[all-commits] [llvm/llvm-project] 3d369c: [AArch64] Prefer SVE2 for fixed-length i64 [S|U][M...

[Benjamin Maxwell via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 3d369ccdf976de00885cc95f8c9ac5a44bc8f16e
      https://github.com/llvm/llvm-project/commit/3d369ccdf976de00885cc95f8c9ac5a44bc8f16e
  Author: Benjamin Maxwell <benjamin.maxwell at arm.com>
  Date:   2026-02-18 (Wed, 18 Feb 2026)

  Changed paths:
    M llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
    M llvm/lib/Target/AArch64/AArch64ISelLowering.h
    M llvm/test/CodeGen/AArch64/aarch64-minmaxv.ll
    M llvm/test/CodeGen/AArch64/sve-fixed-length-int-reduce.ll
    M llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-reduce.ll

  Log Message:
  -----------
  [AArch64] Prefer SVE2 for fixed-length i64 [S|U][MIN|MAX] reductions (#181161)

With SVE2/SME we can lower the v2i64 min/max reductions to an SVE2
pairwise instruction. The throughput is about the same, but the SVE code
is smaller than the NEON expansion.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications