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
| ; RUN: llc < %s -mtriple=armv8 -mattr=+v8.1a | FileCheck %s
;-----------------------------------------------------------------------------
; RDMA Vector
declare <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16>, <4 x i16>)
declare <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16>, <8 x i16>)
declare <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32>, <2 x i32>)
declare <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32>, <4 x i32>)
declare <4 x i16> @llvm.arm.neon.vqadds.v4i16(<4 x i16>, <4 x i16>)
declare <8 x i16> @llvm.arm.neon.vqadds.v8i16(<8 x i16>, <8 x i16>)
declare <2 x i32> @llvm.arm.neon.vqadds.v2i32(<2 x i32>, <2 x i32>)
declare <4 x i32> @llvm.arm.neon.vqadds.v4i32(<4 x i32>, <4 x i32>)
declare <4 x i16> @llvm.arm.neon.vqsubs.v4i16(<4 x i16>, <4 x i16>)
declare <8 x i16> @llvm.arm.neon.vqsubs.v8i16(<8 x i16>, <8 x i16>)
declare <2 x i32> @llvm.arm.neon.vqsubs.v2i32(<2 x i32>, <2 x i32>)
declare <4 x i32> @llvm.arm.neon.vqsubs.v4i32(<4 x i32>, <4 x i32>)
define <4 x i16> @test_vqrdmlah_v4i16(<4 x i16> %acc, <4 x i16> %mhs, <4 x i16> %rhs) {
; CHECK-LABEL: test_vqrdmlah_v4i16:
%prod = call <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16> %mhs, <4 x i16> %rhs)
%retval = call <4 x i16> @llvm.arm.neon.vqadds.v4i16(<4 x i16> %acc, <4 x i16> %prod)
; CHECK: vqrdmlah.s16 {{d[0-9]+}}, {{d[0-9]+}}, {{d[0-9]+}}
ret <4 x i16> %retval
}
define <8 x i16> @test_vqrdmlah_v8i16(<8 x i16> %acc, <8 x i16> %mhs, <8 x i16> %rhs) {
; CHECK-LABEL: test_vqrdmlah_v8i16:
%prod = call <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16> %mhs, <8 x i16> %rhs)
%retval = call <8 x i16> @llvm.arm.neon.vqadds.v8i16(<8 x i16> %acc, <8 x i16> %prod)
; CHECK: vqrdmlah.s16 {{q[0-9]+}}, {{q[0-9]+}}, {{q[0-9]+}}
ret <8 x i16> %retval
}
define <2 x i32> @test_vqrdmlah_v2i32(<2 x i32> %acc, <2 x i32> %mhs, <2 x i32> %rhs) {
; CHECK-LABEL: test_vqrdmlah_v2i32:
%prod = call <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32> %mhs, <2 x i32> %rhs)
%retval = call <2 x i32> @llvm.arm.neon.vqadds.v2i32(<2 x i32> %acc, <2 x i32> %prod)
; CHECK: vqrdmlah.s32 {{d[0-9]+}}, {{d[0-9]+}}, {{d[0-9]+}}
ret <2 x i32> %retval
}
define <4 x i32> @test_vqrdmlah_v4i32(<4 x i32> %acc, <4 x i32> %mhs, <4 x i32> %rhs) {
; CHECK-LABEL: test_vqrdmlah_v4i32:
%prod = call <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32> %mhs, <4 x i32> %rhs)
%retval = call <4 x i32> @llvm.arm.neon.vqadds.v4i32(<4 x i32> %acc, <4 x i32> %prod)
; CHECK: vqrdmlah.s32 {{q[0-9]+}}, {{q[0-9]+}}, {{q[0-9]+}}
ret <4 x i32> %retval
}
define <4 x i16> @test_vqrdmlsh_v4i16(<4 x i16> %acc, <4 x i16> %mhs, <4 x i16> %rhs) {
; CHECK-LABEL: test_vqrdmlsh_v4i16:
%prod = call <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16> %mhs, <4 x i16> %rhs)
%retval = call <4 x i16> @llvm.arm.neon.vqsubs.v4i16(<4 x i16> %acc, <4 x i16> %prod)
; CHECK: vqrdmlsh.s16 {{d[0-9]+}}, {{d[0-9]+}}, {{d[0-9]+}}
ret <4 x i16> %retval
}
define <8 x i16> @test_vqrdmlsh_v8i16(<8 x i16> %acc, <8 x i16> %mhs, <8 x i16> %rhs) {
; CHECK-LABEL: test_vqrdmlsh_v8i16:
%prod = call <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16> %mhs, <8 x i16> %rhs)
%retval = call <8 x i16> @llvm.arm.neon.vqsubs.v8i16(<8 x i16> %acc, <8 x i16> %prod)
; CHECK: vqrdmlsh.s16 {{q[0-9]+}}, {{q[0-9]+}}, {{q[0-9]+}}
ret <8 x i16> %retval
}
define <2 x i32> @test_vqrdmlsh_v2i32(<2 x i32> %acc, <2 x i32> %mhs, <2 x i32> %rhs) {
; CHECK-LABEL: test_vqrdmlsh_v2i32:
%prod = call <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32> %mhs, <2 x i32> %rhs)
%retval = call <2 x i32> @llvm.arm.neon.vqsubs.v2i32(<2 x i32> %acc, <2 x i32> %prod)
; CHECK: vqrdmlsh.s32 {{d[0-9]+}}, {{d[0-9]+}}, {{d[0-9]+}}
ret <2 x i32> %retval
}
define <4 x i32> @test_vqrdmlsh_v4i32(<4 x i32> %acc, <4 x i32> %mhs, <4 x i32> %rhs) {
; CHECK-LABEL: test_vqrdmlsh_v4i32:
%prod = call <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32> %mhs, <4 x i32> %rhs)
%retval = call <4 x i32> @llvm.arm.neon.vqsubs.v4i32(<4 x i32> %acc, <4 x i32> %prod)
; CHECK: vqrdmlsh.s32 {{q[0-9]+}}, {{q[0-9]+}}, {{q[0-9]+}}
ret <4 x i32> %retval
}
;-----------------------------------------------------------------------------
; RDMA Scalar
define <4 x i16> @test_vqrdmlah_lane_s16(<4 x i16> %acc, <4 x i16> %x, <4 x i16> %v) {
; CHECK-LABEL: test_vqrdmlah_lane_s16:
entry:
%shuffle = shufflevector <4 x i16> %v, <4 x i16> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
%prod = call <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16> %x, <4 x i16> %shuffle)
%retval = call <4 x i16> @llvm.arm.neon.vqadds.v4i16(<4 x i16> %acc, <4 x i16> %prod)
; CHECK: vqrdmlah.s16 {{d[0-9]+}}, {{d[0-9]+}}, {{d[0-9]+}}[3]
ret <4 x i16> %retval
}
define <8 x i16> @test_vqrdmlahq_lane_s16(<8 x i16> %acc, <8 x i16> %x, <4 x i16> %v) {
; CHECK-LABEL: test_vqrdmlahq_lane_s16:
entry:
%shuffle = shufflevector <4 x i16> %v, <4 x i16> undef, <8 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
%prod = call <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16> %x, <8 x i16> %shuffle)
%retval = call <8 x i16> @llvm.arm.neon.vqadds.v8i16(<8 x i16> %acc, <8 x i16> %prod)
; CHECK: vqrdmlah.s16 {{q[0-9]+}}, {{q[0-9]+}}, {{d[0-9]+}}[2]
ret <8 x i16> %retval
}
define <2 x i32> @test_vqrdmlah_lane_s32(<2 x i32> %acc, <2 x i32> %x, <2 x i32> %v) {
; CHECK-LABEL: test_vqrdmlah_lane_s32:
entry:
%shuffle = shufflevector <2 x i32> %v, <2 x i32> undef, <2 x i32> <i32 1, i32 1>
%prod = tail call <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32> %x, <2 x i32> %shuffle)
%retval = call <2 x i32> @llvm.arm.neon.vqadds.v2i32(<2 x i32> %acc, <2 x i32> %prod)
; CHECK: vqrdmlah.s32 {{d[0-9]+}}, {{d[0-9]+}}, {{d[0-9]+}}[1]
ret <2 x i32> %retval
}
define <4 x i32> @test_vqrdmlahq_lane_s32(<4 x i32> %acc,<4 x i32> %x, <2 x i32> %v) {
; CHECK-LABEL: test_vqrdmlahq_lane_s32:
entry:
%shuffle = shufflevector <2 x i32> %v, <2 x i32> undef, <4 x i32> zeroinitializer
%prod = tail call <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32> %x, <4 x i32> %shuffle)
%retval = call <4 x i32> @llvm.arm.neon.vqadds.v4i32(<4 x i32> %acc, <4 x i32> %prod)
; CHECK: vqrdmlah.s32 {{q[0-9]+}}, {{q[0-9]+}}, {{d[0-9]+}}[0]
ret <4 x i32> %retval
}
define <4 x i16> @test_vqrdmlsh_lane_s16(<4 x i16> %acc, <4 x i16> %x, <4 x i16> %v) {
; CHECK-LABEL: test_vqrdmlsh_lane_s16:
entry:
%shuffle = shufflevector <4 x i16> %v, <4 x i16> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
%prod = call <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16> %x, <4 x i16> %shuffle)
%retval = call <4 x i16> @llvm.arm.neon.vqsubs.v4i16(<4 x i16> %acc, <4 x i16> %prod)
; CHECK: vqrdmlsh.s16 {{d[0-9]+}}, {{d[0-9]+}}, {{d[0-9]+}}[3]
ret <4 x i16> %retval
}
define <8 x i16> @test_vqrdmlshq_lane_s16(<8 x i16> %acc, <8 x i16> %x, <4 x i16> %v) {
; CHECK-LABEL: test_vqrdmlshq_lane_s16:
entry:
%shuffle = shufflevector <4 x i16> %v, <4 x i16> undef, <8 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
%prod = call <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16> %x, <8 x i16> %shuffle)
%retval = call <8 x i16> @llvm.arm.neon.vqsubs.v8i16(<8 x i16> %acc, <8 x i16> %prod)
; CHECK: vqrdmlsh.s16 {{q[0-9]+}}, {{q[0-9]+}}, {{d[0-9]+}}[2]
ret <8 x i16> %retval
}
define <2 x i32> @test_vqrdmlsh_lane_s32(<2 x i32> %acc, <2 x i32> %x, <2 x i32> %v) {
; CHECK-LABEL: test_vqrdmlsh_lane_s32:
entry:
%shuffle = shufflevector <2 x i32> %v, <2 x i32> undef, <2 x i32> <i32 1, i32 1>
%prod = tail call <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32> %x, <2 x i32> %shuffle)
%retval = call <2 x i32> @llvm.arm.neon.vqsubs.v2i32(<2 x i32> %acc, <2 x i32> %prod)
; CHECK: vqrdmlsh.s32 {{d[0-9]+}}, {{d[0-9]+}}, {{d[0-9]+}}[1]
ret <2 x i32> %retval
}
define <4 x i32> @test_vqrdmlshq_lane_s32(<4 x i32> %acc,<4 x i32> %x, <2 x i32> %v) {
; CHECK-LABEL: test_vqrdmlshq_lane_s32:
entry:
%shuffle = shufflevector <2 x i32> %v, <2 x i32> undef, <4 x i32> zeroinitializer
%prod = tail call <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32> %x, <4 x i32> %shuffle)
%retval = call <4 x i32> @llvm.arm.neon.vqsubs.v4i32(<4 x i32> %acc, <4 x i32> %prod)
; CHECK: vqrdmlsh.s32 {{q[0-9]+}}, {{q[0-9]+}}, {{d[0-9]+}}[0]
ret <4 x i32> %retval
}
|