forked from OSchip/llvm-project
12 lines
388 B
LLVM
12 lines
388 B
LLVM
; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 | grep vnmul.f64
|
|
; RUN: llc < %s -march=arm -mattr=+v6,+vfp2 -enable-sign-dependent-rounding-fp-math | grep vmul.f64
|
|
|
|
|
|
define double @t1(double %a, double %b) {
|
|
entry:
|
|
%tmp2 = fsub double -0.000000e+00, %a ; <double> [#uses=1]
|
|
%tmp4 = fmul double %tmp2, %b ; <double> [#uses=1]
|
|
ret double %tmp4
|
|
}
|
|
|