1; RUN: llc -march=r600 -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s 2 3 4; FUNC-LABEL: @scalar_to_vector_v2i32 5; SI: BUFFER_LOAD_DWORD [[VAL:v[0-9]+]], 6; SI: V_LSHRREV_B32_e32 [[RESULT:v[0-9]+]], 16, [[VAL]] 7; SI: BUFFER_STORE_SHORT [[RESULT]] 8; SI: BUFFER_STORE_SHORT [[RESULT]] 9; SI: BUFFER_STORE_SHORT [[RESULT]] 10; SI: BUFFER_STORE_SHORT [[RESULT]] 11; SI: S_ENDPGM 12define void @scalar_to_vector_v2i32(<4 x i16> addrspace(1)* %out, i32 addrspace(1)* %in) nounwind { 13 %tmp1 = load i32 addrspace(1)* %in, align 4 14 %bc = bitcast i32 %tmp1 to <2 x i16> 15 %tmp2 = shufflevector <2 x i16> %bc, <2 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> 16 store <4 x i16> %tmp2, <4 x i16> addrspace(1)* %out, align 8 17 ret void 18} 19 20; FUNC-LABEL: @scalar_to_vector_v2f32 21; SI: BUFFER_LOAD_DWORD [[VAL:v[0-9]+]], 22; SI: V_LSHRREV_B32_e32 [[RESULT:v[0-9]+]], 16, [[VAL]] 23; SI: BUFFER_STORE_SHORT [[RESULT]] 24; SI: BUFFER_STORE_SHORT [[RESULT]] 25; SI: BUFFER_STORE_SHORT [[RESULT]] 26; SI: BUFFER_STORE_SHORT [[RESULT]] 27; SI: S_ENDPGM 28define void @scalar_to_vector_v2f32(<4 x i16> addrspace(1)* %out, float addrspace(1)* %in) nounwind { 29 %tmp1 = load float addrspace(1)* %in, align 4 30 %bc = bitcast float %tmp1 to <2 x i16> 31 %tmp2 = shufflevector <2 x i16> %bc, <2 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> 32 store <4 x i16> %tmp2, <4 x i16> addrspace(1)* %out, align 8 33 ret void 34} 35 36; Getting a SCALAR_TO_VECTOR seems to be tricky. These cases managed 37; to produce one, but for some reason never made it to selection. 38 39 40; define void @scalar_to_vector_test2(<8 x i8> addrspace(1)* %out, i32 addrspace(1)* %in) nounwind { 41; %tmp1 = load i32 addrspace(1)* %in, align 4 42; %bc = bitcast i32 %tmp1 to <4 x i8> 43 44; %tmp2 = shufflevector <4 x i8> %bc, <4 x i8> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1> 45; store <8 x i8> %tmp2, <8 x i8> addrspace(1)* %out, align 4 46; ret void 47; } 48 49; define void @scalar_to_vector_test3(<4 x i32> addrspace(1)* %out) nounwind { 50; %newvec0 = insertelement <2 x i64> undef, i64 12345, i32 0 51; %newvec1 = insertelement <2 x i64> %newvec0, i64 undef, i32 1 52; %bc = bitcast <2 x i64> %newvec1 to <4 x i32> 53; %add = add <4 x i32> %bc, <i32 1, i32 2, i32 3, i32 4> 54; store <4 x i32> %add, <4 x i32> addrspace(1)* %out, align 16 55; ret void 56; } 57 58; define void @scalar_to_vector_test4(<8 x i16> addrspace(1)* %out) nounwind { 59; %newvec0 = insertelement <4 x i32> undef, i32 12345, i32 0 60; %bc = bitcast <4 x i32> %newvec0 to <8 x i16> 61; %add = add <8 x i16> %bc, <i16 1, i16 2, i16 3, i16 4, i16 1, i16 2, i16 3, i16 4> 62; store <8 x i16> %add, <8 x i16> addrspace(1)* %out, align 16 63; ret void 64; } 65 66; define void @scalar_to_vector_test5(<4 x i16> addrspace(1)* %out) nounwind { 67; %newvec0 = insertelement <2 x i32> undef, i32 12345, i32 0 68; %bc = bitcast <2 x i32> %newvec0 to <4 x i16> 69; %add = add <4 x i16> %bc, <i16 1, i16 2, i16 3, i16 4> 70; store <4 x i16> %add, <4 x i16> addrspace(1)* %out, align 16 71; ret void 72; } 73 74; define void @scalar_to_vector_test6(<4 x i16> addrspace(1)* %out) nounwind { 75; %newvec0 = insertelement <2 x i32> undef, i32 12345, i32 0 76; %bc = bitcast <2 x i32> %newvec0 to <4 x i16> 77; %add = add <4 x i16> %bc, <i16 1, i16 2, i16 3, i16 4> 78; store <4 x i16> %add, <4 x i16> addrspace(1)* %out, align 16 79; ret void 80; } 81