-; RUN: llc < %s -march=x86 -mattr=+sse,+sse2 | FileCheck %s
+; RUN: llc < %s -march=x86 -mattr=+sse,+sse2,+sse41 | FileCheck %s
target datalayout = "e-p:32:32"
target triple = "i686-apple-darwin8.7.2"
%tmp69 = trunc i32 %tmp.upgrd.1 to i16 ; <i16> [#uses=1]
ret i16 %tmp69
; CHECK: test1:
-; CHECK: subss LCPI1_
-; CHECK: mulss LCPI1_
-; CHECK: minss LCPI1_
+; CHECK: subss LCPI0_
+; CHECK: mulss LCPI0_
+; CHECK: minss LCPI0_
}
define i16 @test2(float %f) nounwind {
%tmp69 = trunc i32 %tmp to i16 ; <i16> [#uses=1]
ret i16 %tmp69
; CHECK: test2:
-; CHECK: addss LCPI2_
-; CHECK: mulss LCPI2_
-; CHECK: minss LCPI2_
+; CHECK: addss LCPI1_
+; CHECK: mulss LCPI1_
+; CHECK: minss LCPI1_
}
declare <4 x float> @llvm.x86.sse.sub.ss(<4 x float>, <4 x float>)
declare <4 x float> @llvm.x86.sse.max.ss(<4 x float>, <4 x float>)
declare i32 @llvm.x86.sse.cvttss2si(<4 x float>)
+
+
+declare <4 x float> @llvm.x86.sse41.round.ss(<4 x float>, <4 x float>, i32)
+declare <4 x float> @f()
+
+define <4 x float> @test3(<4 x float> %A, float *%b, i32 %C) nounwind {
+ %a = load float *%b
+ %B = insertelement <4 x float> undef, float %a, i32 0
+ %X = call <4 x float> @llvm.x86.sse41.round.ss(<4 x float> %A, <4 x float> %B, i32 4)
+ ret <4 x float> %X
+; CHECK: test3:
+; CHECK: roundss $4, (%eax), %xmm0
+}
+
+define <4 x float> @test4(<4 x float> %A, float *%b, i32 %C) nounwind {
+ %a = load float *%b
+ %B = insertelement <4 x float> undef, float %a, i32 0
+ %q = call <4 x float> @f()
+ %X = call <4 x float> @llvm.x86.sse41.round.ss(<4 x float> %q, <4 x float> %B, i32 4)
+ ret <4 x float> %X
+; CHECK: test4:
+; CHECK: movss (%eax), %xmm
+; CHECK: call
+; CHECK: roundss $4, %xmm{{.*}}, %xmm0
+}