llvm-project
[AMDGPU] Optimize S_OR_B32 to S_ADDK_I32 where possible
#177949

Merged

[AMDGPU] Optimize S_OR_B32 to S_ADDK_I32 where possible #177949

arsenm merged 11 commits into llvm:main from Iasonaskrpr:AMDGPU-fold-disjoint-s_or_b32

arsenm added backend:AMDGPU

arsenm commented on 2026-02-01

AMDGPU added conversion from disjoint s_or_b32 to s_addk_i32

95a0c4aa

Formatted changes

67c65442

Removed unused variable

4b1744ef

Updated tests to work with transformation

511fa2ac

Correctly formatted file

b38405f2

Updated tests, and all of them pass, also now S_OR to bitset conversi…

33966c0f

Updated tests to not have conflicts

0c6e2158

Iasonaskrpr force pushed to 0c6e2158 139 days ago

Merge branch 'main' into AMDGPU-fold-disjoint-s_or_b32

e4f37a0a

Iasonaskrpr marked this pull request as ready for review 139 days ago

llvmbot added llvm:globalisel

jayfoad approved these changes on 2026-02-05

Reverted change in bswap

3c9e4afd

arsenm commented on 2026-02-05

Merge branch 'llvm:main' into AMDGPU-fold-disjoint-s_or_b32

5fcfca8f

Added test

b3f8f8dc

arsenm approved these changes on 2026-02-07

arsenm enabled auto-merge (squash) 136 days ago

arsenm merged 6c6fb00c into main 136 days ago

Iasonaskrpr deleted the AMDGPU-fold-disjoint-s_or_b32 branch 40 days ago

Reviewers

arsenm

jayfoad

Assignees

No one assigned

Labels

backend:AMDGPU llvm:globalisel

Milestone

No milestone