If a 1-bit value is in a 32-bit VGPR, the scalar opcodes set SCC to
whether the result is 0. If the inputs are SCC, these can be copied to
a 32-bit SGPR to produce an SCC result.
Details
Details
Diff Detail
Diff Detail
Paths
| Differential D64511
AMDGPU/GlobalISel: Allow scalar s1 and/or/xor ClosedPublic Authored by arsenm on Jul 10 2019, 10:05 AM.
Details
Diff Detail Event TimelineHerald added subscribers: Petar.Avramovic, t-tye, tpr and 6 others. · View Herald TranscriptJul 10 2019, 10:05 AM This revision is now accepted and ready to land.Jul 15 2019, 10:32 AM
Revision Contents
Diff 209001 lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp
test/CodeGen/AMDGPU/GlobalISel/regbankselect-and-s1.mir
test/CodeGen/AMDGPU/GlobalISel/regbankselect-and.mir
test/CodeGen/AMDGPU/GlobalISel/regbankselect-or.mir
test/CodeGen/AMDGPU/GlobalISel/regbankselect-phi-s1.mir
test/CodeGen/AMDGPU/GlobalISel/regbankselect-xor.mir
|