[cudax] Make lane_mask part of the mapping result#9140
Merged
davebayer merged 1 commit intoJun 1, 2026
Conversation
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (13)
💤 Files with no reviewable changes (1)
🚧 Files skipped from review as they are similar to previous changes (9)
📝 WalkthroughSummary by CodeRabbit
important: WalkthroughThis PR threads a ::cuda::device::lane_mask through group mapping results: the concept requires lane_mask(), __mapping_result stores and exposes a lane mask, mapping algorithms compute updated masks, synchronizers read masks at sync time, and tests validate mask propagation. ChangesLane Mask Support in Group Mappings
Suggested labels
Suggested reviewers
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 3
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 359ff862-0f7a-40b2-8f54-2077d3eadc7b
📒 Files selected for processing (13)
cudax/include/cuda/experimental/__group/concepts.cuhcudax/include/cuda/experimental/__group/group.cuhcudax/include/cuda/experimental/__group/mapping/group_as.cuhcudax/include/cuda/experimental/__group/mapping/group_by.cuhcudax/include/cuda/experimental/__group/mapping/mapping_result.cuhcudax/include/cuda/experimental/__group/synchronizer/lane_synchronizer.cuhcudax/include/cuda/experimental/__group/this_group.cuhcudax/test/common/group_testing.cuhcudax/test/group/mapping/composite_mapping.cucudax/test/group/mapping/group_as.cucudax/test/group/mapping/group_by.cucudax/test/group/mapping/identity_mapping.cucudax/test/group/synchronizer/lane_synchronizer.cu
💤 Files with no reviewable changes (1)
- cudax/test/group/synchronizer/lane_synchronizer.cu
This comment has been minimized.
This comment has been minimized.
davebayer
changed the title
lane_mask be part of the mapping resultlane_mask part of the mapping result
Contributor
There was a problem hiding this comment.
🧹 Nitpick comments (2)
cudax/include/cuda/experimental/__group/synchronizer/lane_synchronizer.cuh (2)
28-28: 💤 Low valuesuggestion: Drop the unused
<cuda/std/__fwd/span.h>include (only occurrence at line 28) and consider removing__is_supported_count(defined at lines 44-47, with no call sites in this file).
33-33: 💤 Low valuesuggestion: Drop the
#include <cuda/experimental/__group/mapping/group_by.cuh>inlane_synchronizer.cuh(line 33) if it isn’t needed for anything besides_MappingResult’s__group_mapping_resultconcept check.group_bydoesn’t appear elsewhere in the header, and__group_mapping_resultis defined in__group/concepts.cuhwithout referencinggroup_by.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: b9b3ceef-4108-4562-a135-604dc84ff881
📒 Files selected for processing (2)
cudax/include/cuda/experimental/__group/synchronizer/lane_synchronizer.cuhcudax/test/group/mapping/group_by.cu
🚧 Files skipped from review as they are similar to previous changes (1)
- cudax/test/group/mapping/group_by.cu
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
davebayer
force-pushed
the
groups_lane_mask_inside_mapping_result
branch
from
f4da133 to
3408197
Compare
Contributor
🥳 CI Workflow Results🟩 Finished in 33m 02s: Pass: 100%/55 | Total: 16h 15m | Max: 32m 58s | Hits: 31%/98019See results here. |
miscco
approved these changes
Jun 1, 2026
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.
Storing lane mask only inside
lane_synchronizerinstance had several problems:barrier_synchronizer, there was no way to create a sub-group that would uselane_synchronizer, because we couldn't retrieve the lane mask.lane_synchronzierinstance, which made us duplicate the operations.I solved both problems by moving the lane mask into the mapping result. Now, just from the mapping result, the threads are aware which threads from the same group are inside the same warp. I believe it will be useful for some other collective operations even when using
barrier_synchronizer