Skip to content
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.

Sign up for GitHub

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

Jump to bottom

[Neuron][Kernel] NKI-based flash-attention kernel with paged KV cache #11277

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[Neuron][Kernel] NKI-based flash-attention kernel with paged KV cache #11277

wants to merge 1 commit into from
+1,126 −1

Conversation

liangfu
Copy link
Contributor

@liangfu liangfu commented Dec 18, 2024 •
edited by github-actions bot
Loading

Summary

FIX #11152

This PR introduce a NKI-based kernel that brings the support for chunked-prefill with flash-attention.

Co-authored-by: Jiangfei Duan [email protected]

@github-actions GitHub Actions
Copy link

👋 Hi! Thank you for contributing to the vLLM project.
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 do one of these:

  • Add ready label to the PR
  • Enable auto-merge.

🚀

@liangfu liangfu force-pushed the nki-flash-attn branch 3 times, most recently from 188fc99 to b63906b Compare December 19, 2024 06:50
@mergify mergify bot added the ci/build label Jan 6, 2025
@mergify Mergify
Copy link

mergify bot commented Jan 7, 2025

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @liangfu.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

@mergify mergify bot added the needs-rebase label Jan 7, 2025
@liangfu liangfu marked this pull request as ready for review January 7, 2025 06:39
@mergify mergify bot removed the needs-rebase label Jan 7, 2025
@liangfu liangfu changed the title [Draft][Neuron][Kernel] NKI-based flash-attention kernel with paged KV cache [Neuron][Kernel] NKI-based flash-attention kernel with paged KV cache Jan 7, 2025
@liangfu liangfu mentioned this pull request Jan 7, 2025
Open
4 tasks
@liangfu @JF-D
NKI-based flash-attention kernel with paged KV cache
c2af356 
Co-authored-by: Jiangfei Duan <[email protected]>
Signed-off-by: Liangfu Chen <[email protected]>
@simon-mo simon-mo added the ready ONLY add when PR is ready to merge/full CI is needed label Jan 9, 2025
@simon-mo
Copy link
Collaborator

simon-mo commented Jan 9, 2025

@robertgshaw2-neuralmagic @WoosukKwon PTAL

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@JF-D JF-D JF-D left review comments

At least 1 approving review is required to merge this pull request.

Assignees
No one assigned
Labels
ci/build ready ONLY add when PR is ready to merge/full CI is needed
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

[RFC][Exploratory]: vLLM Neuron Backend with V1 Architecture
3 participants
@liangfu @simon-mo @JF-D