OSDN Git Service

[AMDGPU] Better use of llvm::numbers
authorJay Foad <jay.foad@amd.com>
Fri, 29 May 2020 08:42:00 +0000 (09:42 +0100)
committerJay Foad <jay.foad@amd.com>
Fri, 29 May 2020 08:55:36 +0000 (09:55 +0100)
commitb28d038ff34dd54ce8eb9fe83506cc3742e6b85a
tree8e18c70baccb66fc56ca2eea73e94f29059cb95f
parent036d4b0dbfd1bb5f960c620ff899669d85e62e08
[AMDGPU] Better use of llvm::numbers

Tweak a few constant expressions involving numbers::pi etc to avoid
rounding errors. NFCI though it's possible some of these will now be
more accurate in the last bit.
llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp
llvm/lib/Target/AMDGPU/SIISelLowering.cpp