projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: e41e366) | patch
[PowerPC] Exploit the vector min/max instructions
authorNemanja Ivanovic <nemanja.i.ibm@gmail.com>
Thu, 6 Jun 2019 23:49:01 +0000 (23:49 +0000)
committerNemanja Ivanovic <nemanja.i.ibm@gmail.com>
Thu, 6 Jun 2019 23:49:01 +0000 (23:49 +0000)
commitef4a3aa549eadc150d4a31cff655064c3a323d12
treef32e37feaee44ccae31c4574f919d5704b7b186ctree | snapshot
parente41e366ae72741261bfe71ec85cce4e2aed29109commit | diff
[PowerPC] Exploit the vector min/max instructions

Use the PPC vector min/max instructions for computing the corresponding
operation as these should be faster than the compare/select sequences
we currently emit.

Differential revision: https://reviews.llvm.org/D47332

llvm-svn: 362759
llvm/lib/Target/PowerPC/PPCISelLowering.cpp diff | blob | history
llvm/lib/Target/PowerPC/PPCInstrAltivec.td diff | blob | history
llvm/lib/Target/PowerPC/PPCInstrVSX.td diff | blob | history
llvm/test/CodeGen/PowerPC/ctr-minmaxnum.ll diff | blob | history
llvm/test/CodeGen/PowerPC/sat-add.ll diff | blob | history
llvm/test/CodeGen/PowerPC/vec-min-max.ll [new file with mode: 0644] blob
llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll diff | blob | history
Domain: System / Toolchain;