Revert "[Feature] Integrate new deepgemm (#19820)"

[yewentao256]

This reverts commit c6e3bba.

Purpose

#19820

This pr should be merged after deepseek-ai/DeepGEMM#112

I have asked the author to merge quickly, but if they don't, we should temporily revert this pr.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[gemini-code-assist]

Summary of Changes

Hello @yewentao256, 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 is titled as a revert of a previous deepgemm integration, citing an unmerged upstream dependency. However, upon inspection of the changes, it appears to be an update and refinement of the existing deepgemm integration rather than a full undo. The changes primarily involve adapting to new deepgemm API conventions and reorganizing related utility functions within the codebase, particularly within benchmark and test files.

Highlights

• Stated Intent vs. Actual Changes: While this pull request is titled as a revert of a previous deepgemm integration, the actual changes implemented suggest an update and refinement of the existing deepgemm integration, adapting to new API conventions and internal organization, rather than undoing the feature.
• DeepGEMM API Migration: Core deepgemm function calls (e.g., fp8_gemm_nt, fp8_m_grouped_gemm_nt_masked, m_grouped_fp8_gemm_nt_contiguous) have been updated to their new gemm_fp8_fp8_bf16_nt and m_grouped_gemm_fp8_fp8_bf16_nt_contiguous counterparts across various files, reflecting changes in the deepgemm library's API.
• FP8 Utility Refactoring: Common FP8 quantization utility functions like per_block_cast_to_fp8, per_token_cast_to_fp8, ceil_div, and get_col_major_tma_aligned_tensor have been relocated, re-implemented locally within test/benchmark files, or are now directly imported from the deep_gemm module, streamlining their usage and reducing redundancy.
• Benchmark Script Improvements: The benchmark_fp8_block_dense_gemm.py script has been enhanced with expanded benchmark shapes, improved formatting, and additional comments, making the performance evaluation more comprehensive and readable.

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 is currently in preview and 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 to provide feedback.

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. ↩

[WoosukKwon]

Oh sorry I didn't know that the PR was dependent on the upgrade.

[gemini-code-assist]

Code Review

This PR reverts the integration of the deepgemm feature. The changes appear correct, and I've suggested removing some commented-out code.

[yewentao256]

Oh sorry I didn't know that the PR was dependent on the upgrade.

@WoosukKwon No worries, actually it is my fault, I should mark it for not merged.

Let's see if DeepGemm will merge the pr quickly, if they don't and someone met the interface error in vllm, we can directly merge this pr to fix

[yewentao256]

@WoosukKwon They said they will merge the pr after 2~3 weeks, let's merge this pr first, thanks!