projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 11e0876) | patch
R600/SI: Use v_cvt_f32_ubyte* instructions
authorMatt Arsenault <Matthew.Arsenault@amd.com>
Wed, 11 Jun 2014 17:50:44 +0000 (17:50 +0000)
committerMatt Arsenault <Matthew.Arsenault@amd.com>
Wed, 11 Jun 2014 17:50:44 +0000 (17:50 +0000)
commit364a6747aa06ead940783e99f09e4759d271bdb3
tree63a1323b8c9a0dea49763830e795eaf8d24bc210tree | snapshot
parent11e0876bb2098ad734e9f5bbae31ce6ebb9f84cfcommit | diff
R600/SI: Use v_cvt_f32_ubyte* instructions

This eliminates extra extract instructions when loading an i8 vector to
a float vector.

llvm-svn: 210666
llvm/lib/Target/R600/AMDGPUISelLowering.cpp diff | blob | history
llvm/lib/Target/R600/AMDGPUISelLowering.h diff | blob | history
llvm/lib/Target/R600/AMDGPUInstrInfo.td diff | blob | history
llvm/lib/Target/R600/AMDGPUIntrinsics.td diff | blob | history
llvm/lib/Target/R600/SIISelLowering.cpp diff | blob | history
llvm/lib/Target/R600/SIISelLowering.h diff | blob | history
llvm/lib/Target/R600/SIInstructions.td diff | blob | history
llvm/test/CodeGen/R600/bitcast.ll diff | blob | history
llvm/test/CodeGen/R600/cvt_f32_ubyte.ll [new file with mode: 0644] blob
llvm/test/CodeGen/R600/llvm.AMDGPU.cvt_f32_ubyte.ll [new file with mode: 0644] blob
Domain: System / Toolchain;