introduce CudaGuard and cudastreamguard #14901
Merged
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
### Introduce CudaGuard and CudaStreamGuard This diff introduces `CudaGuard` and `CudaStreamGuard` in the Executorch CUDA runtime. These classes provide a convenient way to manage CUDA device and stream selection. #### Changes * Added `CudaGuard` and `CudaStreamGuard` classes in `fbcode/executorch/backends/cuda/runtime/guard.h`. * Implemented `CudaGuard` and `CudaStreamGuard` in `fbcode/executorch/backends/cuda/runtime/guard.cpp`. * Added unit tests for `CudaStreamGuard` in `fbcode/executorch/backends/cuda/runtime/tests/test_cuda_stream_guard.cpp`. * Updated `TARGETS` file to include the new files. #### Purpose The `CudaGuard` class provides a way to select a CUDA device and ensure that it is properly released when the guard goes out of scope. The `CudaStreamGuard` class provides a way to select a CUDA stream and ensure that it is properly synchronized when the guard goes out of scope. #### Usage They will be further used and controled by their shim layer functions. Differential Revision: [D84126481](https://our.internmc.facebook.com/intern/diff/D84126481/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/14901
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit a0077ff with merge base fb87fa6 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
This was referenced Oct 8, 2025
This PR needs a
|
larryliu0820
approved these changes
Oct 8, 2025
meta-codesync
bot
merged commit 7f63c1e
into
gh/gasoonjia/46/base
137 of 147 checks passed
Gasoonjia
added a commit
that referenced
this pull request
Oct 8, 2025
This PR was created by the merge bot to help merge the original PR into the main branch. ghstack PR number: #14901 by @Gasoonjia ^ 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/gasoonjia/46/base ghstack PR head: https://github.com/pytorch/executorch/tree/gh/gasoonjia/46/head Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/gasoonjia/45/orig Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/gasoonjia/46/orig Differential Revision: [D84126481](https://our.internmc.facebook.com/intern/diff/D84126481/) @diff-train-skip-merge --------- Co-authored-by: gasoonjia <gasoonjia@icloud.com>
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):
Introduce CudaGuard and CudaStreamGuard
This diff introduces
CudaGuardandCudaStreamGuardin the Executorch CUDA runtime. These classes provide a convenient way to manage CUDA device and stream selection.Changes
CudaGuardandCudaStreamGuardclasses infbcode/executorch/backends/cuda/runtime/guard.h.CudaGuardandCudaStreamGuardinfbcode/executorch/backends/cuda/runtime/guard.cpp.CudaStreamGuardinfbcode/executorch/backends/cuda/runtime/tests/test_cuda_stream_guard.cpp.TARGETSfile to include the new files.Purpose
The
CudaGuardclass provides a way to select a CUDA device and ensure that it is properly released when the guard goes out of scope. TheCudaStreamGuardclass provides a way to select a CUDA stream and ensure that it is properly synchronized when the guard goes out of scope.Usage
They will be further used and controled by their shim layer functions.
Differential Revision: D84126481