diff options
| author | Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com> | 2017-02-03 20:08:29 +0000 |
|---|---|---|
| committer | Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com> | 2017-02-03 20:08:29 +0000 |
| commit | 81db53109dc782072f79b012db673d0503310523 (patch) | |
| tree | a0f333f40bc5b1fcc71211f9737afc51a742f8e1 | |
| parent | 1fa5eacf9d33933b878b4ab749c118e36bf47835 (diff) | |
| download | bcm5719-llvm-81db53109dc782072f79b012db673d0503310523.tar.gz bcm5719-llvm-81db53109dc782072f79b012db673d0503310523.zip | |
[AMDGPU] Bump -amdgpu-unroll-threshold-private to 2000
This has quite positive performance impact according to measurements.
Before previous fixes to limit the optimization that was too high
and blowed compile time and scratch usage, but now this is gone and
we can bump the threshold.
Differential Revision: https://reviews.llvm.org/D29505
llvm-svn: 294032
| -rw-r--r-- | llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp index 1eeff6e16d2..f3c94182730 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp @@ -32,7 +32,7 @@ using namespace llvm; static cl::opt<unsigned> UnrollThresholdPrivate( "amdgpu-unroll-threshold-private", cl::desc("Unroll threshold for AMDGPU if private memory used in a loop"), - cl::init(800), cl::Hidden); + cl::init(2000), cl::Hidden); void AMDGPUTTIImpl::getUnrollingPreferences(Loop *L, TTI::UnrollingPreferences &UP) { |
