Move launch API from cudax to libcu++ #6667
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
davebayer
approved these changes
Nov 18, 2025
miscco
requested changes
Nov 18, 2025
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
davebayer
requested changes
Nov 19, 2025
| const cudax::device_buffer<int> b = cudax::make_buffer(stream, resource, a.size(), 1); | ||
|
|
||
| cudax::launch(stream, cudax::make_config(cuda::grid_dims<1>, cuda::block_dims<32>), add_kernel{}, a, b); | ||
| cudax::launch(stream, cuda::make_config(cuda::grid_dims<1>, cuda::block_dims<32>), add_kernel{}, a, b); |
Contributor
There was a problem hiding this comment.
Note: cuda::make_config is too generic in my opinion. We should find a better name in a separate PR
| __device__ void operator()(Config config) | ||
| { | ||
| decltype(auto) dynamic_smem = cudax::device::dynamic_shared_memory_view(config); | ||
| decltype(auto) dynamic_smem = cuda::dynamic_shared_memory_view(config); |
Contributor
There was a problem hiding this comment.
Suggested change
| decltype(auto) dynamic_smem = cuda::dynamic_shared_memory_view(config); | |
| decltype(auto) dynamic_smem = cuda::device::dynamic_shared_memory_view(config); |
This comment has been minimized.
This comment has been minimized.
alliepiper
reviewed
Dec 3, 2025
cudax/test/CMakeLists.txt
Outdated
| launch/launch_smoke.cu | ||
| ) | ||
|
|
||
| cudax_add_catch2_test(test_target launch_configuration ${cudax_target} |
Contributor
There was a problem hiding this comment.
Heads up that this will need to be rebased or have main merged in -- we no longer pass cudax_target here.
miscco
approved these changes
Dec 4, 2025
davebayer
reviewed
Dec 4, 2025
Comment on lines
+151
to
+152
| //! Launches a kernel functor object on the specified stream and with specified | ||
| //! configuration. Kernel functor object is a type with __device__ operator(). |
Contributor
There was a problem hiding this comment.
Please, fix your clang-format :)
This comment has been minimized.
This comment has been minimized.
davebayer
approved these changes
Dec 4, 2025
This comment has been minimized.
This comment has been minimized.
Contributor
🥳 CI Workflow Results🟩 Finished in 3h 04m: Pass: 100%/128 | Total: 21h 47m | Max: 1h 14m | Hits: 98%/219845See results here. |
github-actions bot
pushed a commit
that referenced
this pull request
Dec 4, 2025
* Move launch API from cudax to libcu++ * Review feedback and test fixes * More fixes * GCC7 fix * Update libcudacxx/include/cuda/__launch/configuration.h * Fix old GCC --------- Co-authored-by: David Bayer <[email protected]> (cherry picked from commit 4ab39a7)
Contributor
|
Successfully created backport PR for |
davebayer
pushed a commit
that referenced
this pull request
Dec 5, 2025
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.
This PR moves the parts of launch API ready for release to libcu++. Graphs related and
cuda::executionparts will remain experimental.I also commented out extended lambda testing to figure out cmake changes needed to add the compilation flag to that specific test