1; RUN: llc < %s -march=x86 -mattr=+sse42 | FileCheck %s 2 3; widening shuffle v3float and then a add 4define void @shuf(<3 x float>* %dst.addr, <3 x float> %src1,<3 x float> %src2) nounwind { 5entry: 6; CHECK: shuf: 7; CHECK: extractps 8; CHECK: extractps 9 %x = shufflevector <3 x float> %src1, <3 x float> %src2, <3 x i32> < i32 0, i32 1, i32 2> 10 %val = fadd <3 x float> %x, %src2 11 store <3 x float> %val, <3 x float>* %dst.addr 12 ret void 13} 14 15 16; widening shuffle v3float with a different mask and then a add 17define void @shuf2(<3 x float>* %dst.addr, <3 x float> %src1,<3 x float> %src2) nounwind { 18entry: 19; CHECK: shuf2: 20; CHECK: extractps 21; CHECK: extractps 22 %x = shufflevector <3 x float> %src1, <3 x float> %src2, <3 x i32> < i32 0, i32 4, i32 2> 23 %val = fadd <3 x float> %x, %src2 24 store <3 x float> %val, <3 x float>* %dst.addr 25 ret void 26} 27 28; Example of when widening a v3float operation causes the DAG to replace a node 29; with the operation that we are currently widening, i.e. when replacing 30; opA with opB, the DAG will produce new operations with opA. 31define void @shuf3(<4 x float> %tmp10, <4 x float> %vecinit15, <4 x float>* %dst) nounwind { 32entry: 33; CHECK: shuf3: 34; CHECK: pshufd 35 %shuffle.i.i.i12 = shufflevector <4 x float> %tmp10, <4 x float> %vecinit15, <4 x i32> <i32 0, i32 1, i32 4, i32 5> 36 %tmp25.i.i = shufflevector <4 x float> %shuffle.i.i.i12, <4 x float> undef, <3 x i32> <i32 0, i32 1, i32 2> 37 %tmp1.i.i = shufflevector <3 x float> %tmp25.i.i, <3 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3> 38 %tmp3.i13 = shufflevector <4 x float> %tmp1.i.i, <4 x float> undef, <3 x i32> <i32 0, i32 1, i32 2> ; <<3 x float>> 39 %tmp6.i14 = shufflevector <3 x float> %tmp3.i13, <3 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3> 40 %tmp97.i = shufflevector <4 x float> %tmp6.i14, <4 x float> undef, <3 x i32> <i32 0, i32 1, i32 2> 41 %tmp2.i18 = shufflevector <3 x float> %tmp97.i, <3 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 2> 42 %t5 = bitcast <4 x float> %tmp2.i18 to <4 x i32> 43 %shr.i.i19 = lshr <4 x i32> %t5, <i32 19, i32 19, i32 19, i32 19> 44 %and.i.i20 = and <4 x i32> %shr.i.i19, <i32 4080, i32 4080, i32 4080, i32 4080> 45 %shuffle.i.i.i21 = shufflevector <4 x float> %tmp2.i18, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 2, i32 3> 46 store <4 x float> %shuffle.i.i.i21, <4 x float>* %dst 47 ret void 48} 49 50; PR10421: make sure we correctly handle extreme widening with CONCAT_VECTORS 51define <8 x i8> @shuf4(<4 x i8> %a, <4 x i8> %b) nounwind readnone { 52; CHECK: shuf4: 53; CHECK: punpckldq 54 %vshuf = shufflevector <4 x i8> %a, <4 x i8> %b, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> 55 ret <8 x i8> %vshuf 56} 57