[llvm] r352297 - AMDGPU/GlobalISel: Use scalarize instead of clampMaxNumElements
Matt Arsenault via llvm-commits
llvm-commits at lists.llvm.org
Sat Jan 26 15:54:53 PST 2019
Author: arsenm
Date: Sat Jan 26 15:54:53 2019
New Revision: 352297
URL: http://llvm.org/viewvc/llvm-project?rev=352297&view=rev
Log:
AMDGPU/GlobalISel: Use scalarize instead of clampMaxNumElements
Modified:
llvm/trunk/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp
Modified: llvm/trunk/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp?rev=352297&r1=352296&r2=352297&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp (original)
+++ llvm/trunk/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp Sat Jan 26 15:54:53 2019
@@ -201,8 +201,7 @@ AMDGPULegalizerInfo::AMDGPULegalizerInfo
.legalFor({{S1, S32}, {S1, S64}})
.widenScalarToNextPow2(1)
.clampScalar(1, S32, S64)
- .clampMaxNumElements(0, S1, 1)
- .clampMaxNumElements(1, S32, 1);
+ .scalarize(0);
// FIXME: fexp, flog2, flog10 needs to be custom lowered.
getActionDefinitionsBuilder({G_FPOW, G_FEXP, G_FEXP2,
More information about the llvm-commits
mailing list