projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 146df55) | patch
[AMDGPU] Add SIPreEmitPeephole pass.
authorcdevadas <Christudasan.Devadasan@amd.com>
Wed, 25 Mar 2020 14:15:16 +0000 (14:15 +0000)
committercdevadas <Christudasan.Devadasan@amd.com>
Wed, 25 Mar 2020 15:35:35 +0000 (15:35 +0000)
commitce984129eaa54e05b3089bb9ca4aecb3e0bf7b31
treef8b5c11fb848e679758783a59d4d8fe7fd978151tree | snapshot
parent146df5581d9825505f6a4fef976842a1f36bdeb2commit | diff
[AMDGPU] Add SIPreEmitPeephole pass.

This pass can handle all the optimization
opportunities found just before code emission.
Presently it includes the handling of vcc branch
optimization that was handled earlier in SIInsertSkips.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D76712
llvm/lib/Target/AMDGPU/AMDGPU.h diff | blob | history
llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp diff | blob | history
llvm/lib/Target/AMDGPU/CMakeLists.txt diff | blob | history
llvm/lib/Target/AMDGPU/SIInsertSkips.cpp diff | blob | history
llvm/lib/Target/AMDGPU/SIPreEmitPeephole.cpp [new file with mode: 0644] blob
llvm/test/CodeGen/AMDGPU/insert-skip-from-vcc.mir diff | blob | history
Domain: System / Toolchain;