{{ message }}
[ExecuTorch][WebGPU] Graph-owned scratch buffers for fused-op intermediates#20073
Merged
meta-codesync[bot] merged 8 commits intoJun 9, 2026
Merged
Conversation
meta-cla
Bot
added
the
CLA Signed
JulianCloudNTH
added a commit
that referenced
this pull request
Jun 5, 2026
…diates Pull Request resolved: #20073 Add `WebGPUGraph::create_scratch_buffer` for fused-op intermediates (SDPA's `attn_weights`/`attn_weights_softmax`) that are not model tensors and live only between dispatches. Graph-owned, released in the destructor. Vulkan models these as graph tensors; we use raw buffers (buffer-only backend). Consumed by the fused SDPA op above. ghstack-source-id: 390416621 @exported-using-ghexport Differential Revision: [D107543259](https://our.internmc.facebook.com/intern/diff/D107543259/)
JulianCloudNTH
added a commit
that referenced
this pull request
Jun 5, 2026
…diates Pull Request resolved: #20073 Add `WebGPUGraph::create_scratch_buffer` for fused-op intermediates (SDPA's `attn_weights`/`attn_weights_softmax`) that are not model tensors and live only between dispatches. Graph-owned, released in the destructor. Vulkan models these as graph tensors; we use raw buffers (buffer-only backend). Consumed by the fused SDPA op above. ghstack-source-id: 390416621 @exported-using-ghexport Differential Revision: [D107543259](https://our.internmc.facebook.com/intern/diff/D107543259/)
JulianCloudNTH
added a commit
that referenced
this pull request
Jun 5, 2026
…diates Pull Request resolved: #20073 Add `WebGPUGraph::create_scratch_buffer` for fused-op intermediates (SDPA's `attn_weights`/`attn_weights_softmax`) that are not model tensors and live only between dispatches. Graph-owned, released in the destructor. Vulkan models these as graph tensors; we use raw buffers (buffer-only backend). Consumed by the fused SDPA op above. ghstack-source-id: 390416621 @exported-using-ghexport Differential Revision: [D107543259](https://our.internmc.facebook.com/intern/diff/D107543259/)
digantdesai
approved these changes
Jun 5, 2026
digantdesai
left a comment
digantdesai
left a comment
Contributor
There was a problem hiding this comment.
Review automatically exported from Phabricator review in Meta.
JulianCloudNTH
added a commit
that referenced
this pull request
Jun 5, 2026
…diates Pull Request resolved: #20073 Add `WebGPUGraph::create_scratch_buffer` for fused-op intermediates (SDPA's `attn_weights`/`attn_weights_softmax`) that are not model tensors and live only between dispatches. Graph-owned, released in the destructor. Vulkan models these as graph tensors; we use raw buffers (buffer-only backend). Consumed by the fused SDPA op above. ghstack-source-id: 390440808 @exported-using-ghexport Differential Revision: [D107543259](https://our.internmc.facebook.com/intern/diff/D107543259/)
This was referenced Jun 5, 2026
Merged
This was referenced Jun 6, 2026
JulianCloudNTH
added a commit
that referenced
this pull request
Jun 9, 2026
…diates Pull Request resolved: #20073 Add `WebGPUGraph::create_scratch_buffer` for fused-op intermediates (SDPA's `attn_weights`/`attn_weights_softmax`) that are not model tensors and live only between dispatches. Graph-owned, released in the destructor. Vulkan models these as graph tensors; we use raw buffers (buffer-only backend). Consumed by the fused SDPA op above. ghstack-source-id: 391378805 @exported-using-ghexport Differential Revision: [D107543259](https://our.internmc.facebook.com/intern/diff/D107543259/)
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.
Stack from ghstack (oldest at bottom):
Add
WebGPUGraph::create_scratch_bufferfor fused-op intermediates (SDPA'sattn_weights/attn_weights_softmax) that are not model tensors and live only between dispatches. Graph-owned, released in the destructor. Vulkan models these as graph tensors; we use raw buffers (buffer-only backend). Consumed by the fused SDPA op above.@exported-using-ghexport
Differential Revision: D107543259
Differential Revision: D107543259