[ET-VK][ez] Update SDPA test to be able to test different SDPA modes #15598
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
added 2 commits
November 4, 2025 11:52
Title says it all! Use the new combos codegen API which makes it easier to express generating storage type combinations. Differential Revision: [D86226138](https://our.internmc.facebook.com/intern/diff/D86226138/) ghstack-source-id: 320850476 Pull Request resolved: #15576
Title says it all! The purpose of this diff is twofold: 1. Test SDPA as both a fused operator (sdpa_with_kv_cache) and decomposed update_cache and custom_sdpa ops in order to detect possible regressions with being able to support older models 2. Make it easier to debug issues with SDPA by exposing a mode that tests only the attention weight computation. Title says it all! Update SDPA op to use buffer storage for cache tensors if projected tensors are buffer. Also included is a small change to ensure that cache tensors use the same storage type as input tensors. Differential Revision: [D86226135](https://our.internmc.facebook.com/intern/diff/D86226135/) ghstack-source-id: 320850473 Pull Request resolved: #15577
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15598
Note: Links to docs will display an error until the docs builds have been completed. ⏳ No Failures, 31 PendingAs of commit cf49a75 with merge base bde6b11 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
SS-JIA
approved these changes
Nov 5, 2025
abhinaykukkadapu
pushed a commit
to abhinaykukkadapu/executorch
that referenced
this pull request
Nov 6, 2025
…ytorch#15598) Title says it all! The purpose of this diff is twofold: 1. Test SDPA as both a fused operator (sdpa_with_kv_cache) and decomposed update_cache and custom_sdpa ops in order to detect possible regressions with being able to support older models 2. Make it easier to debug issues with SDPA by exposing a mode that tests only the attention weight computation. Title says it all! Update SDPA op to use buffer storage for cache tensors if projected tensors are buffer. Also included is a small change to ensure that cache tensors use the same storage type as input tensors. Differential Revision: [D86226135](https://our.internmc.facebook.com/intern/diff/D86226135/)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This PR was created by the merge bot to help merge the original PR into the main branch.
ghstack PR number: #15577 by @SS-JIA
^ Please use this as the source of truth for the PR details, comments, and reviews
ghstack PR base: https://github.com/pytorch/executorch/tree/gh/SS-JIA/348/base
ghstack PR head: https://github.com/pytorch/executorch/tree/gh/SS-JIA/348/head
Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/SS-JIA/347/orig
Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/SS-JIA/348/orig
Differential Revision: D86226135
@diff-train-skip-merge