[all-commits] [llvm/llvm-project] 1b1297: [AArch64][AMDGPU][GlobalISel] Remove vector handli...

David Green via All-commits all-commits at lists.llvm.org
Tue Feb 20 02:54:56 PST 2024


  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 1b12974ccba6478b27d7ba13ceead2864ab4be20
      https://github.com/llvm/llvm-project/commit/1b12974ccba6478b27d7ba13ceead2864ab4be20
  Author: David Green <david.green at arm.com>
  Date:   2024-02-20 (Tue, 20 Feb 2024)

  Changed paths:
    M llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp
    M llvm/test/CodeGen/AArch64/GlobalISel/combine-unmerge.mir
    M llvm/test/CodeGen/AArch64/aarch64-bif-gen.ll
    M llvm/test/CodeGen/AArch64/aarch64-bit-gen.ll
    M llvm/test/CodeGen/AArch64/abs.ll
    M llvm/test/CodeGen/AArch64/arm64-neon-copy.ll
    M llvm/test/CodeGen/AArch64/arm64-vcvt_f.ll
    M llvm/test/CodeGen/AArch64/bswap.ll
    M llvm/test/CodeGen/AArch64/fpext.ll
    M llvm/test/CodeGen/AArch64/fptoi.ll
    M llvm/test/CodeGen/AArch64/llvm.exp10.ll
    M llvm/test/CodeGen/AArch64/reduce-and.ll
    M llvm/test/CodeGen/AArch64/reduce-or.ll
    M llvm/test/CodeGen/AArch64/reduce-xor.ll
    M llvm/test/CodeGen/AArch64/shift.ll
    M llvm/test/CodeGen/AArch64/vecreduce-fmax-legalization-nan.ll
    M llvm/test/CodeGen/AArch64/vecreduce-fmax-legalization.ll
    M llvm/test/CodeGen/AArch64/vecreduce-fmaximum.ll
    M llvm/test/CodeGen/AArch64/vecreduce-fmin-legalization.ll
    M llvm/test/CodeGen/AArch64/vecreduce-fminimum.ll
    M llvm/test/CodeGen/AArch64/vecreduce-umax-legalization.ll
    M llvm/test/CodeGen/AMDGPU/GlobalISel/combine-fma-add-mul-post-legalize.mir
    M llvm/test/CodeGen/AMDGPU/GlobalISel/combine-fma-add-mul-pre-legalize.mir

  Log Message:
  -----------
  [AArch64][AMDGPU][GlobalISel] Remove vector handling from unmerge_dead_to_trunc (#82224)

This combine transforms an unmerge where only the first element is used
into a truncate. That works OK for scalar but for vector needs to insert
a bitcast to integers, perform the truncate then bitcast back to
vectors. This generates more awkward code than using an Unmerge.



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


More information about the All-commits mailing list