[OpenReg][distributed] Refactor OCCL backend registration#183257
Closed
KarhouTam wants to merge 2 commits into
Closed
[OpenReg][distributed] Refactor OCCL backend registration#183257KarhouTam wants to merge 2 commits into
KarhouTam wants to merge 2 commits into
Conversation
Separate distributed bindings from device module and remove unnecessary factory API. Address review comments from PR pytorch#171250: - Remove torch_python_library link from backend CMake (not needed) - Move ProcessGroupOCCL bindings to dedicated ProcessGroupInit.cpp - Remove _createProcessGroupOCCL factory, expose constructor directly - Clean up unnecessary imports
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/183257
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit dc9f630 with merge base f8b2b60 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
fffrog
reviewed
May 11, 2026
fffrog
reviewed
May 11, 2026
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 |
dsashidh
pushed a commit
to dsashidh/pytorch
that referenced
this pull request
May 13, 2026
…3257) ## Summary Refactors OCCL distributed backend registration to address code review feedback from PR pytorch#171250. ## Changes 1. **Remove unnecessary library dependency**: `torch_python_library` is no longer linked to the backend library - pybind11 bindings are handled by the separate `torch_bindings` library. 2. **Separate distributed bindings**: ProcessGroupOCCL bindings moved from inline in `Module.cpp` to dedicated `ProcessGroupInit.cpp/hpp` files. 3. **Remove factory function**: `_createProcessGroupOCCL` factory function removed - now exposing `ProcessGroupOCCL` constructor directly via pybind11. 4. **Clean up imports**: Removed unnecessary `<pybind11/chrono.h>` from Module.cpp and `<c10/util/intrusive_ptr.h>` from ProcessGroupInit.cpp. Pull Request resolved: pytorch#183257 Approved by: https://github.com/fffrog
Alokksinha00
pushed a commit
to Alokksinha00/pytorch
that referenced
this pull request
May 15, 2026
…3257) ## Summary Refactors OCCL distributed backend registration to address code review feedback from PR pytorch#171250. ## Changes 1. **Remove unnecessary library dependency**: `torch_python_library` is no longer linked to the backend library - pybind11 bindings are handled by the separate `torch_bindings` library. 2. **Separate distributed bindings**: ProcessGroupOCCL bindings moved from inline in `Module.cpp` to dedicated `ProcessGroupInit.cpp/hpp` files. 3. **Remove factory function**: `_createProcessGroupOCCL` factory function removed - now exposing `ProcessGroupOCCL` constructor directly via pybind11. 4. **Clean up imports**: Removed unnecessary `<pybind11/chrono.h>` from Module.cpp and `<c10/util/intrusive_ptr.h>` from ProcessGroupInit.cpp. Pull Request resolved: pytorch#183257 Approved by: https://github.com/fffrog
mansiag05
added a commit
to mansiag05/pytorch
that referenced
this pull request
May 17, 2026
The OCCL backend registration was refactored: the standalone createProcessGroupOCCL factory function was removed, pybind11 bindings moved from Module.cpp to a dedicated init.cpp, and Python registration now imports ProcessGroupOCCL directly. Update the doc to match.
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
4 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.
Summary
Refactors OCCL distributed backend registration to address code review feedback from PR #171250.
Changes
Remove unnecessary library dependency:
torch_python_libraryis no longer linked to the backend library - pybind11 bindings are handled by the separatetorch_bindingslibrary.Separate distributed bindings: ProcessGroupOCCL bindings moved from inline in
Module.cppto dedicatedProcessGroupInit.cpp/hppfiles.Remove factory function:
_createProcessGroupOCCLfactory function removed - now exposingProcessGroupOCCLconstructor directly via pybind11.Clean up imports: Removed unnecessary
<pybind11/chrono.h>from Module.cpp and<c10/util/intrusive_ptr.h>from ProcessGroupInit.cpp.cc @fffrog