[all-commits] [llvm/llvm-project] 58d70d: [AArch64] Keep floating-point conversion in SIMD (...

[Guy David via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 58d70dc62b219cd89ba434c96928a0d9c1b23a60
      https://github.com/llvm/llvm-project/commit/58d70dc62b219cd89ba434c96928a0d9c1b23a60
  Author: Guy David <49722543+guy-david at users.noreply.github.com>
  Date:   2025-07-30 (Wed, 30 Jul 2025)

  Changed paths:
    M llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
    M llvm/lib/Target/AArch64/AArch64InstrInfo.td
    M llvm/test/CodeGen/AArch64/selectopt-const.ll
    A llvm/test/CodeGen/AArch64/store-float-conversion.ll
    M llvm/test/CodeGen/AArch64/tbl-loops.ll

  Log Message:
  -----------
  [AArch64] Keep floating-point conversion in SIMD (#147707)

Stores can be issued faster if the result is kept in the SIMD/FP
registers.
The `HasOneUse` guards against creating two floating point conversions,
if for example there's some arithmetic done on the converted value as
well. Another approach would be to inspect the user instructions during
lowering, but I don't see that type of check in the lowering too often.



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