[x86] eliminate unnecessary shuffling/moves with unary scalar math ops (PR21507)
[oota-llvm.git] / docs / gcc-loops.png
docs/gcc-loops.png