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
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mattr=+sve -force-streaming-compatible < %s | FileCheck %s
; RUN: llc -mattr=+sme -force-streaming < %s | FileCheck %s
; RUN: llc -force-streaming-compatible < %s | FileCheck %s --check-prefix=NONEON-NOSVE
target triple = "aarch64-unknown-linux-gnu"
define <4 x i32> @test(ptr %arg1, ptr %arg2) {
; CHECK-LABEL: test:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: ldp q0, q1, [x0, #32]
; CHECK-NEXT: ldp q3, q4, [x0]
; CHECK-NEXT: add z2.s, z0.s, z0.s
; CHECK-NEXT: add z5.s, z1.s, z1.s
; CHECK-NEXT: mov z0.s, z1.s[2]
; CHECK-NEXT: add z1.s, z3.s, z3.s
; CHECK-NEXT: add z3.s, z4.s, z4.s
; CHECK-NEXT: // kill: def $q0 killed $q0 killed $z0
; CHECK-NEXT: stp q2, q5, [x0, #32]
; CHECK-NEXT: stp q1, q3, [x0]
; CHECK-NEXT: ret
;
; NONEON-NOSVE-LABEL: test:
; NONEON-NOSVE: // %bb.0: // %entry
; NONEON-NOSVE-NEXT: sub sp, sp, #144
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 144
; NONEON-NOSVE-NEXT: ldp q1, q0, [x0, #32]
; NONEON-NOSVE-NEXT: ldp q2, q3, [x0]
; NONEON-NOSVE-NEXT: str q1, [sp, #64]
; NONEON-NOSVE-NEXT: stp q0, q3, [sp, #32]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #60]
; NONEON-NOSVE-NEXT: str q2, [sp, #16]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #40]
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: stp w8, w8, [sp, #8]
; NONEON-NOSVE-NEXT: str w9, [sp, #124]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #56]
; NONEON-NOSVE-NEXT: stp w8, w8, [sp]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: ldr q0, [sp]
; NONEON-NOSVE-NEXT: str w9, [sp, #120]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #52]
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: str w9, [sp, #116]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #48]
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: str w9, [sp, #112]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #28]
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: str w9, [sp, #92]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #24]
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: str w9, [sp, #88]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #20]
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: str w9, [sp, #84]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #16]
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: str w9, [sp, #80]
; NONEON-NOSVE-NEXT: ldr w9, [sp, #44]
; NONEON-NOSVE-NEXT: add w9, w9, w9
; NONEON-NOSVE-NEXT: stp w8, w9, [sp, #104]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #36]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #100]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #32]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #96]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #76]
; NONEON-NOSVE-NEXT: ldp q4, q2, [sp, #80]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #140]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #72]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #136]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #68]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #132]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #64]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #128]
; NONEON-NOSVE-NEXT: ldp q3, q1, [sp, #112]
; NONEON-NOSVE-NEXT: stp q4, q3, [x0]
; NONEON-NOSVE-NEXT: stp q1, q2, [x0, #32]
; NONEON-NOSVE-NEXT: add sp, sp, #144
; NONEON-NOSVE-NEXT: ret
entry:
%0 = load <16 x i32>, ptr %arg1, align 256
%1 = load <16 x i32>, ptr %arg2, align 256
%shvec = shufflevector <16 x i32> %0, <16 x i32> %1, <4 x i32> <i32 14, i32 14, i32 14, i32 14>
%2 = add <16 x i32> %0, %0
store <16 x i32> %2, ptr %arg1, align 256
ret <4 x i32> %shvec
}
define <2 x i32> @test2(ptr %arg1, ptr %arg2) {
; CHECK-LABEL: test2:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: ldp q1, q0, [x0, #32]
; CHECK-NEXT: ldp q3, q4, [x0]
; CHECK-NEXT: add z2.s, z0.s, z0.s
; CHECK-NEXT: ext z0.b, z0.b, z0.b, #8
; CHECK-NEXT: add z1.s, z1.s, z1.s
; CHECK-NEXT: add z3.s, z3.s, z3.s
; CHECK-NEXT: add z4.s, z4.s, z4.s
; CHECK-NEXT: mov z0.s, s0
; CHECK-NEXT: stp q1, q2, [x0, #32]
; CHECK-NEXT: stp q3, q4, [x0]
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $z0
; CHECK-NEXT: ret
;
; NONEON-NOSVE-LABEL: test2:
; NONEON-NOSVE: // %bb.0: // %entry
; NONEON-NOSVE-NEXT: sub sp, sp, #144
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 144
; NONEON-NOSVE-NEXT: ldp q1, q0, [x0, #32]
; NONEON-NOSVE-NEXT: ldp q2, q3, [x0]
; NONEON-NOSVE-NEXT: str q1, [sp, #64]
; NONEON-NOSVE-NEXT: stp q0, q3, [sp, #32]
; NONEON-NOSVE-NEXT: ldr d0, [sp, #40]
; NONEON-NOSVE-NEXT: str q2, [sp, #16]
; NONEON-NOSVE-NEXT: str d0, [sp]
; NONEON-NOSVE-NEXT: ldr w8, [sp]
; NONEON-NOSVE-NEXT: stp w8, w8, [sp, #8]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #60]
; NONEON-NOSVE-NEXT: ldr d0, [sp, #8]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #124]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #56]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #120]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #52]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #116]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #48]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #112]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #28]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #92]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #24]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #88]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #20]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #84]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #16]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #80]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #44]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #108]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #40]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #104]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #36]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #100]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #32]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #96]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #76]
; NONEON-NOSVE-NEXT: ldp q4, q2, [sp, #80]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #140]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #72]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #136]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #68]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #132]
; NONEON-NOSVE-NEXT: ldr w8, [sp, #64]
; NONEON-NOSVE-NEXT: add w8, w8, w8
; NONEON-NOSVE-NEXT: str w8, [sp, #128]
; NONEON-NOSVE-NEXT: ldp q3, q1, [sp, #112]
; NONEON-NOSVE-NEXT: stp q4, q3, [x0]
; NONEON-NOSVE-NEXT: stp q1, q2, [x0, #32]
; NONEON-NOSVE-NEXT: add sp, sp, #144
; NONEON-NOSVE-NEXT: ret
entry:
%0 = load <16 x i32>, ptr %arg1, align 256
%1 = load <16 x i32>, ptr %arg2, align 256
%shvec = shufflevector <16 x i32> %0, <16 x i32> %1, <2 x i32> <i32 14, i32 14>
%2 = add <16 x i32> %0, %0
store <16 x i32> %2, ptr %arg1, align 256
ret <2 x i32> %shvec
}
|
