projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 01a2cb5) | patch
[X86][SSE] Let 64-bit targets use the fast 2i32-2f32 UINT_TO_FP conversion as well...
authorSimon Pilgrim <llvm-dev@redking.me.uk>
Sat, 30 Jul 2016 14:06:59 +0000 (14:06 +0000)
committerSimon Pilgrim <llvm-dev@redking.me.uk>
Sat, 30 Jul 2016 14:06:59 +0000 (14:06 +0000)
commitcf49fa3251ae5f0be823eaa167741333710d134f
treec60878bd914536af88387250c0b07c6cab75b952tree | snapshot
parent01a2cb55f1ea6b04cb6f5b514d302179ce9fe811commit | diff
[X86][SSE] Let 64-bit targets use the fast 2i32-2f32 UINT_TO_FP conversion as well as 32-bit

The 2i32-2i64 legalization means that we can use the slightly quicker double bits + fptrunc approach for the same results

llvm-svn: 277271
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/test/CodeGen/X86/cvtv2f32.ll diff | blob | history
Domain: System / Toolchain;