store-local.96.ll
9.75 KB
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
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -global-isel -mtriple=amdgcn-amd-amdpal -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX9 %s
; RUN: llc -global-isel -mtriple=amdgcn-amd-amdpal -mcpu=hawaii -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX7 %s
; FIXME:
; XUN: llc -global-isel -mtriple=amdgcn-amd-amdpal -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX6 %s
define amdgpu_kernel void @store_lds_v3i32(<3 x i32> addrspace(3)* %out, <3 x i32> %x) {
; GFX9-LABEL: store_lds_v3i32:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s4, s[0:1], 0x24
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x34
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v3, s4
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: v_mov_b32_e32 v1, s1
; GFX9-NEXT: v_mov_b32_e32 v2, s2
; GFX9-NEXT: ds_write_b96 v3, v[0:2]
; GFX9-NEXT: s_endpgm
;
; GFX7-LABEL: store_lds_v3i32:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_load_dword s4, s[0:1], 0x9
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0xd
; GFX7-NEXT: s_mov_b32 m0, -1
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
; GFX7-NEXT: v_mov_b32_e32 v3, s4
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: v_mov_b32_e32 v1, s1
; GFX7-NEXT: v_mov_b32_e32 v2, s2
; GFX7-NEXT: ds_write_b96 v3, v[0:2]
; GFX7-NEXT: s_endpgm
store <3 x i32> %x, <3 x i32> addrspace(3)* %out
ret void
}
define amdgpu_kernel void @store_lds_v3i32_align1(<3 x i32> addrspace(3)* %out, <3 x i32> %x) {
; GFX9-LABEL: store_lds_v3i32_align1:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s4, s[0:1], 0x24
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x34
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v1, s4
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: s_lshr_b32 s3, s0, 8
; GFX9-NEXT: ds_write_b8 v1, v0
; GFX9-NEXT: v_mov_b32_e32 v0, s3
; GFX9-NEXT: s_lshr_b32 s5, s0, 16
; GFX9-NEXT: s_lshr_b32 s6, s0, 24
; GFX9-NEXT: ds_write_b8 v1, v0 offset:1
; GFX9-NEXT: v_mov_b32_e32 v0, s5
; GFX9-NEXT: ds_write_b8 v1, v0 offset:2
; GFX9-NEXT: v_mov_b32_e32 v0, s6
; GFX9-NEXT: ds_write_b8 v1, v0 offset:3
; GFX9-NEXT: v_mov_b32_e32 v0, s1
; GFX9-NEXT: s_lshr_b32 s0, s1, 8
; GFX9-NEXT: ds_write_b8 v1, v0 offset:4
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: s_lshr_b32 s3, s1, 16
; GFX9-NEXT: s_lshr_b32 s4, s1, 24
; GFX9-NEXT: ds_write_b8 v1, v0 offset:5
; GFX9-NEXT: v_mov_b32_e32 v0, s3
; GFX9-NEXT: ds_write_b8 v1, v0 offset:6
; GFX9-NEXT: v_mov_b32_e32 v0, s4
; GFX9-NEXT: ds_write_b8 v1, v0 offset:7
; GFX9-NEXT: v_mov_b32_e32 v0, s2
; GFX9-NEXT: s_lshr_b32 s0, s2, 8
; GFX9-NEXT: ds_write_b8 v1, v0 offset:8
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: s_lshr_b32 s1, s2, 16
; GFX9-NEXT: ds_write_b8 v1, v0 offset:9
; GFX9-NEXT: v_mov_b32_e32 v0, s1
; GFX9-NEXT: s_lshr_b32 s3, s2, 24
; GFX9-NEXT: ds_write_b8 v1, v0 offset:10
; GFX9-NEXT: v_mov_b32_e32 v0, s3
; GFX9-NEXT: ds_write_b8 v1, v0 offset:11
; GFX9-NEXT: s_endpgm
;
; GFX7-LABEL: store_lds_v3i32_align1:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_load_dword s4, s[0:1], 0x9
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0xd
; GFX7-NEXT: s_mov_b32 m0, -1
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
; GFX7-NEXT: v_mov_b32_e32 v1, s4
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: s_lshr_b32 s3, s0, 8
; GFX7-NEXT: ds_write_b8 v1, v0
; GFX7-NEXT: v_mov_b32_e32 v0, s3
; GFX7-NEXT: s_lshr_b32 s5, s0, 16
; GFX7-NEXT: s_lshr_b32 s6, s0, 24
; GFX7-NEXT: ds_write_b8 v1, v0 offset:1
; GFX7-NEXT: v_mov_b32_e32 v0, s5
; GFX7-NEXT: ds_write_b8 v1, v0 offset:2
; GFX7-NEXT: v_mov_b32_e32 v0, s6
; GFX7-NEXT: ds_write_b8 v1, v0 offset:3
; GFX7-NEXT: v_mov_b32_e32 v0, s1
; GFX7-NEXT: s_lshr_b32 s0, s1, 8
; GFX7-NEXT: ds_write_b8 v1, v0 offset:4
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: s_lshr_b32 s3, s1, 16
; GFX7-NEXT: s_lshr_b32 s4, s1, 24
; GFX7-NEXT: ds_write_b8 v1, v0 offset:5
; GFX7-NEXT: v_mov_b32_e32 v0, s3
; GFX7-NEXT: ds_write_b8 v1, v0 offset:6
; GFX7-NEXT: v_mov_b32_e32 v0, s4
; GFX7-NEXT: ds_write_b8 v1, v0 offset:7
; GFX7-NEXT: v_mov_b32_e32 v0, s2
; GFX7-NEXT: s_lshr_b32 s0, s2, 8
; GFX7-NEXT: ds_write_b8 v1, v0 offset:8
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: s_lshr_b32 s1, s2, 16
; GFX7-NEXT: ds_write_b8 v1, v0 offset:9
; GFX7-NEXT: v_mov_b32_e32 v0, s1
; GFX7-NEXT: s_lshr_b32 s3, s2, 24
; GFX7-NEXT: ds_write_b8 v1, v0 offset:10
; GFX7-NEXT: v_mov_b32_e32 v0, s3
; GFX7-NEXT: ds_write_b8 v1, v0 offset:11
; GFX7-NEXT: s_endpgm
store <3 x i32> %x, <3 x i32> addrspace(3)* %out, align 1
ret void
}
define amdgpu_kernel void @store_lds_v3i32_align2(<3 x i32> addrspace(3)* %out, <3 x i32> %x) {
; GFX9-LABEL: store_lds_v3i32_align2:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s4, s[0:1], 0x24
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x34
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v1, s4
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: s_lshr_b32 s3, s0, 16
; GFX9-NEXT: ds_write_b16 v1, v0
; GFX9-NEXT: v_mov_b32_e32 v0, s3
; GFX9-NEXT: ds_write_b16 v1, v0 offset:2
; GFX9-NEXT: v_mov_b32_e32 v0, s1
; GFX9-NEXT: s_lshr_b32 s0, s1, 16
; GFX9-NEXT: ds_write_b16 v1, v0 offset:4
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: ds_write_b16 v1, v0 offset:6
; GFX9-NEXT: v_mov_b32_e32 v0, s2
; GFX9-NEXT: s_lshr_b32 s0, s2, 16
; GFX9-NEXT: ds_write_b16 v1, v0 offset:8
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: ds_write_b16 v1, v0 offset:10
; GFX9-NEXT: s_endpgm
;
; GFX7-LABEL: store_lds_v3i32_align2:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_load_dword s4, s[0:1], 0x9
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0xd
; GFX7-NEXT: s_mov_b32 m0, -1
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
; GFX7-NEXT: v_mov_b32_e32 v1, s4
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: s_lshr_b32 s3, s0, 16
; GFX7-NEXT: ds_write_b16 v1, v0
; GFX7-NEXT: v_mov_b32_e32 v0, s3
; GFX7-NEXT: ds_write_b16 v1, v0 offset:2
; GFX7-NEXT: v_mov_b32_e32 v0, s1
; GFX7-NEXT: s_lshr_b32 s0, s1, 16
; GFX7-NEXT: ds_write_b16 v1, v0 offset:4
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: ds_write_b16 v1, v0 offset:6
; GFX7-NEXT: v_mov_b32_e32 v0, s2
; GFX7-NEXT: s_lshr_b32 s0, s2, 16
; GFX7-NEXT: ds_write_b16 v1, v0 offset:8
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: ds_write_b16 v1, v0 offset:10
; GFX7-NEXT: s_endpgm
store <3 x i32> %x, <3 x i32> addrspace(3)* %out, align 2
ret void
}
define amdgpu_kernel void @store_lds_v3i32_align4(<3 x i32> addrspace(3)* %out, <3 x i32> %x) {
; GFX9-LABEL: store_lds_v3i32_align4:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s4, s[0:1], 0x24
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x34
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v2, s4
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: v_mov_b32_e32 v1, s1
; GFX9-NEXT: v_mov_b32_e32 v3, s2
; GFX9-NEXT: ds_write2_b32 v2, v0, v1 offset1:1
; GFX9-NEXT: ds_write_b32 v2, v3 offset:8
; GFX9-NEXT: s_endpgm
;
; GFX7-LABEL: store_lds_v3i32_align4:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_load_dword s4, s[0:1], 0x9
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0xd
; GFX7-NEXT: s_mov_b32 m0, -1
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
; GFX7-NEXT: v_mov_b32_e32 v2, s4
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: v_mov_b32_e32 v1, s1
; GFX7-NEXT: v_mov_b32_e32 v3, s2
; GFX7-NEXT: ds_write2_b32 v2, v0, v1 offset1:1
; GFX7-NEXT: ds_write_b32 v2, v3 offset:8
; GFX7-NEXT: s_endpgm
store <3 x i32> %x, <3 x i32> addrspace(3)* %out, align 4
ret void
}
define amdgpu_kernel void @store_lds_v3i32_align8(<3 x i32> addrspace(3)* %out, <3 x i32> %x) {
; GFX9-LABEL: store_lds_v3i32_align8:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s4, s[0:1], 0x24
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x34
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v2, s4
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: v_mov_b32_e32 v1, s1
; GFX9-NEXT: v_mov_b32_e32 v3, s2
; GFX9-NEXT: ds_write_b64 v2, v[0:1]
; GFX9-NEXT: ds_write_b32 v2, v3 offset:8
; GFX9-NEXT: s_endpgm
;
; GFX7-LABEL: store_lds_v3i32_align8:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_load_dword s4, s[0:1], 0x9
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0xd
; GFX7-NEXT: s_mov_b32 m0, -1
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
; GFX7-NEXT: v_mov_b32_e32 v2, s4
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: v_mov_b32_e32 v1, s1
; GFX7-NEXT: v_mov_b32_e32 v3, s2
; GFX7-NEXT: ds_write_b64 v2, v[0:1]
; GFX7-NEXT: ds_write_b32 v2, v3 offset:8
; GFX7-NEXT: s_endpgm
store <3 x i32> %x, <3 x i32> addrspace(3)* %out, align 8
ret void
}
define amdgpu_kernel void @store_lds_v3i32_align16(<3 x i32> addrspace(3)* %out, <3 x i32> %x) {
; GFX9-LABEL: store_lds_v3i32_align16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_load_dword s4, s[0:1], 0x24
; GFX9-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0x34
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v3, s4
; GFX9-NEXT: v_mov_b32_e32 v0, s0
; GFX9-NEXT: v_mov_b32_e32 v1, s1
; GFX9-NEXT: v_mov_b32_e32 v2, s2
; GFX9-NEXT: ds_write_b96 v3, v[0:2]
; GFX9-NEXT: s_endpgm
;
; GFX7-LABEL: store_lds_v3i32_align16:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_load_dword s4, s[0:1], 0x9
; GFX7-NEXT: s_load_dwordx4 s[0:3], s[0:1], 0xd
; GFX7-NEXT: s_mov_b32 m0, -1
; GFX7-NEXT: s_waitcnt lgkmcnt(0)
; GFX7-NEXT: v_mov_b32_e32 v3, s4
; GFX7-NEXT: v_mov_b32_e32 v0, s0
; GFX7-NEXT: v_mov_b32_e32 v1, s1
; GFX7-NEXT: v_mov_b32_e32 v2, s2
; GFX7-NEXT: ds_write_b96 v3, v[0:2]
; GFX7-NEXT: s_endpgm
store <3 x i32> %x, <3 x i32> addrspace(3)* %out, align 16
ret void
}