1 ; RUN: llc -mtriple=arm-eabi -mattr=+neon %s -o - | FileCheck %s
3 define <8 x i8> @vhadds8(ptr %A, ptr %B) nounwind {
6 %tmp1 = load <8 x i8>, ptr %A
7 %tmp2 = load <8 x i8>, ptr %B
8 %tmp3 = call <8 x i8> @llvm.arm.neon.vhadds.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
12 define <4 x i16> @vhadds16(ptr %A, ptr %B) nounwind {
13 ;CHECK-LABEL: vhadds16:
15 %tmp1 = load <4 x i16>, ptr %A
16 %tmp2 = load <4 x i16>, ptr %B
17 %tmp3 = call <4 x i16> @llvm.arm.neon.vhadds.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
21 define <2 x i32> @vhadds32(ptr %A, ptr %B) nounwind {
22 ;CHECK-LABEL: vhadds32:
24 %tmp1 = load <2 x i32>, ptr %A
25 %tmp2 = load <2 x i32>, ptr %B
26 %tmp3 = call <2 x i32> @llvm.arm.neon.vhadds.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
30 define <8 x i8> @vhaddu8(ptr %A, ptr %B) nounwind {
31 ;CHECK-LABEL: vhaddu8:
33 %tmp1 = load <8 x i8>, ptr %A
34 %tmp2 = load <8 x i8>, ptr %B
35 %tmp3 = call <8 x i8> @llvm.arm.neon.vhaddu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
39 define <4 x i16> @vhaddu16(ptr %A, ptr %B) nounwind {
40 ;CHECK-LABEL: vhaddu16:
42 %tmp1 = load <4 x i16>, ptr %A
43 %tmp2 = load <4 x i16>, ptr %B
44 %tmp3 = call <4 x i16> @llvm.arm.neon.vhaddu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
48 define <2 x i32> @vhaddu32(ptr %A, ptr %B) nounwind {
49 ;CHECK-LABEL: vhaddu32:
51 %tmp1 = load <2 x i32>, ptr %A
52 %tmp2 = load <2 x i32>, ptr %B
53 %tmp3 = call <2 x i32> @llvm.arm.neon.vhaddu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
57 define <16 x i8> @vhaddQs8(ptr %A, ptr %B) nounwind {
58 ;CHECK-LABEL: vhaddQs8:
60 %tmp1 = load <16 x i8>, ptr %A
61 %tmp2 = load <16 x i8>, ptr %B
62 %tmp3 = call <16 x i8> @llvm.arm.neon.vhadds.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
66 define <8 x i16> @vhaddQs16(ptr %A, ptr %B) nounwind {
67 ;CHECK-LABEL: vhaddQs16:
69 %tmp1 = load <8 x i16>, ptr %A
70 %tmp2 = load <8 x i16>, ptr %B
71 %tmp3 = call <8 x i16> @llvm.arm.neon.vhadds.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
75 define <4 x i32> @vhaddQs32(ptr %A, ptr %B) nounwind {
76 ;CHECK-LABEL: vhaddQs32:
78 %tmp1 = load <4 x i32>, ptr %A
79 %tmp2 = load <4 x i32>, ptr %B
80 %tmp3 = call <4 x i32> @llvm.arm.neon.vhadds.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
84 define <16 x i8> @vhaddQu8(ptr %A, ptr %B) nounwind {
85 ;CHECK-LABEL: vhaddQu8:
87 %tmp1 = load <16 x i8>, ptr %A
88 %tmp2 = load <16 x i8>, ptr %B
89 %tmp3 = call <16 x i8> @llvm.arm.neon.vhaddu.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
93 define <8 x i16> @vhaddQu16(ptr %A, ptr %B) nounwind {
94 ;CHECK-LABEL: vhaddQu16:
96 %tmp1 = load <8 x i16>, ptr %A
97 %tmp2 = load <8 x i16>, ptr %B
98 %tmp3 = call <8 x i16> @llvm.arm.neon.vhaddu.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
102 define <4 x i32> @vhaddQu32(ptr %A, ptr %B) nounwind {
103 ;CHECK-LABEL: vhaddQu32:
105 %tmp1 = load <4 x i32>, ptr %A
106 %tmp2 = load <4 x i32>, ptr %B
107 %tmp3 = call <4 x i32> @llvm.arm.neon.vhaddu.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
111 declare <8 x i8> @llvm.arm.neon.vhadds.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
112 declare <4 x i16> @llvm.arm.neon.vhadds.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
113 declare <2 x i32> @llvm.arm.neon.vhadds.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
115 declare <8 x i8> @llvm.arm.neon.vhaddu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
116 declare <4 x i16> @llvm.arm.neon.vhaddu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
117 declare <2 x i32> @llvm.arm.neon.vhaddu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
119 declare <16 x i8> @llvm.arm.neon.vhadds.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
120 declare <8 x i16> @llvm.arm.neon.vhadds.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
121 declare <4 x i32> @llvm.arm.neon.vhadds.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
123 declare <16 x i8> @llvm.arm.neon.vhaddu.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
124 declare <8 x i16> @llvm.arm.neon.vhaddu.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
125 declare <4 x i32> @llvm.arm.neon.vhaddu.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
127 define <8 x i8> @vrhadds8(ptr %A, ptr %B) nounwind {
128 ;CHECK-LABEL: vrhadds8:
130 %tmp1 = load <8 x i8>, ptr %A
131 %tmp2 = load <8 x i8>, ptr %B
132 %tmp3 = call <8 x i8> @llvm.arm.neon.vrhadds.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
136 define <4 x i16> @vrhadds16(ptr %A, ptr %B) nounwind {
137 ;CHECK-LABEL: vrhadds16:
139 %tmp1 = load <4 x i16>, ptr %A
140 %tmp2 = load <4 x i16>, ptr %B
141 %tmp3 = call <4 x i16> @llvm.arm.neon.vrhadds.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
145 define <2 x i32> @vrhadds32(ptr %A, ptr %B) nounwind {
146 ;CHECK-LABEL: vrhadds32:
148 %tmp1 = load <2 x i32>, ptr %A
149 %tmp2 = load <2 x i32>, ptr %B
150 %tmp3 = call <2 x i32> @llvm.arm.neon.vrhadds.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
154 define <8 x i8> @vrhaddu8(ptr %A, ptr %B) nounwind {
155 ;CHECK-LABEL: vrhaddu8:
157 %tmp1 = load <8 x i8>, ptr %A
158 %tmp2 = load <8 x i8>, ptr %B
159 %tmp3 = call <8 x i8> @llvm.arm.neon.vrhaddu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
163 define <4 x i16> @vrhaddu16(ptr %A, ptr %B) nounwind {
164 ;CHECK-LABEL: vrhaddu16:
166 %tmp1 = load <4 x i16>, ptr %A
167 %tmp2 = load <4 x i16>, ptr %B
168 %tmp3 = call <4 x i16> @llvm.arm.neon.vrhaddu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
172 define <2 x i32> @vrhaddu32(ptr %A, ptr %B) nounwind {
173 ;CHECK-LABEL: vrhaddu32:
175 %tmp1 = load <2 x i32>, ptr %A
176 %tmp2 = load <2 x i32>, ptr %B
177 %tmp3 = call <2 x i32> @llvm.arm.neon.vrhaddu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
181 define <16 x i8> @vrhaddQs8(ptr %A, ptr %B) nounwind {
182 ;CHECK-LABEL: vrhaddQs8:
184 %tmp1 = load <16 x i8>, ptr %A
185 %tmp2 = load <16 x i8>, ptr %B
186 %tmp3 = call <16 x i8> @llvm.arm.neon.vrhadds.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
190 define <8 x i16> @vrhaddQs16(ptr %A, ptr %B) nounwind {
191 ;CHECK-LABEL: vrhaddQs16:
193 %tmp1 = load <8 x i16>, ptr %A
194 %tmp2 = load <8 x i16>, ptr %B
195 %tmp3 = call <8 x i16> @llvm.arm.neon.vrhadds.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
199 define <4 x i32> @vrhaddQs32(ptr %A, ptr %B) nounwind {
200 ;CHECK-LABEL: vrhaddQs32:
202 %tmp1 = load <4 x i32>, ptr %A
203 %tmp2 = load <4 x i32>, ptr %B
204 %tmp3 = call <4 x i32> @llvm.arm.neon.vrhadds.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
208 define <16 x i8> @vrhaddQu8(ptr %A, ptr %B) nounwind {
209 ;CHECK-LABEL: vrhaddQu8:
211 %tmp1 = load <16 x i8>, ptr %A
212 %tmp2 = load <16 x i8>, ptr %B
213 %tmp3 = call <16 x i8> @llvm.arm.neon.vrhaddu.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
217 define <8 x i16> @vrhaddQu16(ptr %A, ptr %B) nounwind {
218 ;CHECK-LABEL: vrhaddQu16:
220 %tmp1 = load <8 x i16>, ptr %A
221 %tmp2 = load <8 x i16>, ptr %B
222 %tmp3 = call <8 x i16> @llvm.arm.neon.vrhaddu.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
226 define <4 x i32> @vrhaddQu32(ptr %A, ptr %B) nounwind {
227 ;CHECK-LABEL: vrhaddQu32:
229 %tmp1 = load <4 x i32>, ptr %A
230 %tmp2 = load <4 x i32>, ptr %B
231 %tmp3 = call <4 x i32> @llvm.arm.neon.vrhaddu.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
235 declare <8 x i8> @llvm.arm.neon.vrhadds.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
236 declare <4 x i16> @llvm.arm.neon.vrhadds.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
237 declare <2 x i32> @llvm.arm.neon.vrhadds.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
239 declare <8 x i8> @llvm.arm.neon.vrhaddu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
240 declare <4 x i16> @llvm.arm.neon.vrhaddu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
241 declare <2 x i32> @llvm.arm.neon.vrhaddu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
243 declare <16 x i8> @llvm.arm.neon.vrhadds.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
244 declare <8 x i16> @llvm.arm.neon.vrhadds.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
245 declare <4 x i32> @llvm.arm.neon.vrhadds.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
247 declare <16 x i8> @llvm.arm.neon.vrhaddu.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
248 declare <8 x i16> @llvm.arm.neon.vrhaddu.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
249 declare <4 x i32> @llvm.arm.neon.vrhaddu.v4i32(<4 x i32>, <4 x i32>) nounwind readnone