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
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -march=amdgcn -mcpu=gfx908 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN %s
define amdgpu_kernel void @divergent_or3_b32(ptr addrspace(1) %arg) {
; GCN-LABEL: divergent_or3_b32:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x24
; GCN-NEXT: v_lshlrev_b32_e32 v3, 4, v0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: global_load_dwordx3 v[0:2], v3, s[0:1]
; GCN-NEXT: s_waitcnt vmcnt(0)
; GCN-NEXT: v_or3_b32 v0, v1, v0, v2
; GCN-NEXT: v_not_b32_e32 v0, v0
; GCN-NEXT: global_store_dword v3, v0, s[0:1]
; GCN-NEXT: s_endpgm
bb:
%i = tail call i32 @llvm.amdgcn.workitem.id.x()
%i1 = zext i32 %i to i64
%i2 = getelementptr inbounds <3 x i32>, ptr addrspace(1) %arg, i64 %i1
%i3 = load <3 x i32>, ptr addrspace(1) %i2, align 16
%i4 = extractelement <3 x i32> %i3, i64 0
%i5 = extractelement <3 x i32> %i3, i64 1
%i6 = extractelement <3 x i32> %i3, i64 2
%i7 = or i32 %i5, %i4
%i8 = or i32 %i7, %i6
%i9 = xor i32 %i8, -1
store i32 %i9, ptr addrspace(1) %i2, align 16
ret void
}
define amdgpu_kernel void @divergent_or3_b64(ptr addrspace(1) %arg) {
; GCN-LABEL: divergent_or3_b64:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x24
; GCN-NEXT: v_lshlrev_b32_e32 v6, 5, v0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: global_load_dwordx2 v[4:5], v6, s[0:1] offset:16
; GCN-NEXT: global_load_dwordx4 v[0:3], v6, s[0:1]
; GCN-NEXT: s_waitcnt vmcnt(0)
; GCN-NEXT: v_or3_b32 v1, v3, v1, v5
; GCN-NEXT: v_or3_b32 v0, v2, v0, v4
; GCN-NEXT: v_not_b32_e32 v1, v1
; GCN-NEXT: v_not_b32_e32 v0, v0
; GCN-NEXT: global_store_dwordx2 v6, v[0:1], s[0:1]
; GCN-NEXT: s_endpgm
bb:
%i = tail call i32 @llvm.amdgcn.workitem.id.x()
%i1 = zext i32 %i to i64
%i2 = getelementptr inbounds <3 x i64>, ptr addrspace(1) %arg, i64 %i1
%i3 = load <3 x i64>, ptr addrspace(1) %i2, align 32
%i4 = extractelement <3 x i64> %i3, i64 0
%i5 = extractelement <3 x i64> %i3, i64 1
%i6 = extractelement <3 x i64> %i3, i64 2
%i7 = or i64 %i5, %i4
%i8 = or i64 %i7, %i6
%i9 = xor i64 %i8, -1
store i64 %i9, ptr addrspace(1) %i2, align 32
ret void
}
define amdgpu_kernel void @divergent_and3_b32(ptr addrspace(1) %arg) {
; GCN-LABEL: divergent_and3_b32:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x24
; GCN-NEXT: v_lshlrev_b32_e32 v3, 4, v0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: global_load_dwordx3 v[0:2], v3, s[0:1]
; GCN-NEXT: s_waitcnt vmcnt(0)
; GCN-NEXT: v_and_b32_e32 v0, v1, v0
; GCN-NEXT: v_and_b32_e32 v0, v0, v2
; GCN-NEXT: v_not_b32_e32 v0, v0
; GCN-NEXT: global_store_dword v3, v0, s[0:1]
; GCN-NEXT: s_endpgm
bb:
%i = tail call i32 @llvm.amdgcn.workitem.id.x()
%i1 = zext i32 %i to i64
%i2 = getelementptr inbounds <3 x i32>, ptr addrspace(1) %arg, i64 %i1
%i3 = load <3 x i32>, ptr addrspace(1) %i2, align 16
%i4 = extractelement <3 x i32> %i3, i64 0
%i5 = extractelement <3 x i32> %i3, i64 1
%i6 = extractelement <3 x i32> %i3, i64 2
%i7 = and i32 %i5, %i4
%i8 = and i32 %i7, %i6
%i9 = xor i32 %i8, -1
store i32 %i9, ptr addrspace(1) %i2, align 16
ret void
}
define amdgpu_kernel void @divergent_and3_b64(ptr addrspace(1) %arg) {
; GCN-LABEL: divergent_and3_b64:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x24
; GCN-NEXT: v_lshlrev_b32_e32 v6, 5, v0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: global_load_dwordx4 v[0:3], v6, s[0:1]
; GCN-NEXT: global_load_dwordx2 v[4:5], v6, s[0:1] offset:16
; GCN-NEXT: s_waitcnt vmcnt(1)
; GCN-NEXT: v_and_b32_e32 v1, v3, v1
; GCN-NEXT: v_and_b32_e32 v0, v2, v0
; GCN-NEXT: s_waitcnt vmcnt(0)
; GCN-NEXT: v_and_b32_e32 v1, v1, v5
; GCN-NEXT: v_and_b32_e32 v0, v0, v4
; GCN-NEXT: v_not_b32_e32 v1, v1
; GCN-NEXT: v_not_b32_e32 v0, v0
; GCN-NEXT: global_store_dwordx2 v6, v[0:1], s[0:1]
; GCN-NEXT: s_endpgm
bb:
%i = tail call i32 @llvm.amdgcn.workitem.id.x()
%i1 = zext i32 %i to i64
%i2 = getelementptr inbounds <3 x i64>, ptr addrspace(1) %arg, i64 %i1
%i3 = load <3 x i64>, ptr addrspace(1) %i2, align 32
%i4 = extractelement <3 x i64> %i3, i64 0
%i5 = extractelement <3 x i64> %i3, i64 1
%i6 = extractelement <3 x i64> %i3, i64 2
%i7 = and i64 %i5, %i4
%i8 = and i64 %i7, %i6
%i9 = xor i64 %i8, -1
store i64 %i9, ptr addrspace(1) %i2, align 32
ret void
}
define amdgpu_kernel void @divergent_xor3_b32(ptr addrspace(1) %arg) {
; GCN-LABEL: divergent_xor3_b32:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x24
; GCN-NEXT: v_lshlrev_b32_e32 v3, 4, v0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: global_load_dwordx3 v[0:2], v3, s[0:1]
; GCN-NEXT: s_waitcnt vmcnt(0)
; GCN-NEXT: v_xor_b32_e32 v0, v1, v0
; GCN-NEXT: v_xnor_b32_e32 v0, v0, v2
; GCN-NEXT: global_store_dword v3, v0, s[0:1]
; GCN-NEXT: s_endpgm
bb:
%i = tail call i32 @llvm.amdgcn.workitem.id.x()
%i1 = zext i32 %i to i64
%i2 = getelementptr inbounds <3 x i32>, ptr addrspace(1) %arg, i64 %i1
%i3 = load <3 x i32>, ptr addrspace(1) %i2, align 16
%i4 = extractelement <3 x i32> %i3, i64 0
%i5 = extractelement <3 x i32> %i3, i64 1
%i6 = extractelement <3 x i32> %i3, i64 2
%i7 = xor i32 %i5, %i4
%i8 = xor i32 %i7, %i6
%i9 = xor i32 %i8, -1
store i32 %i9, ptr addrspace(1) %i2, align 16
ret void
}
define amdgpu_kernel void @divergent_xor3_b64(ptr addrspace(1) %arg) {
; GCN-LABEL: divergent_xor3_b64:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x24
; GCN-NEXT: v_lshlrev_b32_e32 v6, 5, v0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: global_load_dwordx4 v[0:3], v6, s[0:1]
; GCN-NEXT: global_load_dwordx2 v[4:5], v6, s[0:1] offset:16
; GCN-NEXT: s_waitcnt vmcnt(1)
; GCN-NEXT: v_xor_b32_e32 v1, v3, v1
; GCN-NEXT: v_xor_b32_e32 v0, v2, v0
; GCN-NEXT: s_waitcnt vmcnt(0)
; GCN-NEXT: v_xnor_b32_e32 v1, v1, v5
; GCN-NEXT: v_xnor_b32_e32 v0, v0, v4
; GCN-NEXT: global_store_dwordx2 v6, v[0:1], s[0:1]
; GCN-NEXT: s_endpgm
bb:
%i = tail call i32 @llvm.amdgcn.workitem.id.x()
%i1 = zext i32 %i to i64
%i2 = getelementptr inbounds <3 x i64>, ptr addrspace(1) %arg, i64 %i1
%i3 = load <3 x i64>, ptr addrspace(1) %i2, align 32
%i4 = extractelement <3 x i64> %i3, i64 0
%i5 = extractelement <3 x i64> %i3, i64 1
%i6 = extractelement <3 x i64> %i3, i64 2
%i7 = xor i64 %i5, %i4
%i8 = xor i64 %i7, %i6
%i9 = xor i64 %i8, -1
store i64 %i9, ptr addrspace(1) %i2, align 32
ret void
}
define amdgpu_kernel void @uniform_or3_b32(ptr addrspace(1) %arg) {
; GCN-LABEL: uniform_or3_b32:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[4:5], s[0:1], 0x24
; GCN-NEXT: v_mov_b32_e32 v0, 0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_or_b32 s0, s1, s0
; GCN-NEXT: s_nor_b32 s0, s0, s2
; GCN-NEXT: v_mov_b32_e32 v1, s0
; GCN-NEXT: global_store_dword v0, v1, s[4:5]
; GCN-NEXT: s_endpgm
bb:
%i3 = load <3 x i32>, ptr addrspace(1) %arg, align 16
%i4 = extractelement <3 x i32> %i3, i64 0
%i5 = extractelement <3 x i32> %i3, i64 1
%i6 = extractelement <3 x i32> %i3, i64 2
%i7 = or i32 %i5, %i4
%i8 = or i32 %i7, %i6
%i9 = xor i32 %i8, -1
store i32 %i9, ptr addrspace(1) %arg, align 16
ret void
}
define amdgpu_kernel void @uniform_or3_b64(ptr addrspace(1) %arg) {
; GCN-LABEL: uniform_or3_b64:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[4:5], s[0:1], 0x24
; GCN-NEXT: v_mov_b32_e32 v2, 0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GCN-NEXT: s_load_dwordx2 s[6:7], s[4:5], 0x10
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_or_b64 s[0:1], s[2:3], s[0:1]
; GCN-NEXT: s_nor_b64 s[0:1], s[0:1], s[6:7]
; GCN-NEXT: v_mov_b32_e32 v0, s0
; GCN-NEXT: v_mov_b32_e32 v1, s1
; GCN-NEXT: global_store_dwordx2 v2, v[0:1], s[4:5]
; GCN-NEXT: s_endpgm
bb:
%i3 = load <3 x i64>, ptr addrspace(1) %arg, align 32
%i4 = extractelement <3 x i64> %i3, i64 0
%i5 = extractelement <3 x i64> %i3, i64 1
%i6 = extractelement <3 x i64> %i3, i64 2
%i7 = or i64 %i5, %i4
%i8 = or i64 %i7, %i6
%i9 = xor i64 %i8, -1
store i64 %i9, ptr addrspace(1) %arg, align 32
ret void
}
define amdgpu_kernel void @uniform_and3_b32(ptr addrspace(1) %arg) {
; GCN-LABEL: uniform_and3_b32:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[4:5], s[0:1], 0x24
; GCN-NEXT: v_mov_b32_e32 v0, 0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_and_b32 s0, s1, s0
; GCN-NEXT: s_nand_b32 s0, s0, s2
; GCN-NEXT: v_mov_b32_e32 v1, s0
; GCN-NEXT: global_store_dword v0, v1, s[4:5]
; GCN-NEXT: s_endpgm
bb:
%i3 = load <3 x i32>, ptr addrspace(1) %arg, align 16
%i4 = extractelement <3 x i32> %i3, i64 0
%i5 = extractelement <3 x i32> %i3, i64 1
%i6 = extractelement <3 x i32> %i3, i64 2
%i7 = and i32 %i5, %i4
%i8 = and i32 %i7, %i6
%i9 = xor i32 %i8, -1
store i32 %i9, ptr addrspace(1) %arg, align 16
ret void
}
define amdgpu_kernel void @uniform_and3_b64(ptr addrspace(1) %arg) {
; GCN-LABEL: uniform_and3_b64:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[4:5], s[0:1], 0x24
; GCN-NEXT: v_mov_b32_e32 v2, 0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GCN-NEXT: s_load_dwordx2 s[6:7], s[4:5], 0x10
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_and_b64 s[0:1], s[2:3], s[0:1]
; GCN-NEXT: s_nand_b64 s[0:1], s[0:1], s[6:7]
; GCN-NEXT: v_mov_b32_e32 v0, s0
; GCN-NEXT: v_mov_b32_e32 v1, s1
; GCN-NEXT: global_store_dwordx2 v2, v[0:1], s[4:5]
; GCN-NEXT: s_endpgm
bb:
%i3 = load <3 x i64>, ptr addrspace(1) %arg, align 32
%i4 = extractelement <3 x i64> %i3, i64 0
%i5 = extractelement <3 x i64> %i3, i64 1
%i6 = extractelement <3 x i64> %i3, i64 2
%i7 = and i64 %i5, %i4
%i8 = and i64 %i7, %i6
%i9 = xor i64 %i8, -1
store i64 %i9, ptr addrspace(1) %arg, align 32
ret void
}
define amdgpu_kernel void @uniform_xor3_b32(ptr addrspace(1) %arg) {
; GCN-LABEL: uniform_xor3_b32:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[4:5], s[0:1], 0x24
; GCN-NEXT: v_mov_b32_e32 v0, 0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_xor_b32 s0, s1, s0
; GCN-NEXT: s_xnor_b32 s0, s0, s2
; GCN-NEXT: v_mov_b32_e32 v1, s0
; GCN-NEXT: global_store_dword v0, v1, s[4:5]
; GCN-NEXT: s_endpgm
bb:
%i3 = load <3 x i32>, ptr addrspace(1) %arg, align 16
%i4 = extractelement <3 x i32> %i3, i64 0
%i5 = extractelement <3 x i32> %i3, i64 1
%i6 = extractelement <3 x i32> %i3, i64 2
%i7 = xor i32 %i5, %i4
%i8 = xor i32 %i7, %i6
%i9 = xor i32 %i8, -1
store i32 %i9, ptr addrspace(1) %arg, align 16
ret void
}
define amdgpu_kernel void @uniform_xor3_b64(ptr addrspace(1) %arg) {
; GCN-LABEL: uniform_xor3_b64:
; GCN: ; %bb.0: ; %bb
; GCN-NEXT: s_load_dwordx2 s[4:5], s[0:1], 0x24
; GCN-NEXT: v_mov_b32_e32 v2, 0
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_load_dwordx4 s[0:3], s[4:5], 0x0
; GCN-NEXT: s_load_dwordx2 s[6:7], s[4:5], 0x10
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_xor_b64 s[0:1], s[2:3], s[0:1]
; GCN-NEXT: s_xnor_b64 s[0:1], s[0:1], s[6:7]
; GCN-NEXT: v_mov_b32_e32 v0, s0
; GCN-NEXT: v_mov_b32_e32 v1, s1
; GCN-NEXT: global_store_dwordx2 v2, v[0:1], s[4:5]
; GCN-NEXT: s_endpgm
bb:
%i3 = load <3 x i64>, ptr addrspace(1) %arg, align 32
%i4 = extractelement <3 x i64> %i3, i64 0
%i5 = extractelement <3 x i64> %i3, i64 1
%i6 = extractelement <3 x i64> %i3, i64 2
%i7 = xor i64 %i5, %i4
%i8 = xor i64 %i7, %i6
%i9 = xor i64 %i8, -1
store i64 %i9, ptr addrspace(1) %arg, align 32
ret void
}
declare i32 @llvm.amdgcn.workitem.id.x()
|
