-
Notifications
You must be signed in to change notification settings - Fork 2.8k
Remove 200us slow concat kernel (part 1: kernel) #7145
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Summary of Changes
Hello @fzyzcjy, 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 the first part of an effort to remove a performance bottleneck identified as a slow concatenation operation within the CUTLASS MLA kernel. It refactors the kernel and its surrounding code to accept the query tensor split into its non-positional (q_nope
) and positional (q_pe
) components upfront, eliminating the need for concatenation inside the kernel.
Highlights
- Kernel Signature Change: The
cutlass_mla_decode
kernel and its interfaces (Python wrapper, C++ binding, header) have been updated to accept two separate query tensors (q_nope
andq_pe
) instead of a single concatenatedq_nope_and_q_pe
tensor. - Benchmark and Test Updates: The benchmark script (
bench_cutlass_mla.py
) and the test script (test_cutlass_mla.py
) have been modified to generate and pass the split query tensors (q_nope
,q_pe
) to the updatedcutlass_mla_decode
function. - CUDA Version Check Removed: The conditional compilation block checking for CUDA version >= 12.4 in
cutlass_mla_kernel.cu
has been commented out, effectively making the kernel always available.
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
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configureGemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
831fc43
to
2f4ec75
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Code Review
The pull request refactors the cutlass_mla_decode
function to accept separate q_nope
and q_pe
tensors, enhancing code modularity and potentially improving performance. The changes are consistently applied across the codebase, including benchmarks and tests. Key areas for review include ensuring type consistency, optimizing tensor creation, and clarifying device initialization.
Motivation
Modifications
Checklist