-; RUN: llvm-as < %s | llc -march=x86-64 -mattr=+mmx | grep paddusw
-@R = external global <1 x i64> ; <<1 x i64>*> [#uses=1]
+; RUN: llc < %s -march=x86-64 -mattr=+mmx | grep paddusw
+@R = external global x86_mmx ; <x86_mmx*> [#uses=1]
define void @foo(<1 x i64> %A, <1 x i64> %B) {
entry:
- %tmp4 = bitcast <1 x i64> %B to <4 x i16> ; <<4 x i16>> [#uses=1]
- %tmp6 = bitcast <1 x i64> %A to <4 x i16> ; <<4 x i16>> [#uses=1]
- %tmp7 = tail call <4 x i16> @llvm.x86.mmx.paddus.w( <4 x i16> %tmp6, <4 x i16> %tmp4 ) ; <<4 x i16>> [#uses=1]
- %tmp8 = bitcast <4 x i16> %tmp7 to <1 x i64> ; <<1 x i64>> [#uses=1]
- store <1 x i64> %tmp8, <1 x i64>* @R
+ %tmp2 = bitcast <1 x i64> %A to x86_mmx
+ %tmp3 = bitcast <1 x i64> %B to x86_mmx
+ %tmp7 = tail call x86_mmx @llvm.x86.mmx.paddus.w( x86_mmx %tmp2, x86_mmx %tmp3 ) ; <x86_mmx> [#uses=1]
+ store x86_mmx %tmp7, x86_mmx* @R
tail call void @llvm.x86.mmx.emms( )
ret void
}
-declare <4 x i16> @llvm.x86.mmx.paddus.w(<4 x i16>, <4 x i16>)
+declare x86_mmx @llvm.x86.mmx.paddus.w(x86_mmx, x86_mmx)
declare void @llvm.x86.mmx.emms()