[CP] Introduce flex_cp_forward custom op for FlexAttention CP #163185
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/163185
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ✅ No FailuresAs of commit db5a09d with merge base d41aa18 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
pytorch-bot
bot
added
ciflow/inductor
oncall: distributed
This was referenced Sep 17, 2025
fegin
added a commit
that referenced
this pull request
Sep 17, 2025
The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd. For the next step, we should explore how to interpolate the required communication based on the information from BlockMask. ghstack-source-id: a2547aa Pull-Request-resolved: #163185
This was referenced Sep 17, 2025
fegin
added
topic: not user facing
fegin
added a commit
that referenced
this pull request
Sep 18, 2025
The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd. For the next step, we should explore how to interpolate the required communication based on the information from BlockMask. ghstack-source-id: 26feaeb Pull-Request-resolved: #163185
fegin
added a commit
that referenced
this pull request
Sep 18, 2025
The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd. For the next step, we should explore how to interpolate the required communication based on the information from BlockMask. ghstack-source-id: d5259e1 Pull-Request-resolved: #163185
fegin
added a commit
that referenced
this pull request
Sep 29, 2025
The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd. For the next step, we should explore how to interpolate the required communication based on the information from BlockMask. ghstack-source-id: 453b2e8 Pull-Request-resolved: #163185
fegin
added a commit
that referenced
this pull request
Oct 2, 2025
The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd. For the next step, we should explore how to interpolate the required communication based on the information from BlockMask. ghstack-source-id: 0d3a95e Pull-Request-resolved: #163185
fegin
added a commit
that referenced
this pull request
Oct 3, 2025
The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd. For the next step, we should explore how to interpolate the required communication based on the information from BlockMask. ghstack-source-id: 9b8c540 Pull-Request-resolved: #163185
fegin
added a commit
that referenced
this pull request
Oct 9, 2025
The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd. For the next step, we should explore how to interpolate the required communication based on the information from BlockMask. ghstack-source-id: 4b720b1 Pull-Request-resolved: #163185
Contributor
Author
|
@pytorchbot merge |
Collaborator
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Collaborator
|
The merge job was canceled or timed out. This most often happen if two merge requests were issued for the same PR, or if merge job was waiting for more than 6 hours for tests to finish. In later case, please do not hesitate to reissue the merge command |
Collaborator
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Collaborator
Merge failedReason: Comment with id 3396438666 not found Details for Dev Infra teamRaised by workflow job |
Collaborator
|
@pytorchbot merge |
Collaborator
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
zhudada0120
pushed a commit
to zhudada0120/pytorch
that referenced
this pull request
Oct 15, 2025
…orch#165039) No logic change, just polish the docstrings, comments and remove unused variables Pull Request resolved: pytorch#165039 Approved by: https://github.com/XilunWu ghstack dependencies: pytorch#162542, pytorch#164500, pytorch#163185
Chao1Han
pushed a commit
to Chao1Han/pytorch
that referenced
this pull request
Oct 21, 2025
…h#163185) The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd. For the next step, we should explore how to interpolate the required communication based on the information from BlockMask. Pull Request resolved: pytorch#163185 Approved by: https://github.com/XilunWu ghstack dependencies: pytorch#162542, pytorch#164500
Chao1Han
pushed a commit
to Chao1Han/pytorch
that referenced
this pull request
Oct 21, 2025
…orch#165039) No logic change, just polish the docstrings, comments and remove unused variables Pull Request resolved: pytorch#165039 Approved by: https://github.com/XilunWu ghstack dependencies: pytorch#162542, pytorch#164500, pytorch#163185
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.
Stack from ghstack (oldest at bottom):
The custom op will fetch the required K and V. Currently, the forward pass is just an all-gather, and the backward pass is a reduce-scatter. While the logic is the same as all_gather_tensor_autograd, the custom op avoids the Autograd warning that wait_tensor() is registered to autograd.
For the next step, we should explore how to interpolate the required communication based on the information from BlockMask.
cc @H-Huang @awgu @wanchaol @fduwjj @wz337 @wconstab @d4l3k @pragupta @ezyang @msaroufim @dcci