[WIP,Cute,Flex,Sm100] vectorized mask mod application#2261
Draft
reubenconducts wants to merge 1 commit intoDao-AILab:mainfrom
Draft
[WIP,Cute,Flex,Sm100] vectorized mask mod application#2261reubenconducts wants to merge 1 commit intoDao-AILab:mainfrom
reubenconducts wants to merge 1 commit intoDao-AILab:mainfrom
Conversation
drisspg
reviewed
Feb 28, 2026
| # 2: application, where it is applied to compile down to r2p | ||
| # | ||
| # evaluation | ||
| num_mask_vals = (ncol + 32 - 1) // 32 |
Collaborator
There was a problem hiding this comment.
I think we need R2P width global constant, this is where the 32 comes from right?
Contributor
Author
There was a problem hiding this comment.
the 32 does come from the R2P width yes
Contributor
Author
There was a problem hiding this comment.
actually sorry that's not the case - it was just my choice to keep the bitmask in Uint32s
drisspg
reviewed
Feb 28, 2026
Collaborator
drisspg
left a comment
drisspg
left a comment
There was a problem hiding this comment.
couple clarifying questions but this looks good, I just put up autotuning PR: pytorch/pytorch#176055
helps alot in some cases
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Follow-up to #2236. The approach to vectorizing is bipartite:
r2pThe latter is important for example in situations where
mask_moddepends onaux_tensorsthat are contiguous in the kv idx, or whenaux_tensorsdon't depend on kv index at all.mask_mods still emitTensorSSAs, but they need not be single values. These are treated as bit-packed masks.cc @drisspg