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
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sme2 -mattr=+sme-i16i64 -mattr=+sme-f64f64 -verify-machineinstrs < %s | FileCheck %s
;
; SUB Multi-Single x2
;
define void @multi_vector_sub_write_single_za_vg1x2_i32(i32 %slice, <vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1, <vscale x 4 x i32> %zm) {
; CHECK-LABEL: multi_vector_sub_write_single_za_vg1x2_i32:
; CHECK: // %bb.0:
; CHECK-NEXT: mov w8, w0
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT: sub za.s[w8, 0, vgx2], { z0.s, z1.s }, z2.s
; CHECK-NEXT: sub za.s[w8, 7, vgx2], { z0.s, z1.s }, z2.s
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.sub.write.single.za.vg1x2.nxv4i32(i32 %slice,
<vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
<vscale x 4 x i32> %zm)
%slice.7 = add i32 %slice, 7
call void @llvm.aarch64.sme.sub.write.single.za.vg1x2.nxv4i32(i32 %slice.7,
<vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
<vscale x 4 x i32> %zm)
ret void
}
define void @multi_vector_sub_write_single_za_vg1x2_i64(i32 %slice, <vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1, <vscale x 2 x i64> %zm) {
; CHECK-LABEL: multi_vector_sub_write_single_za_vg1x2_i64:
; CHECK: // %bb.0:
; CHECK-NEXT: mov w8, w0
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT: sub za.d[w8, 0, vgx2], { z0.d, z1.d }, z2.d
; CHECK-NEXT: sub za.d[w8, 7, vgx2], { z0.d, z1.d }, z2.d
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.sub.write.single.za.vg1x2.nxv2i64(i32 %slice,
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zm)
%slice.7 = add i32 %slice, 7
call void @llvm.aarch64.sme.sub.write.single.za.vg1x2.nxv2i64(i32 %slice.7,
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zm)
ret void
}
;
; SUB Multi-Single x4
;
define void @multi_vector_sub_write_single_za_vg1x4_i32(i32 %slice, <vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
; CHECK-LABEL: multi_vector_sub_write_single_za_vg1x4_i32:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: mov w8, w0
; CHECK-NEXT: // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: sub za.s[w8, 0, vgx4], { z0.s - z3.s }, z4.s
; CHECK-NEXT: sub za.s[w8, 7, vgx4], { z0.s - z3.s }, z4.s
; CHECK-NEXT: ret
<vscale x 4 x i32> %zn2, <vscale x 4 x i32> %zn3,
<vscale x 4 x i32> %zm) {
call void @llvm.aarch64.sme.sub.write.single.za.vg1x4.nxv4i32(i32 %slice,
<vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
<vscale x 4 x i32> %zn2, <vscale x 4 x i32> %zn3,
<vscale x 4 x i32> %zm)
%slice.7 = add i32 %slice, 7
call void @llvm.aarch64.sme.sub.write.single.za.vg1x4.nxv4i32(i32 %slice.7,
<vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
<vscale x 4 x i32> %zn2, <vscale x 4 x i32> %zn3,
<vscale x 4 x i32> %zm)
ret void
}
define void @multi_vector_sub_write_single_za_vg1x4_i64(i32 %slice,
; CHECK-LABEL: multi_vector_sub_write_single_za_vg1x4_i64:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: mov w8, w0
; CHECK-NEXT: // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: sub za.d[w8, 0, vgx4], { z0.d - z3.d }, z4.d
; CHECK-NEXT: sub za.d[w8, 7, vgx4], { z0.d - z3.d }, z4.d
; CHECK-NEXT: ret
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zn2, <vscale x 2 x i64> %zn3,
<vscale x 2 x i64> %zm) {
call void @llvm.aarch64.sme.sub.write.single.za.vg1x4.nxv2i64(i32 %slice,
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zn2, <vscale x 2 x i64> %zn3,
<vscale x 2 x i64> %zm)
%slice.7 = add i32 %slice, 7
call void @llvm.aarch64.sme.sub.write.single.za.vg1x4.nxv2i64(i32 %slice.7,
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zn2, <vscale x 2 x i64> %zn3,
<vscale x 2 x i64> %zm)
ret void
}
;
; SUB Multi-Multi x2
;
define void @multi_vector_sub_write_za_vg1x2_i32(i32 %slice, <vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
; CHECK-LABEL: multi_vector_sub_write_za_vg1x2_i32:
; CHECK: // %bb.0:
; CHECK-NEXT: mov w8, w0
; CHECK-NEXT: // kill: def $z3 killed $z3 killed $z2_z3 def $z2_z3
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT: // kill: def $z2 killed $z2 killed $z2_z3 def $z2_z3
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT: sub za.s[w8, 0, vgx2], { z0.s, z1.s }, { z2.s, z3.s }
; CHECK-NEXT: sub za.s[w8, 7, vgx2], { z0.s, z1.s }, { z2.s, z3.s }
; CHECK-NEXT: ret
<vscale x 4 x i32> %zm1, <vscale x 4 x i32> %zm2) {
call void @llvm.aarch64.sme.sub.write.za.vg1x2.nxv4i32(i32 %slice,
<vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
<vscale x 4 x i32> %zm1, <vscale x 4 x i32> %zm2)
%slice.7 = add i32 %slice, 7
call void @llvm.aarch64.sme.sub.write.za.vg1x2.nxv4i32(i32 %slice.7,
<vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
<vscale x 4 x i32> %zm1, <vscale x 4 x i32> %zm2)
ret void
}
define void @multi_vector_sub_write_za_vg1x2_i64(i32 %slice, <vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
; CHECK-LABEL: multi_vector_sub_write_za_vg1x2_i64:
; CHECK: // %bb.0:
; CHECK-NEXT: mov w8, w0
; CHECK-NEXT: // kill: def $z3 killed $z3 killed $z2_z3 def $z2_z3
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT: // kill: def $z2 killed $z2 killed $z2_z3 def $z2_z3
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT: sub za.d[w8, 0, vgx2], { z0.d, z1.d }, { z2.d, z3.d }
; CHECK-NEXT: sub za.d[w8, 7, vgx2], { z0.d, z1.d }, { z2.d, z3.d }
; CHECK-NEXT: ret
<vscale x 2 x i64> %zm1, <vscale x 2 x i64> %zm2) {
call void @llvm.aarch64.sme.sub.write.za.vg1x2.nxv2i64(i32 %slice,
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zm1, <vscale x 2 x i64> %zm2)
%slice.7 = add i32 %slice, 7
call void @llvm.aarch64.sme.sub.write.za.vg1x2.nxv2i64(i32 %slice.7,
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zm1, <vscale x 2 x i64> %zm2)
ret void
}
;
; SUB Multi-Multi x4
;
define void @multi_vector_sub_write_za_vg1x4_i32(i32 %slice, <vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
; CHECK-LABEL: multi_vector_sub_write_za_vg1x4_i32:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $z7 killed $z7 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT: // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: mov w8, w0
; CHECK-NEXT: // kill: def $z6 killed $z6 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT: // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: // kill: def $z5 killed $z5 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: // kill: def $z4 killed $z4 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: sub za.s[w8, 0, vgx4], { z0.s - z3.s }, { z4.s - z7.s }
; CHECK-NEXT: sub za.s[w8, 7, vgx4], { z0.s - z3.s }, { z4.s - z7.s }
; CHECK-NEXT: ret
<vscale x 4 x i32> %zn2, <vscale x 4 x i32> %zn3,
<vscale x 4 x i32> %zm0, <vscale x 4 x i32> %zm1,
<vscale x 4 x i32> %zm2, <vscale x 4 x i32> %zm3) {
call void @llvm.aarch64.sme.sub.write.za.vg1x4.nxv4i32(i32 %slice,
<vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
<vscale x 4 x i32> %zn2, <vscale x 4 x i32> %zn3,
<vscale x 4 x i32> %zm0, <vscale x 4 x i32> %zm1,
<vscale x 4 x i32> %zm2, <vscale x 4 x i32> %zm3)
%slice.7 = add i32 %slice, 7
call void @llvm.aarch64.sme.sub.write.za.vg1x4.nxv4i32(i32 %slice.7,
<vscale x 4 x i32> %zn0, <vscale x 4 x i32> %zn1,
<vscale x 4 x i32> %zn2, <vscale x 4 x i32> %zn3,
<vscale x 4 x i32> %zm0, <vscale x 4 x i32> %zm1,
<vscale x 4 x i32> %zm2, <vscale x 4 x i32> %zm3)
ret void
}
define void @multi_vector_sub_write_za_vg1x4_i64(i32 %slice, <vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
; CHECK-LABEL: multi_vector_sub_write_za_vg1x4_i64:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $z7 killed $z7 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT: // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: mov w8, w0
; CHECK-NEXT: // kill: def $z6 killed $z6 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT: // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: // kill: def $z5 killed $z5 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: // kill: def $z4 killed $z4 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT: sub za.d[w8, 0, vgx4], { z0.d - z3.d }, { z4.d - z7.d }
; CHECK-NEXT: sub za.d[w8, 7, vgx4], { z0.d - z3.d }, { z4.d - z7.d }
; CHECK-NEXT: ret
<vscale x 2 x i64> %zn2, <vscale x 2 x i64> %zn3,
<vscale x 2 x i64> %zm0, <vscale x 2 x i64> %zm1,
<vscale x 2 x i64> %zm2, <vscale x 2 x i64> %zm3) {
call void @llvm.aarch64.sme.sub.write.za.vg1x4.nxv2i64(i32 %slice,
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zn2, <vscale x 2 x i64> %zn3,
<vscale x 2 x i64> %zm0, <vscale x 2 x i64> %zm1,
<vscale x 2 x i64> %zm2, <vscale x 2 x i64> %zm3)
%slice.7 = add i32 %slice, 7
call void @llvm.aarch64.sme.sub.write.za.vg1x4.nxv2i64(i32 %slice.7,
<vscale x 2 x i64> %zn0, <vscale x 2 x i64> %zn1,
<vscale x 2 x i64> %zn2, <vscale x 2 x i64> %zn3,
<vscale x 2 x i64> %zm0, <vscale x 2 x i64> %zm1,
<vscale x 2 x i64> %zm2, <vscale x 2 x i64> %zm3)
ret void
}
declare void@llvm.aarch64.sme.sub.write.single.za.vg1x2.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>)
declare void@llvm.aarch64.sme.sub.write.single.za.vg1x2.nxv2i64(i32, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>)
declare void@llvm.aarch64.sme.sub.write.single.za.vg1x4.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>,
<vscale x 4 x i32>, <vscale x 4 x i32>)
declare void@llvm.aarch64.sme.sub.write.single.za.vg1x4.nxv2i64(i32, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>,
<vscale x 2 x i64>, <vscale x 2 x i64>)
declare void@llvm.aarch64.sme.sub.write.za.vg1x2.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>)
declare void@llvm.aarch64.sme.sub.write.za.vg1x2.nxv2i64(i32, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>)
declare void@llvm.aarch64.sme.sub.write.za.vg1x4.nxv4i32(i32, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>)
declare void@llvm.aarch64.sme.sub.write.za.vg1x4.nxv2i64(i32, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>)
|
