Change tensor alignment method to mn major

[mmangkad]

Modifications

Change tensor alignment method to mn major (from changes on #9340).

[gemini-code-assist]

Summary of Changes

Hello @mmangkad, 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 updates the internal tensor alignment strategy used in specific deep learning operations. The change ensures that tensors are aligned using the MN-major convention, which can be crucial for optimizing memory access patterns and computational efficiency in certain hardware architectures or deep learning frameworks.

Highlights

• Tensor Alignment Update: Updated the tensor alignment method within the forward_deepgemm function from column-major to MN-major for TMA-aligned tensors, specifically for gateup_input_scale and down_input_scale. This change is a follow-up to modifications introduced in PR #9340.

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 updates the tensor alignment method from get_col_major_tma_aligned_tensor to get_mn_major_tma_aligned_tensor in EPMoE.forward_deepgemm. This change aligns the implementation with DeepEPMoE.forward_deepgemm_masked and likely reflects an update in the underlying deep_gemm library for handling TMA-aligned tensors. The changes are straightforward and look correct.

[zhaotyer]

Modifications

Change tensor alignment method to mn major (from changes on #9340).

It's work on H100

[zhyncs]

Thanks for this quick fix!