llvm-project/llvm/test/CodeGen/Hexagon/dsub.ll

19 lines
541 B
LLVM

; RUN: llc -march=hexagon -mcpu=hexagonv5 < %s | FileCheck %s
; Check that we generate double precision floating point subtract in V5.
; CHECK: call __hexagon_subdf3
define i32 @main() nounwind {
entry:
%a = alloca double, align 8
%b = alloca double, align 8
%c = alloca double, align 8
store double 1.540000e+01, double* %a, align 8
store double 9.100000e+00, double* %b, align 8
%0 = load double* %b, align 8
%1 = load double* %a, align 8
%sub = fsub double %0, %1
store double %sub, double* %c, align 8
ret i32 0
}