1 ; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
3 %struct.__neon_int8x8x2_t = type { <8 x i8>, <8 x i8> }
4 %struct.__neon_int16x4x2_t = type { <4 x i16>, <4 x i16> }
5 %struct.__neon_int32x2x2_t = type { <2 x i32>, <2 x i32> }
6 %struct.__neon_float32x2x2_t = type { <2 x float>, <2 x float> }
8 %struct.__neon_int16x8x2_t = type { <8 x i16>, <8 x i16> }
9 %struct.__neon_int32x4x2_t = type { <4 x i32>, <4 x i32> }
10 %struct.__neon_float32x4x2_t = type { <4 x float>, <4 x float> }
12 define <8 x i8> @vld2lanei8(i8* %A, <8 x i8>* %B) nounwind {
15 %tmp1 = load <8 x i8>* %B
16 %tmp2 = call %struct.__neon_int8x8x2_t @llvm.arm.neon.vld2lane.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 1)
17 %tmp3 = extractvalue %struct.__neon_int8x8x2_t %tmp2, 0
18 %tmp4 = extractvalue %struct.__neon_int8x8x2_t %tmp2, 1
19 %tmp5 = add <8 x i8> %tmp3, %tmp4
23 define <4 x i16> @vld2lanei16(i16* %A, <4 x i16>* %B) nounwind {
26 %tmp1 = load <4 x i16>* %B
27 %tmp2 = call %struct.__neon_int16x4x2_t @llvm.arm.neon.vld2lane.v4i16(i16* %A, <4 x i16> %tmp1, <4 x i16> %tmp1, i32 1)
28 %tmp3 = extractvalue %struct.__neon_int16x4x2_t %tmp2, 0
29 %tmp4 = extractvalue %struct.__neon_int16x4x2_t %tmp2, 1
30 %tmp5 = add <4 x i16> %tmp3, %tmp4
34 define <2 x i32> @vld2lanei32(i32* %A, <2 x i32>* %B) nounwind {
37 %tmp1 = load <2 x i32>* %B
38 %tmp2 = call %struct.__neon_int32x2x2_t @llvm.arm.neon.vld2lane.v2i32(i32* %A, <2 x i32> %tmp1, <2 x i32> %tmp1, i32 1)
39 %tmp3 = extractvalue %struct.__neon_int32x2x2_t %tmp2, 0
40 %tmp4 = extractvalue %struct.__neon_int32x2x2_t %tmp2, 1
41 %tmp5 = add <2 x i32> %tmp3, %tmp4
45 define <2 x float> @vld2lanef(float* %A, <2 x float>* %B) nounwind {
48 %tmp1 = load <2 x float>* %B
49 %tmp2 = call %struct.__neon_float32x2x2_t @llvm.arm.neon.vld2lane.v2f32(float* %A, <2 x float> %tmp1, <2 x float> %tmp1, i32 1)
50 %tmp3 = extractvalue %struct.__neon_float32x2x2_t %tmp2, 0
51 %tmp4 = extractvalue %struct.__neon_float32x2x2_t %tmp2, 1
52 %tmp5 = add <2 x float> %tmp3, %tmp4
56 define <8 x i16> @vld2laneQi16(i16* %A, <8 x i16>* %B) nounwind {
59 %tmp1 = load <8 x i16>* %B
60 %tmp2 = call %struct.__neon_int16x8x2_t @llvm.arm.neon.vld2lane.v8i16(i16* %A, <8 x i16> %tmp1, <8 x i16> %tmp1, i32 1)
61 %tmp3 = extractvalue %struct.__neon_int16x8x2_t %tmp2, 0
62 %tmp4 = extractvalue %struct.__neon_int16x8x2_t %tmp2, 1
63 %tmp5 = add <8 x i16> %tmp3, %tmp4
67 define <4 x i32> @vld2laneQi32(i32* %A, <4 x i32>* %B) nounwind {
70 %tmp1 = load <4 x i32>* %B
71 %tmp2 = call %struct.__neon_int32x4x2_t @llvm.arm.neon.vld2lane.v4i32(i32* %A, <4 x i32> %tmp1, <4 x i32> %tmp1, i32 2)
72 %tmp3 = extractvalue %struct.__neon_int32x4x2_t %tmp2, 0
73 %tmp4 = extractvalue %struct.__neon_int32x4x2_t %tmp2, 1
74 %tmp5 = add <4 x i32> %tmp3, %tmp4
78 define <4 x float> @vld2laneQf(float* %A, <4 x float>* %B) nounwind {
81 %tmp1 = load <4 x float>* %B
82 %tmp2 = call %struct.__neon_float32x4x2_t @llvm.arm.neon.vld2lane.v4f32(float* %A, <4 x float> %tmp1, <4 x float> %tmp1, i32 1)
83 %tmp3 = extractvalue %struct.__neon_float32x4x2_t %tmp2, 0
84 %tmp4 = extractvalue %struct.__neon_float32x4x2_t %tmp2, 1
85 %tmp5 = add <4 x float> %tmp3, %tmp4
89 declare %struct.__neon_int8x8x2_t @llvm.arm.neon.vld2lane.v8i8(i8*, <8 x i8>, <8 x i8>, i32) nounwind readonly
90 declare %struct.__neon_int16x4x2_t @llvm.arm.neon.vld2lane.v4i16(i8*, <4 x i16>, <4 x i16>, i32) nounwind readonly
91 declare %struct.__neon_int32x2x2_t @llvm.arm.neon.vld2lane.v2i32(i8*, <2 x i32>, <2 x i32>, i32) nounwind readonly
92 declare %struct.__neon_float32x2x2_t @llvm.arm.neon.vld2lane.v2f32(i8*, <2 x float>, <2 x float>, i32) nounwind readonly
94 declare %struct.__neon_int16x8x2_t @llvm.arm.neon.vld2lane.v8i16(i8*, <8 x i16>, <8 x i16>, i32) nounwind readonly
95 declare %struct.__neon_int32x4x2_t @llvm.arm.neon.vld2lane.v4i32(i8*, <4 x i32>, <4 x i32>, i32) nounwind readonly
96 declare %struct.__neon_float32x4x2_t @llvm.arm.neon.vld2lane.v4f32(i8*, <4 x float>, <4 x float>, i32) nounwind readonly
98 %struct.__neon_int8x8x3_t = type { <8 x i8>, <8 x i8>, <8 x i8> }
99 %struct.__neon_int16x4x3_t = type { <4 x i16>, <4 x i16>, <4 x i16> }
100 %struct.__neon_int32x2x3_t = type { <2 x i32>, <2 x i32>, <2 x i32> }
101 %struct.__neon_float32x2x3_t = type { <2 x float>, <2 x float>, <2 x float> }
103 define <8 x i8> @vld3lanei8(i8* %A, <8 x i8>* %B) nounwind {
106 %tmp1 = load <8 x i8>* %B
107 %tmp2 = call %struct.__neon_int8x8x3_t @llvm.arm.neon.vld3lane.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 1)
108 %tmp3 = extractvalue %struct.__neon_int8x8x3_t %tmp2, 0
109 %tmp4 = extractvalue %struct.__neon_int8x8x3_t %tmp2, 1
110 %tmp5 = extractvalue %struct.__neon_int8x8x3_t %tmp2, 2
111 %tmp6 = add <8 x i8> %tmp3, %tmp4
112 %tmp7 = add <8 x i8> %tmp5, %tmp6
116 define <4 x i16> @vld3lanei16(i16* %A, <4 x i16>* %B) nounwind {
119 %tmp1 = load <4 x i16>* %B
120 %tmp2 = call %struct.__neon_int16x4x3_t @llvm.arm.neon.vld3lane.v4i16(i16* %A, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, i32 1)
121 %tmp3 = extractvalue %struct.__neon_int16x4x3_t %tmp2, 0
122 %tmp4 = extractvalue %struct.__neon_int16x4x3_t %tmp2, 1
123 %tmp5 = extractvalue %struct.__neon_int16x4x3_t %tmp2, 2
124 %tmp6 = add <4 x i16> %tmp3, %tmp4
125 %tmp7 = add <4 x i16> %tmp5, %tmp6
129 define <2 x i32> @vld3lanei32(i32* %A, <2 x i32>* %B) nounwind {
132 %tmp1 = load <2 x i32>* %B
133 %tmp2 = call %struct.__neon_int32x2x3_t @llvm.arm.neon.vld3lane.v2i32(i32* %A, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, i32 1)
134 %tmp3 = extractvalue %struct.__neon_int32x2x3_t %tmp2, 0
135 %tmp4 = extractvalue %struct.__neon_int32x2x3_t %tmp2, 1
136 %tmp5 = extractvalue %struct.__neon_int32x2x3_t %tmp2, 2
137 %tmp6 = add <2 x i32> %tmp3, %tmp4
138 %tmp7 = add <2 x i32> %tmp5, %tmp6
142 define <2 x float> @vld3lanef(float* %A, <2 x float>* %B) nounwind {
145 %tmp1 = load <2 x float>* %B
146 %tmp2 = call %struct.__neon_float32x2x3_t @llvm.arm.neon.vld3lane.v2f32(float* %A, <2 x float> %tmp1, <2 x float> %tmp1, <2 x float> %tmp1, i32 1)
147 %tmp3 = extractvalue %struct.__neon_float32x2x3_t %tmp2, 0
148 %tmp4 = extractvalue %struct.__neon_float32x2x3_t %tmp2, 1
149 %tmp5 = extractvalue %struct.__neon_float32x2x3_t %tmp2, 2
150 %tmp6 = add <2 x float> %tmp3, %tmp4
151 %tmp7 = add <2 x float> %tmp5, %tmp6
152 ret <2 x float> %tmp7
155 declare %struct.__neon_int8x8x3_t @llvm.arm.neon.vld3lane.v8i8(i8*, <8 x i8>, <8 x i8>, <8 x i8>, i32) nounwind readonly
156 declare %struct.__neon_int16x4x3_t @llvm.arm.neon.vld3lane.v4i16(i8*, <4 x i16>, <4 x i16>, <4 x i16>, i32) nounwind readonly
157 declare %struct.__neon_int32x2x3_t @llvm.arm.neon.vld3lane.v2i32(i8*, <2 x i32>, <2 x i32>, <2 x i32>, i32) nounwind readonly
158 declare %struct.__neon_float32x2x3_t @llvm.arm.neon.vld3lane.v2f32(i8*, <2 x float>, <2 x float>, <2 x float>, i32) nounwind readonly
160 %struct.__neon_int8x8x4_t = type { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> }
161 %struct.__neon_int16x4x4_t = type { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }
162 %struct.__neon_int32x2x4_t = type { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> }
163 %struct.__neon_float32x2x4_t = type { <2 x float>, <2 x float>, <2 x float>, <2 x float> }
165 define <8 x i8> @vld4lanei8(i8* %A, <8 x i8>* %B) nounwind {
168 %tmp1 = load <8 x i8>* %B
169 %tmp2 = call %struct.__neon_int8x8x4_t @llvm.arm.neon.vld4lane.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 1)
170 %tmp3 = extractvalue %struct.__neon_int8x8x4_t %tmp2, 0
171 %tmp4 = extractvalue %struct.__neon_int8x8x4_t %tmp2, 1
172 %tmp5 = extractvalue %struct.__neon_int8x8x4_t %tmp2, 2
173 %tmp6 = extractvalue %struct.__neon_int8x8x4_t %tmp2, 3
174 %tmp7 = add <8 x i8> %tmp3, %tmp4
175 %tmp8 = add <8 x i8> %tmp5, %tmp6
176 %tmp9 = add <8 x i8> %tmp7, %tmp8
180 define <4 x i16> @vld4lanei16(i16* %A, <4 x i16>* %B) nounwind {
183 %tmp1 = load <4 x i16>* %B
184 %tmp2 = call %struct.__neon_int16x4x4_t @llvm.arm.neon.vld4lane.v4i16(i16* %A, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, i32 1)
185 %tmp3 = extractvalue %struct.__neon_int16x4x4_t %tmp2, 0
186 %tmp4 = extractvalue %struct.__neon_int16x4x4_t %tmp2, 1
187 %tmp5 = extractvalue %struct.__neon_int16x4x4_t %tmp2, 2
188 %tmp6 = extractvalue %struct.__neon_int16x4x4_t %tmp2, 3
189 %tmp7 = add <4 x i16> %tmp3, %tmp4
190 %tmp8 = add <4 x i16> %tmp5, %tmp6
191 %tmp9 = add <4 x i16> %tmp7, %tmp8
195 define <2 x i32> @vld4lanei32(i32* %A, <2 x i32>* %B) nounwind {
198 %tmp1 = load <2 x i32>* %B
199 %tmp2 = call %struct.__neon_int32x2x4_t @llvm.arm.neon.vld4lane.v2i32(i32* %A, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, i32 1)
200 %tmp3 = extractvalue %struct.__neon_int32x2x4_t %tmp2, 0
201 %tmp4 = extractvalue %struct.__neon_int32x2x4_t %tmp2, 1
202 %tmp5 = extractvalue %struct.__neon_int32x2x4_t %tmp2, 2
203 %tmp6 = extractvalue %struct.__neon_int32x2x4_t %tmp2, 3
204 %tmp7 = add <2 x i32> %tmp3, %tmp4
205 %tmp8 = add <2 x i32> %tmp5, %tmp6
206 %tmp9 = add <2 x i32> %tmp7, %tmp8
210 define <2 x float> @vld4lanef(float* %A, <2 x float>* %B) nounwind {
213 %tmp1 = load <2 x float>* %B
214 %tmp2 = call %struct.__neon_float32x2x4_t @llvm.arm.neon.vld4lane.v2f32(float* %A, <2 x float> %tmp1, <2 x float> %tmp1, <2 x float> %tmp1, <2 x float> %tmp1, i32 1)
215 %tmp3 = extractvalue %struct.__neon_float32x2x4_t %tmp2, 0
216 %tmp4 = extractvalue %struct.__neon_float32x2x4_t %tmp2, 1
217 %tmp5 = extractvalue %struct.__neon_float32x2x4_t %tmp2, 2
218 %tmp6 = extractvalue %struct.__neon_float32x2x4_t %tmp2, 3
219 %tmp7 = add <2 x float> %tmp3, %tmp4
220 %tmp8 = add <2 x float> %tmp5, %tmp6
221 %tmp9 = add <2 x float> %tmp7, %tmp8
222 ret <2 x float> %tmp9
225 declare %struct.__neon_int8x8x4_t @llvm.arm.neon.vld4lane.v8i8(i8*, <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8>, i32) nounwind readonly
226 declare %struct.__neon_int16x4x4_t @llvm.arm.neon.vld4lane.v4i16(i8*, <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16>, i32) nounwind readonly
227 declare %struct.__neon_int32x2x4_t @llvm.arm.neon.vld4lane.v2i32(i8*, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, i32) nounwind readonly
228 declare %struct.__neon_float32x2x4_t @llvm.arm.neon.vld4lane.v2f32(i8*, <2 x float>, <2 x float>, <2 x float>, <2 x float>, i32) nounwind readonly