[all-commits] [llvm/llvm-project] abedb3: [CodeGenTypes] Speed up getVectorElementType and g...

[Jay Foad via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: abedb3b8356d5d56f1c575c4f7682fba2cb19787
      https://github.com/llvm/llvm-project/commit/abedb3b8356d5d56f1c575c4f7682fba2cb19787
  Author: Jay Foad <jay.foad at amd.com>
  Date:   2024-06-12 (Wed, 12 Jun 2024)

  Changed paths:
    M llvm/include/llvm/CodeGenTypes/MachineValueType.h
    M llvm/utils/TableGen/Common/CodeGenTarget.cpp
    M llvm/utils/TableGen/VTEmitter.cpp

  Log Message:
  -----------
  [CodeGenTypes] Speed up getVectorElementType and getVectorMinNumElements (#95282)

Implement MVT::getVectorElementType and MVT::getVectorMinNumElements
with table lookup instead of switch.

This speeds up "check-llvm-codegen-amdgpu" by about 7% in my Release
build.



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