projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: fb51aac) | patch
[PowerPC] Make use of VSX f64 <-> i64 conversion instructions
authorHal Finkel <hfinkel@anl.gov>
Sun, 23 Mar 2014 05:35:00 +0000 (05:35 +0000)
committerHal Finkel <hfinkel@anl.gov>
Sun, 23 Mar 2014 05:35:00 +0000 (05:35 +0000)
commitb6cbecd27298ca795eba350ee68fb82c4f4e422a
treec2f979a264e331db74bbefe555181b44f9f0745etree | snapshot
parentfb51aac12966d07cf512e17ddae9183517d7798ecommit | diff
[PowerPC] Make use of VSX f64 <-> i64 conversion instructions

When VSX is available, these instructions should be used in preference to the
older variants that only have access to the scalar floating-point registers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204559 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/PowerPC/PPCInstrVSX.td diff | blob | history
test/CodeGen/PowerPC/float-to-int.ll diff | blob | history
test/CodeGen/PowerPC/i32-to-float.ll diff | blob | history
test/CodeGen/PowerPC/i64-to-float.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors