1; RUN: llc < %s -march=x86 -mcpu=penryn -mattr=sse4.1 | FileCheck %s 2 3; ModuleID = 'vec_shuffle-27.bc' 4target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32" 5target triple = "i686-apple-cl.1.0" 6 7define <8 x float> @my2filter4_1d(<4 x float> %a, <8 x float> %T0, <8 x float> %T1) nounwind readnone { 8entry: 9; CHECK: subps 10; CHECK: subps 11; CHECK: mulps 12; CHECK: mulps 13; CHECK: addps 14; CHECK: addps 15 %tmp7 = shufflevector <4 x float> %a, <4 x float> undef, <8 x i32> < i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3 > ; <<8 x float>> [#uses=1] 16 %sub = fsub <8 x float> %T1, %T0 ; <<8 x float>> [#uses=1] 17 %mul = fmul <8 x float> %sub, %tmp7 ; <<8 x float>> [#uses=1] 18 %add = fadd <8 x float> %mul, %T0 ; <<8 x float>> [#uses=1] 19 ret <8 x float> %add 20} 21 22; Test case for r122206 23define void @test2(<4 x i64>* %ap, <4 x i64>* %bp) nounwind { 24entry: 25; CHECK: movdqa 26 %a = load <4 x i64> * %ap 27 %b = load <4 x i64> * %bp 28 %mulaa = mul <4 x i64> %a, %a 29 %mulbb = mul <4 x i64> %b, %b 30 %mulab = mul <4 x i64> %a, %b 31 %vect1271 = shufflevector <4 x i64> %mulaa, <4 x i64> %mulbb, <4 x i32> <i32 0, i32 4, i32 undef, i32 undef> 32 %vect1272 = shufflevector <4 x i64> %mulaa, <4 x i64> %mulbb, <4 x i32> <i32 1, i32 5, i32 undef, i32 undef> 33 %vect1487 = shufflevector <4 x i64> %vect1271, <4 x i64> %mulab, <4 x i32> <i32 0, i32 1, i32 2, i32 4> 34 %vect1488 = shufflevector <4 x i64> %vect1272, <4 x i64> %mulab, <4 x i32> <i32 0, i32 1, i32 2, i32 5> 35 store <4 x i64> %vect1487, <4 x i64>* %ap 36 store <4 x i64> %vect1488, <4 x i64>* %bp 37 ret void; 38} 39