Add document for vllm paged attention kernel. #2978
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Collaborator
|
For reviewers, rendered here: https://vllm--2978.org.readthedocs.build/en/2978/dev/kernel/paged_attention.html |
Collaborator
|
Thank you for this great write up! |
Contributor
|
This doc is very usefull. Hope to be merged soon. |
LiuXiaoxuanPKU
approved these changes
Mar 4, 2024
Collaborator
LiuXiaoxuanPKU
left a comment
LiuXiaoxuanPKU
left a comment
There was a problem hiding this comment.
Looks great! Thanks! Some minor comments.
Collaborator
|
@pian13131 This is AWESOME! Thanks for your contribution! |
dtransposed
pushed a commit
to afeldman-nm/vllm
that referenced
this pull request
Mar 26, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hello, I am currently studying the vLLM paged attention kernel, and I've found that the implementation can be quite complex for newcomers. After thoroughly reviewing the primary implementation of the kernel in
csrc/attention/attention_kernels.cu, I have composed this document to provide a high-level understanding of the paged attention kernel. The document covers explanations on memory layout, read patterns, and step-by-step calculations, accompanied by diagrams and pseudo-code. It is intended to serve as a valuable reference for individuals interested in the implementation of the paged attention kernel.Given that I am still a novice in this subject, there may be some misunderstandings in the document. I welcome any comments and advice to enhance its accuracy and clarity. Your feedback is highly appreciated! Wish this document can be merged to help other people!