Feature/amdgpu warp shfl ops#5
Draft
diptorupd wants to merge 2 commits intoROCm:amd-integrationfrom
Draft
Conversation
Phase 0 of the monolith-warp-butterfly plan did not touch any quadrants
source; this commit only ignores the container bootstrap marker file
(.dev-installed) so the working tree stays clean after first-run of the
dev container.
The wheel produced from this commit with
QUADRANTS_CMAKE_ARGS="-DQD_WITH_VULKAN:BOOL=ON -DQD_WITH_AMDGPU:BOOL=ON \
-DQD_WITH_CUDA:BOOL=ON -DQD_BUILD_TESTS:BOOL=OFF"
is what generated the Genesis-side `baselines/phase0/` numbers. Phase 1a
branches off of here.
Made-with: Cursor
Add AMDGPU wave64 warp shuffle lowering for MI325X (gfx942/CDNA3): - New op: cuda_shfl_xor_sync_f32 - AMDGPU lowering for shfl_xor/down/sync (i32 and f32) via llvm.amdgcn.ds.bpermute - shfl_xor_f32 raises NotImplementedError on CUDA (NVPTX lowering not implemented) 8 wave64 tests validating butterfly reductions and broadcasts: - test_shfl_xor_f32_wave64_butterfly (integer values) - test_shfl_xor_f32_wave64_butterfly_fractional (IEEE 754 bitcast verification) - test_shfl_xor_i32_wave64_butterfly - test_shfl_down_f32/i32_wave64_butterfly - test_shfl_sync_f32/i32_wave64 (broadcast) - test_shfl_xor_f32_wave64_asymmetric (wrap-around)
Collaborator
|
How similar is this to: Genesis-Embodied-AI#510 |
Closed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Add AMDGPU wave64 warp shuffle lowering for MI325X (gfx942/CDNA3):
8 wave64 tests validating butterfly reductions and broadcasts: