llvm-project
[AArch64] Avoid folding sign-extend of vector extracts into ALU ops
#183522

Merged

[AArch64] Avoid folding sign-extend of vector extracts into ALU ops #183522

MacDue merged 4 commits into llvm:main from MacDue:saddlv

Precommit tests

cf1b1ffb

[AArch64] Prefer `SADDLV` over `ADDV` for vector mask reductions

fecd62c7

MacDue requested a review from

paulwalker-arm 113 days ago

llvmbot added backend:AArch64

paulwalker-arm commented on 2026-03-03

Fixups

252cced8

paulwalker-arm approved these changes on 2026-03-03

davemgreen commented on 2026-03-03

Jettison combine

4a1e7c80

MacDue changed the title ~~[AArch64] Prefer `SADDLV` over `ADDV` for vector mask reductions~~ [AArch64] Avoid folding sign-extend of vector extracts into ALU ops 108 days ago

davemgreen approved these changes on 2026-03-04

MacDue merged c62d5f35 into main 107 days ago

MacDue deleted the saddlv branch 107 days ago

Reviewers

davemgreen

paulwalker-arm

Assignees

No one assigned

Labels

backend:AArch64

Milestone

No milestone

llvm-project [AArch64] Avoid folding sign-extend of vector extracts into ALU ops #183522 Merged

[AArch64] Avoid folding sign-extend of vector extracts into ALU ops #183522

llvm-project
[AArch64] Avoid folding sign-extend of vector extracts into ALU ops
#183522

Merged