expressions. I've also suggested a change to improve robustness by handling invalid inputs defensively. Finally, there's a critical fix for the tests to ensure the new autotuning capabilities are properly verified.

[nvpohanh]

@nvjullin please comment if this works. thanks!

[yongwww]

overall looks good to me

[ttyio]

@yongwww can I get this merged? thank you!

[yzh119]

LGTM