1 ; RUN: llc -march=amdgcn -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
2 ; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
5 ; FUNC-LABEL: {{^}}scalar_to_vector_v2i32:
6 ; SI: buffer_load_dword [[VAL:v[0-9]+]],
7 ; SI: v_lshrrev_b32_e32 [[RESULT:v[0-9]+]], 16, [[VAL]]
8 ; SI: buffer_store_short [[RESULT]]
9 ; SI: buffer_store_short [[RESULT]]
10 ; SI: buffer_store_short [[RESULT]]
11 ; SI: buffer_store_short [[RESULT]]
13 define void @scalar_to_vector_v2i32(<4 x i16> addrspace(1)* %out, i32 addrspace(1)* %in) nounwind {
14 %tmp1 = load i32, i32 addrspace(1)* %in, align 4
15 %bc = bitcast i32 %tmp1 to <2 x i16>
16 %tmp2 = shufflevector <2 x i16> %bc, <2 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
17 store <4 x i16> %tmp2, <4 x i16> addrspace(1)* %out, align 8
21 ; FUNC-LABEL: {{^}}scalar_to_vector_v2f32:
22 ; SI: buffer_load_dword [[VAL:v[0-9]+]],
23 ; SI: v_lshrrev_b32_e32 [[RESULT:v[0-9]+]], 16, [[VAL]]
24 ; SI: buffer_store_short [[RESULT]]
25 ; SI: buffer_store_short [[RESULT]]
26 ; SI: buffer_store_short [[RESULT]]
27 ; SI: buffer_store_short [[RESULT]]
29 define void @scalar_to_vector_v2f32(<4 x i16> addrspace(1)* %out, float addrspace(1)* %in) nounwind {
30 %tmp1 = load float, float addrspace(1)* %in, align 4
31 %bc = bitcast float %tmp1 to <2 x i16>
32 %tmp2 = shufflevector <2 x i16> %bc, <2 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
33 store <4 x i16> %tmp2, <4 x i16> addrspace(1)* %out, align 8
37 ; Getting a SCALAR_TO_VECTOR seems to be tricky. These cases managed
38 ; to produce one, but for some reason never made it to selection.
41 ; define void @scalar_to_vector_test2(<8 x i8> addrspace(1)* %out, i32 addrspace(1)* %in) nounwind {
42 ; %tmp1 = load i32, i32 addrspace(1)* %in, align 4
43 ; %bc = bitcast i32 %tmp1 to <4 x i8>
45 ; %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>
46 ; store <8 x i8> %tmp2, <8 x i8> addrspace(1)* %out, align 4
50 ; define void @scalar_to_vector_test3(<4 x i32> addrspace(1)* %out) nounwind {
51 ; %newvec0 = insertelement <2 x i64> undef, i64 12345, i32 0
52 ; %newvec1 = insertelement <2 x i64> %newvec0, i64 undef, i32 1
53 ; %bc = bitcast <2 x i64> %newvec1 to <4 x i32>
54 ; %add = add <4 x i32> %bc, <i32 1, i32 2, i32 3, i32 4>
55 ; store <4 x i32> %add, <4 x i32> addrspace(1)* %out, align 16
59 ; define void @scalar_to_vector_test4(<8 x i16> addrspace(1)* %out) nounwind {
60 ; %newvec0 = insertelement <4 x i32> undef, i32 12345, i32 0
61 ; %bc = bitcast <4 x i32> %newvec0 to <8 x i16>
62 ; %add = add <8 x i16> %bc, <i16 1, i16 2, i16 3, i16 4, i16 1, i16 2, i16 3, i16 4>
63 ; store <8 x i16> %add, <8 x i16> addrspace(1)* %out, align 16
67 ; define void @scalar_to_vector_test5(<4 x i16> addrspace(1)* %out) nounwind {
68 ; %newvec0 = insertelement <2 x i32> undef, i32 12345, i32 0
69 ; %bc = bitcast <2 x i32> %newvec0 to <4 x i16>
70 ; %add = add <4 x i16> %bc, <i16 1, i16 2, i16 3, i16 4>
71 ; store <4 x i16> %add, <4 x i16> addrspace(1)* %out, align 16
75 ; define void @scalar_to_vector_test6(<4 x i16> addrspace(1)* %out) nounwind {
76 ; %newvec0 = insertelement <2 x i32> undef, i32 12345, i32 0
77 ; %bc = bitcast <2 x i32> %newvec0 to <4 x i16>
78 ; %add = add <4 x i16> %bc, <i16 1, i16 2, i16 3, i16 4>
79 ; store <4 x i16> %add, <4 x i16> addrspace(1)* %out, align 16