Programming Languages Research Group: Git

author	Tom Stellard <thomas.stellard@amd.com>
	Sat, 27 Jul 2013 00:01:07 +0000 (00:01 +0000)
committer	Tom Stellard <thomas.stellard@amd.com>
	Sat, 27 Jul 2013 00:01:07 +0000 (00:01 +0000)
commit	57e6b2d1f3de0bf459e96f7038e692d624f7e580
tree	42204d51445bcb178a24fa07fe9da2d1d22818ff	tree \| snapshot
parent	018696db4587dc6a0d8081d0627543f5c08b6c97	commit \| diff

SimplifyCFG: Use parallel-and and parallel-or mode to consolidate branch conditions

Merge consecutive if-regions if they contain identical statements.
Both transformations reduce number of branches. The transformation
is guarded by a target-hook, and is currently enabled only for +R600,
but the correctness has been tested on X86 target using a variety of
CPU benchmarks.

Patch by: Mei Ye

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187278 91177308-0d34-0410-b5e6-96231b3b80d8

17 files changed:

include/llvm/Analysis/TargetTransformInfo.h		diff \| blob \| history
include/llvm/InitializePasses.h		diff \| blob \| history
include/llvm/Transforms/Scalar.h		diff \| blob \| history
include/llvm/Transforms/Utils/Local.h		diff \| blob \| history
lib/Analysis/TargetTransformInfo.cpp		diff \| blob \| history
lib/CodeGen/BasicTargetTransformInfo.cpp		diff \| blob \| history
lib/Target/R600/AMDGPU.h		diff \| blob \| history
lib/Target/R600/AMDGPUTargetMachine.cpp		diff \| blob \| history
lib/Target/R600/AMDGPUTargetMachine.h		diff \| blob \| history
lib/Target/R600/AMDGPUTargetTransformInfo.cpp	[new file with mode: 0644]	blob
lib/Target/R600/CMakeLists.txt		diff \| blob \| history
lib/Transforms/IPO/PassManagerBuilder.cpp		diff \| blob \| history
lib/Transforms/Scalar/Scalar.cpp		diff \| blob \| history
lib/Transforms/Scalar/SimplifyCFGPass.cpp		diff \| blob \| history
lib/Transforms/Utils/SimplifyCFG.cpp		diff \| blob \| history
tools/lto/LTOCodeGenerator.cpp		diff \| blob \| history
tools/opt/opt.cpp		diff \| blob \| history

C/C++ LLVM-based compilers that forbids OOTA behaviors

RSS Atom