1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
|
; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
define <8 x i8> @vqshrns8(<8 x i16>* %A) nounwind {
;CHECK: vqshrns8:
;CHECK: vqshrn.s16
%tmp1 = load <8 x i16>* %A
%tmp2 = call <8 x i8> @llvm.arm.neon.vqshiftns.v8i8(<8 x i16> %tmp1, <8 x i16> < i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8 >)
ret <8 x i8> %tmp2
}
define <4 x i16> @vqshrns16(<4 x i32>* %A) nounwind {
;CHECK: vqshrns16:
;CHECK: vqshrn.s32
%tmp1 = load <4 x i32>* %A
%tmp2 = call <4 x i16> @llvm.arm.neon.vqshiftns.v4i16(<4 x i32> %tmp1, <4 x i32> < i32 -16, i32 -16, i32 -16, i32 -16 >)
ret <4 x i16> %tmp2
}
define <2 x i32> @vqshrns32(<2 x i64>* %A) nounwind {
;CHECK: vqshrns32:
;CHECK: vqshrn.s64
%tmp1 = load <2 x i64>* %A
%tmp2 = call <2 x i32> @llvm.arm.neon.vqshiftns.v2i32(<2 x i64> %tmp1, <2 x i64> < i64 -32, i64 -32 >)
ret <2 x i32> %tmp2
}
define <8 x i8> @vqshrnu8(<8 x i16>* %A) nounwind {
;CHECK: vqshrnu8:
;CHECK: vqshrn.u16
%tmp1 = load <8 x i16>* %A
%tmp2 = call <8 x i8> @llvm.arm.neon.vqshiftnu.v8i8(<8 x i16> %tmp1, <8 x i16> < i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8 >)
ret <8 x i8> %tmp2
}
define <4 x i16> @vqshrnu16(<4 x i32>* %A) nounwind {
;CHECK: vqshrnu16:
;CHECK: vqshrn.u32
%tmp1 = load <4 x i32>* %A
%tmp2 = call <4 x i16> @llvm.arm.neon.vqshiftnu.v4i16(<4 x i32> %tmp1, <4 x i32> < i32 -16, i32 -16, i32 -16, i32 -16 >)
ret <4 x i16> %tmp2
}
define <2 x i32> @vqshrnu32(<2 x i64>* %A) nounwind {
;CHECK: vqshrnu32:
;CHECK: vqshrn.u64
%tmp1 = load <2 x i64>* %A
%tmp2 = call <2 x i32> @llvm.arm.neon.vqshiftnu.v2i32(<2 x i64> %tmp1, <2 x i64> < i64 -32, i64 -32 >)
ret <2 x i32> %tmp2
}
define <8 x i8> @vqshruns8(<8 x i16>* %A) nounwind {
;CHECK: vqshruns8:
;CHECK: vqshrun.s16
%tmp1 = load <8 x i16>* %A
%tmp2 = call <8 x i8> @llvm.arm.neon.vqshiftnsu.v8i8(<8 x i16> %tmp1, <8 x i16> < i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8 >)
ret <8 x i8> %tmp2
}
define <4 x i16> @vqshruns16(<4 x i32>* %A) nounwind {
;CHECK: vqshruns16:
;CHECK: vqshrun.s32
%tmp1 = load <4 x i32>* %A
%tmp2 = call <4 x i16> @llvm.arm.neon.vqshiftnsu.v4i16(<4 x i32> %tmp1, <4 x i32> < i32 -16, i32 -16, i32 -16, i32 -16 >)
ret <4 x i16> %tmp2
}
define <2 x i32> @vqshruns32(<2 x i64>* %A) nounwind {
;CHECK: vqshruns32:
;CHECK: vqshrun.s64
%tmp1 = load <2 x i64>* %A
%tmp2 = call <2 x i32> @llvm.arm.neon.vqshiftnsu.v2i32(<2 x i64> %tmp1, <2 x i64> < i64 -32, i64 -32 >)
ret <2 x i32> %tmp2
}
declare <8 x i8> @llvm.arm.neon.vqshiftns.v8i8(<8 x i16>, <8 x i16>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vqshiftns.v4i16(<4 x i32>, <4 x i32>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vqshiftns.v2i32(<2 x i64>, <2 x i64>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vqshiftnu.v8i8(<8 x i16>, <8 x i16>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vqshiftnu.v4i16(<4 x i32>, <4 x i32>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vqshiftnu.v2i32(<2 x i64>, <2 x i64>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vqshiftnsu.v8i8(<8 x i16>, <8 x i16>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vqshiftnsu.v4i16(<4 x i32>, <4 x i32>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vqshiftnsu.v2i32(<2 x i64>, <2 x i64>) nounwind readnone
define <8 x i8> @vqrshrns8(<8 x i16>* %A) nounwind {
;CHECK: vqrshrns8:
;CHECK: vqrshrn.s16
%tmp1 = load <8 x i16>* %A
%tmp2 = call <8 x i8> @llvm.arm.neon.vqrshiftns.v8i8(<8 x i16> %tmp1, <8 x i16> < i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8 >)
ret <8 x i8> %tmp2
}
define <4 x i16> @vqrshrns16(<4 x i32>* %A) nounwind {
;CHECK: vqrshrns16:
;CHECK: vqrshrn.s32
%tmp1 = load <4 x i32>* %A
%tmp2 = call <4 x i16> @llvm.arm.neon.vqrshiftns.v4i16(<4 x i32> %tmp1, <4 x i32> < i32 -16, i32 -16, i32 -16, i32 -16 >)
ret <4 x i16> %tmp2
}
define <2 x i32> @vqrshrns32(<2 x i64>* %A) nounwind {
;CHECK: vqrshrns32:
;CHECK: vqrshrn.s64
%tmp1 = load <2 x i64>* %A
%tmp2 = call <2 x i32> @llvm.arm.neon.vqrshiftns.v2i32(<2 x i64> %tmp1, <2 x i64> < i64 -32, i64 -32 >)
ret <2 x i32> %tmp2
}
define <8 x i8> @vqrshrnu8(<8 x i16>* %A) nounwind {
;CHECK: vqrshrnu8:
;CHECK: vqrshrn.u16
%tmp1 = load <8 x i16>* %A
%tmp2 = call <8 x i8> @llvm.arm.neon.vqrshiftnu.v8i8(<8 x i16> %tmp1, <8 x i16> < i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8 >)
ret <8 x i8> %tmp2
}
define <4 x i16> @vqrshrnu16(<4 x i32>* %A) nounwind {
;CHECK: vqrshrnu16:
;CHECK: vqrshrn.u32
%tmp1 = load <4 x i32>* %A
%tmp2 = call <4 x i16> @llvm.arm.neon.vqrshiftnu.v4i16(<4 x i32> %tmp1, <4 x i32> < i32 -16, i32 -16, i32 -16, i32 -16 >)
ret <4 x i16> %tmp2
}
define <2 x i32> @vqrshrnu32(<2 x i64>* %A) nounwind {
|
