fix _block_scaled_block_gemm kernel, remove config class#1239
fix _block_scaled_block_gemm kernel, remove config class#1239hiworldwzj merged 1 commit intomainfrom
Conversation
Summary of ChangesHello, 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 streamlines the Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe 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
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 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. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request refactors the kernel configuration logic by removing the Fp8BlockMMKernelConfig class and relying on the @autotune decorator for performance tuning. This is a good simplification that centralizes configuration management. However, I've identified a potential bug in the new fallback configuration where the block sizes for N and K dimensions appear to be swapped. This could lead to suboptimal performance or errors when non-square block sizes are used. Please see the specific comment for details.
| "BLOCK_N": block_size[0], | ||
| "BLOCK_K": block_size[1], |
There was a problem hiding this comment.
There seems to be a mix-up in assigning BLOCK_N and BLOCK_K from block_size. block_size[0] is block_k (for the K dimension) and block_size[1] is block_n (for the N dimension). However, BLOCK_N is being assigned block_size[0] and BLOCK_K is being assigned block_size[1]. This should be swapped to align the tiling dimensions correctly with the quantization block dimensions.
| "BLOCK_N": block_size[0], | |
| "BLOCK_K": block_size[1], | |
| "BLOCK_N": block_size[1], | |
| "BLOCK_K": block_size[0], |
1 participant
No description provided.