Repository navigation
test(rccl): add LL/LL128 vector ops coverage and op128 refactor - #10674
Closed
alex-breslow-amd wants to merge 4 commits into
Closed
alex-breslow-amd wants to merge 4 commits into
alex-breslow-amd wants to merge 4 commits into
Conversation
Replace paired 64-bit nontemporal loads and plain stores in LL FIFO access with single v4u vector operations, matching the LL128 load128NT/store128Plain approach. Preserves the gfx1200/gfx1201 dual atomic-store fallback when b128 builtins are unavailable. Co-authored-by: Cursor <cursoragent@cursor.com>
Add device fixture tests that mirror loadLLLine/storeLLLine and LL128 vector helpers, plus MPI AllReduce tests that force LL and registered LL128 paths (including gfx1250 cooperative atomics). Co-authored-by: Cursor <cursoragent@cursor.com>
Move load128NT, store128Plain, and store128LLFifo into op128.h so unit and MPI tests exercise production code instead of duplicated helpers. Thin wrappers remain in prims_ll.h; prims_ll128.h uses op128 via common.h. MPI tests now assert NCCL_TUNING logs show the forced LL/LL128 protocol. Co-authored-by: Cursor <cursoragent@cursor.com>
alex-breslow-amd
changed the title
Comment on lines
+83
to
+84
| __hip_atomic_store((u64_gptr)v, v0, __ATOMIC_RELAXED, __HIP_MEMORY_SCOPE_SYSTEM); | ||
| __hip_atomic_store((u64_gptr)v + 1, v1, __ATOMIC_RELEASE, __HIP_MEMORY_SCOPE_SYSTEM); |
Contributor
Author
There was a problem hiding this comment.
TODO: scoped clang atomic instead
alex-breslow-amd
force-pushed
the
users/abreslow/ll_cleanup
branch
from
September 24, 2026 18:50
d4b8ddd to
808dc4f
Compare
alex-breslow-amd
changed the base branch from
users/abreslow/ll_cleanup
to
develop
September 24, 2026 18:55
✅ All Policy Checks Passed
📖 Need help? See the Policy FAQ for details on every check and how to fix failures. |
Contributor
|
This pull request has been inactive for 25 days and will be marked as stale. If you would like to keep this PR open, please:
This PR will be automatically closed in 5 days if no further activity occurs. |
Contributor
|
This pull request has been automatically closed due to inactivity (30 days with no updates). If you'd like to continue working on this, feel free to reopen the PR or create a new one. |
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
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
load128NT,store128Plain,store128LLFifo) intoop128.has the single source of truthprims_ll.h;prims_ll128.husesop128.hvia existing include chain#include "op128.h"directly (no duplicated helper source)Files changed
op128.h,prims_ll.h,prims_ll128.htest/device/TestLLFifo.cpp,test/device/TestLL128Vector.cpptest/LL128VectorMPITests.cpp,test/CMakeLists.txtTest plan
./test/rccl-UnitTestsFixtures --gtest_filter='LLFifoVectorTest.*:LL128VectorTest.*'mpirun -np 2 ./test/rccl-UnitTestsMPI --gtest_filter=LL128VectorMPITest.*Load128Store128Gfx1250CooperativePathandRegisteredAllReduce_LL128_Gfx1250Onlypass