[all-commits] [llvm/llvm-project] 39e4b4: TTI: Check legalization cost of abs nodes (#100523)

[Shilei Tian via All-commits]

  Branch: refs/heads/users/shiltian/amdgpu-attributor-full-lto
  Home:   https://github.com/llvm/llvm-project
  Commit: 39e4b4af033ee39038e7a0d265bebcfdb3c213cc
      https://github.com/llvm/llvm-project/commit/39e4b4af033ee39038e7a0d265bebcfdb3c213cc
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-08-09 (Fri, 09 Aug 2024)

  Changed paths:
    M llvm/test/Analysis/CostModel/AMDGPU/arith-ssat.ll
    M llvm/test/Analysis/CostModel/AMDGPU/arith-usat.ll

  Log Message:
  -----------
  TTI: Check legalization cost of abs nodes (#100523)


  Commit: 110403c257821985c61b857cc954efb3f1ffd598
      https://github.com/llvm/llvm-project/commit/110403c257821985c61b857cc954efb3f1ffd598
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-08-09 (Fri, 09 Aug 2024)

  Changed paths:
    M llvm/test/Analysis/CostModel/AMDGPU/arith-ssat.ll
    M llvm/test/Analysis/CostModel/AMDGPU/arith-usat.ll

  Log Message:
  -----------
  AMDGPU: Correct costs of saturating add/sub intrinsics (#100808)

These are directly legal with fast instructions.


  Commit: af0b8691fca6de138bff2756e431e71febe80258
      https://github.com/llvm/llvm-project/commit/af0b8691fca6de138bff2756e431e71febe80258
  Author: Matthew Devereau <matthew.devereau at arm.com>
  Date:   2024-08-09 (Fri, 09 Aug 2024)

  Changed paths:
    M llvm/CMakeLists.txt
    M llvm/cmake/config-ix.cmake
    M llvm/include/llvm/ADT/APFloat.h
    M llvm/include/llvm/Support/float128.h
    M llvm/lib/Analysis/CMakeLists.txt
    M llvm/lib/Analysis/ConstantFolding.cpp
    M llvm/lib/Support/APFloat.cpp

  Log Message:
  -----------
  Enable logf128 constant folding for hosts with 128bit floats (#96287)

Hosts which support a float size of 128 bits can benefit from constant
fp128 folding.


  Commit: 59fb1dd4990c919b5507de64ec40aa0d6f9f9ec2
      https://github.com/llvm/llvm-project/commit/59fb1dd4990c919b5507de64ec40aa0d6f9f9ec2
  Author: Nikita Popov <npopov at redhat.com>
  Date:   2024-08-09 (Fri, 09 Aug 2024)

  Changed paths:
    M llvm/CMakeLists.txt
    M llvm/cmake/config-ix.cmake
    M llvm/include/llvm/ADT/APFloat.h
    M llvm/include/llvm/Support/float128.h
    M llvm/lib/Analysis/CMakeLists.txt
    M llvm/lib/Analysis/ConstantFolding.cpp
    M llvm/lib/Support/APFloat.cpp

  Log Message:
  -----------
  Revert "Enable logf128 constant folding for hosts with 128bit floats (#96287)"

This reverts commit ccb2b011e577e861254f61df9c59494e9e122b38.

Causes buildbot failures, e.g. on ppc64le builders.


  Commit: 659844f5dc23246fd06fd8b9ca667c2a13b0849b
      https://github.com/llvm/llvm-project/commit/659844f5dc23246fd06fd8b9ca667c2a13b0849b
  Author: Shilei Tian <i at tianshilei.me>
  Date:   2024-08-09 (Fri, 09 Aug 2024)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
    M llvm/test/CodeGen/AMDGPU/amdgpu-simplify-libcall-sincos.ll
    M llvm/test/CodeGen/AMDGPU/attributor-noopt.ll
    M llvm/test/CodeGen/AMDGPU/inline-attr.ll
    A llvm/test/CodeGen/AMDGPU/print-pipeline-passes.ll
    M llvm/test/CodeGen/AMDGPU/simplify-libcalls.ll

  Log Message:
  -----------
  [AMDGPU] Add an option to run `AMDGPUAttributorPass` in full LTO

Currently `AMDGPUAttributorPass` is registered in default optimizer pipeline.
This will allow the pass to run in default pipeline as well as at thinLTO post
link stage. However, it will not run in full LTO post link stage. This patch
adds an option to run it in full LTO. It is set to false by default.


Compare: https://github.com/llvm/llvm-project/compare/ee478979b44a...659844f5dc23

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