1 ; RUN: llc -mtriple=arm-eabi -mattr=+neon %s -o - | FileCheck %s
3 define <8 x i8> @vsli8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
6 %tmp1 = load <8 x i8>, <8 x i8>* %A
7 %tmp2 = load <8 x i8>, <8 x i8>* %B
8 %tmp3 = call <8 x i8> @llvm.arm.neon.vshiftins.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2, <8 x i8> < i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7 >)
12 define <4 x i16> @vsli16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
15 %tmp1 = load <4 x i16>, <4 x i16>* %A
16 %tmp2 = load <4 x i16>, <4 x i16>* %B
17 %tmp3 = call <4 x i16> @llvm.arm.neon.vshiftins.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2, <4 x i16> < i16 15, i16 15, i16 15, i16 15 >)
21 define <2 x i32> @vsli32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
24 %tmp1 = load <2 x i32>, <2 x i32>* %A
25 %tmp2 = load <2 x i32>, <2 x i32>* %B
26 %tmp3 = call <2 x i32> @llvm.arm.neon.vshiftins.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2, <2 x i32> < i32 31, i32 31 >)
30 define <1 x i64> @vsli64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
33 %tmp1 = load <1 x i64>, <1 x i64>* %A
34 %tmp2 = load <1 x i64>, <1 x i64>* %B
35 %tmp3 = call <1 x i64> @llvm.arm.neon.vshiftins.v1i64(<1 x i64> %tmp1, <1 x i64> %tmp2, <1 x i64> < i64 63 >)
39 define <16 x i8> @vsliQ8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
42 %tmp1 = load <16 x i8>, <16 x i8>* %A
43 %tmp2 = load <16 x i8>, <16 x i8>* %B
44 %tmp3 = call <16 x i8> @llvm.arm.neon.vshiftins.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2, <16 x i8> < i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7 >)
48 define <8 x i16> @vsliQ16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
49 ;CHECK-LABEL: vsliQ16:
51 %tmp1 = load <8 x i16>, <8 x i16>* %A
52 %tmp2 = load <8 x i16>, <8 x i16>* %B
53 %tmp3 = call <8 x i16> @llvm.arm.neon.vshiftins.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2, <8 x i16> < i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15 >)
57 define <4 x i32> @vsliQ32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
58 ;CHECK-LABEL: vsliQ32:
60 %tmp1 = load <4 x i32>, <4 x i32>* %A
61 %tmp2 = load <4 x i32>, <4 x i32>* %B
62 %tmp3 = call <4 x i32> @llvm.arm.neon.vshiftins.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2, <4 x i32> < i32 31, i32 31, i32 31, i32 31 >)
66 define <2 x i64> @vsliQ64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
67 ;CHECK-LABEL: vsliQ64:
69 %tmp1 = load <2 x i64>, <2 x i64>* %A
70 %tmp2 = load <2 x i64>, <2 x i64>* %B
71 %tmp3 = call <2 x i64> @llvm.arm.neon.vshiftins.v2i64(<2 x i64> %tmp1, <2 x i64> %tmp2, <2 x i64> < i64 63, i64 63 >)
75 define <8 x i8> @vsri8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
78 %tmp1 = load <8 x i8>, <8 x i8>* %A
79 %tmp2 = load <8 x i8>, <8 x i8>* %B
80 %tmp3 = call <8 x i8> @llvm.arm.neon.vshiftins.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2, <8 x i8> < i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8 >)
84 define <4 x i16> @vsri16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
87 %tmp1 = load <4 x i16>, <4 x i16>* %A
88 %tmp2 = load <4 x i16>, <4 x i16>* %B
89 %tmp3 = call <4 x i16> @llvm.arm.neon.vshiftins.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2, <4 x i16> < i16 -16, i16 -16, i16 -16, i16 -16 >)
93 define <2 x i32> @vsri32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
96 %tmp1 = load <2 x i32>, <2 x i32>* %A
97 %tmp2 = load <2 x i32>, <2 x i32>* %B
98 %tmp3 = call <2 x i32> @llvm.arm.neon.vshiftins.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2, <2 x i32> < i32 -32, i32 -32 >)
102 define <1 x i64> @vsri64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
103 ;CHECK-LABEL: vsri64:
105 %tmp1 = load <1 x i64>, <1 x i64>* %A
106 %tmp2 = load <1 x i64>, <1 x i64>* %B
107 %tmp3 = call <1 x i64> @llvm.arm.neon.vshiftins.v1i64(<1 x i64> %tmp1, <1 x i64> %tmp2, <1 x i64> < i64 -64 >)
111 define <16 x i8> @vsriQ8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
112 ;CHECK-LABEL: vsriQ8:
114 %tmp1 = load <16 x i8>, <16 x i8>* %A
115 %tmp2 = load <16 x i8>, <16 x i8>* %B
116 %tmp3 = call <16 x i8> @llvm.arm.neon.vshiftins.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2, <16 x i8> < i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8 >)
120 define <8 x i16> @vsriQ16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
121 ;CHECK-LABEL: vsriQ16:
123 %tmp1 = load <8 x i16>, <8 x i16>* %A
124 %tmp2 = load <8 x i16>, <8 x i16>* %B
125 %tmp3 = call <8 x i16> @llvm.arm.neon.vshiftins.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2, <8 x i16> < i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16 >)
129 define <4 x i32> @vsriQ32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
130 ;CHECK-LABEL: vsriQ32:
132 %tmp1 = load <4 x i32>, <4 x i32>* %A
133 %tmp2 = load <4 x i32>, <4 x i32>* %B
134 %tmp3 = call <4 x i32> @llvm.arm.neon.vshiftins.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2, <4 x i32> < i32 -32, i32 -32, i32 -32, i32 -32 >)
138 define <2 x i64> @vsriQ64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
139 ;CHECK-LABEL: vsriQ64:
141 %tmp1 = load <2 x i64>, <2 x i64>* %A
142 %tmp2 = load <2 x i64>, <2 x i64>* %B
143 %tmp3 = call <2 x i64> @llvm.arm.neon.vshiftins.v2i64(<2 x i64> %tmp1, <2 x i64> %tmp2, <2 x i64> < i64 -64, i64 -64 >)
147 declare <8 x i8> @llvm.arm.neon.vshiftins.v8i8(<8 x i8>, <8 x i8>, <8 x i8>) nounwind readnone
148 declare <4 x i16> @llvm.arm.neon.vshiftins.v4i16(<4 x i16>, <4 x i16>, <4 x i16>) nounwind readnone
149 declare <2 x i32> @llvm.arm.neon.vshiftins.v2i32(<2 x i32>, <2 x i32>, <2 x i32>) nounwind readnone
150 declare <1 x i64> @llvm.arm.neon.vshiftins.v1i64(<1 x i64>, <1 x i64>, <1 x i64>) nounwind readnone
152 declare <16 x i8> @llvm.arm.neon.vshiftins.v16i8(<16 x i8>, <16 x i8>, <16 x i8>) nounwind readnone
153 declare <8 x i16> @llvm.arm.neon.vshiftins.v8i16(<8 x i16>, <8 x i16>, <8 x i16>) nounwind readnone
154 declare <4 x i32> @llvm.arm.neon.vshiftins.v4i32(<4 x i32>, <4 x i32>, <4 x i32>) nounwind readnone
155 declare <2 x i64> @llvm.arm.neon.vshiftins.v2i64(<2 x i64>, <2 x i64>, <2 x i64>) nounwind readnone