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
|
; RUN: llvm-as < %s | llvm-dis | FileCheck %s
define void @atomic_inc(ptr %ptr0, ptr addrspace(1) %ptr1, ptr addrspace(3) %ptr3) {
; CHECK: atomicrmw uinc_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result0 = call i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr %ptr0, i32 42, i32 0, i32 0, i1 false)
; CHECK: atomicrmw uinc_wrap ptr addrspace(1) %ptr1, i32 43 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result1 = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %ptr1, i32 43, i32 0, i32 0, i1 false)
; CHECK: atomicrmw uinc_wrap ptr addrspace(3) %ptr3, i32 46 syncscope("agent") seq_cst, align 4{{$}}
%result2 = call i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) %ptr3, i32 46, i32 0, i32 0, i1 false)
; CHECK: atomicrmw uinc_wrap ptr %ptr0, i64 48 syncscope("agent") seq_cst, align 8
%result3 = call i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr %ptr0, i64 48, i32 0, i32 0, i1 false)
; CHECK: atomicrmw uinc_wrap ptr addrspace(1) %ptr1, i64 45 syncscope("agent") seq_cst, align 8
%result4 = call i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) %ptr1, i64 45, i32 0, i32 0, i1 false)
; CHECK: atomicrmw uinc_wrap ptr addrspace(3) %ptr3, i64 4345 syncscope("agent") seq_cst, align 8
%result5 = call i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) %ptr3, i64 4345, i32 0, i32 0, i1 false)
; CHECK: atomicrmw volatile uinc_wrap ptr addrspace(3) %ptr3, i64 4345 syncscope("agent") seq_cst, align 8
%result6 = call i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) %ptr3, i64 4345, i32 0, i32 0, i1 true)
ret void
}
define void @atomic_dec(ptr %ptr0, ptr addrspace(1) %ptr1, ptr addrspace(3) %ptr3) {
; CHECK: atomicrmw udec_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result0 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 0, i32 0, i1 false)
; CHECK: atomicrmw udec_wrap ptr addrspace(1) %ptr1, i32 43 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result1 = call i32 @llvm.amdgcn.atomic.dec.i32.p1(ptr addrspace(1) %ptr1, i32 43, i32 0, i32 0, i1 false)
; CHECK: atomicrmw udec_wrap ptr addrspace(3) %ptr3, i32 46 syncscope("agent") seq_cst, align 4{{$}}
%result2 = call i32 @llvm.amdgcn.atomic.dec.i32.p3(ptr addrspace(3) %ptr3, i32 46, i32 0, i32 0, i1 false)
; CHECK: atomicrmw udec_wrap ptr %ptr0, i64 48 syncscope("agent") seq_cst, align 8
%result3 = call i64 @llvm.amdgcn.atomic.dec.i64.p0(ptr %ptr0, i64 48, i32 0, i32 0, i1 false)
; CHECK: atomicrmw udec_wrap ptr addrspace(1) %ptr1, i64 45 syncscope("agent") seq_cst, align 8
%result4 = call i64 @llvm.amdgcn.atomic.dec.i64.p1(ptr addrspace(1) %ptr1, i64 45, i32 0, i32 0, i1 false)
; CHECK: atomicrmw udec_wrap ptr addrspace(3) %ptr3, i64 4345 syncscope("agent") seq_cst, align 8
%result5 = call i64 @llvm.amdgcn.atomic.dec.i64.p3(ptr addrspace(3) %ptr3, i64 4345, i32 0, i32 0, i1 false)
; CHECK: atomicrmw volatile udec_wrap ptr addrspace(3) %ptr3, i64 4345 syncscope("agent") seq_cst, align 8
%result6 = call i64 @llvm.amdgcn.atomic.dec.i64.p3(ptr addrspace(3) %ptr3, i64 4345, i32 0, i32 0, i1 true)
ret void
}
; Test some invalid ordering handling
define void @ordering(ptr %ptr0, ptr addrspace(1) %ptr1, ptr addrspace(3) %ptr3) {
; CHECK: atomicrmw volatile uinc_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result0 = call i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr %ptr0, i32 42, i32 -1, i32 0, i1 true)
; CHECK: atomicrmw volatile uinc_wrap ptr addrspace(1) %ptr1, i32 43 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result1 = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %ptr1, i32 43, i32 0, i32 0, i1 true)
; CHECK: atomicrmw uinc_wrap ptr addrspace(1) %ptr1, i32 43 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result2 = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %ptr1, i32 43, i32 1, i32 0, i1 false)
; CHECK: atomicrmw volatile uinc_wrap ptr addrspace(1) %ptr1, i32 43 syncscope("agent") monotonic, align 4, !amdgpu.no.fine.grained.memory !0
%result3 = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %ptr1, i32 43, i32 2, i32 0, i1 true)
; CHECK: atomicrmw uinc_wrap ptr addrspace(1) %ptr1, i32 43 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result4 = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %ptr1, i32 43, i32 3, i32 0, i1 false)
; CHECK: atomicrmw volatile udec_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result5 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 0, i32 4, i1 true)
; CHECK: atomicrmw udec_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result6 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 0, i32 5, i1 false)
; CHECK: atomicrmw volatile udec_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result7 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 0, i32 6, i1 true)
; CHECK: atomicrmw udec_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result8 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 0, i32 7, i1 false)
; CHECK:= atomicrmw volatile udec_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result9 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 0, i32 8, i1 true)
; CHECK:= atomicrmw volatile udec_wrap ptr addrspace(1) %ptr1, i32 43 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result10 = call i32 @llvm.amdgcn.atomic.dec.i32.p1(ptr addrspace(1) %ptr1, i32 43, i32 3, i32 0, i1 true)
ret void
}
define void @immarg_violations(ptr %ptr0, i32 %val32, i1 %val1) {
; CHECK: atomicrmw udec_wrap ptr %ptr0, i32 42 syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
%result0 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 %val32, i32 0, i1 false)
; CHECK: atomicrmw udec_wrap ptr %ptr0, i32 42 syncscope("agent") monotonic, align 4, !amdgpu.no.fine.grained.memory !0
%result1 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 2, i32 %val32, i1 false)
; CHECK: atomicrmw volatile udec_wrap ptr %ptr0, i32 42 syncscope("agent") monotonic, align 4, !amdgpu.no.fine.grained.memory !0
%result2 = call i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr %ptr0, i32 42, i32 2, i32 0, i1 %val1)
ret void
}
declare i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) nocapture, i32, i32 immarg, i32 immarg, i1 immarg) #0
declare i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) nocapture, i32, i32 immarg, i32 immarg, i1 immarg) #0
declare i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr nocapture, i32, i32 immarg, i32 immarg, i1 immarg) #0
declare i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) nocapture, i64, i32 immarg, i32 immarg, i1 immarg) #0
declare i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) nocapture, i64, i32 immarg, i32 immarg, i1 immarg) #0
declare i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr nocapture, i64, i32 immarg, i32 immarg, i1 immarg) #0
declare i32 @llvm.amdgcn.atomic.dec.i32.p1(ptr addrspace(1) nocapture, i32, i32 immarg, i32 immarg, i1 immarg) #0
declare i32 @llvm.amdgcn.atomic.dec.i32.p3(ptr addrspace(3) nocapture, i32, i32 immarg, i32 immarg, i1 immarg) #0
declare i32 @llvm.amdgcn.atomic.dec.i32.p0(ptr nocapture, i32, i32 immarg, i32 immarg, i1 immarg) #0
declare i64 @llvm.amdgcn.atomic.dec.i64.p1(ptr addrspace(1) nocapture, i64, i32 immarg, i32 immarg, i1 immarg) #0
declare i64 @llvm.amdgcn.atomic.dec.i64.p3(ptr addrspace(3) nocapture, i64, i32 immarg, i32 immarg, i1 immarg) #0
declare i64 @llvm.amdgcn.atomic.dec.i64.p0(ptr nocapture, i64, i32 immarg, i32 immarg, i1 immarg) #0
; ptr, rmw_value, ordering, scope, isVolatile)
declare float @llvm.amdgcn.ds.fadd.f32(ptr addrspace(3) nocapture, float, i32 immarg, i32 immarg, i1 immarg)
declare double @llvm.amdgcn.ds.fadd.f64(ptr addrspace(3) nocapture, double, i32 immarg, i32 immarg, i1 immarg)
declare <2 x half> @llvm.amdgcn.ds.fadd.v2f16(ptr addrspace(3) nocapture, <2 x half>, i32 immarg, i32 immarg, i1 immarg)
declare <2 x i16> @llvm.amdgcn.ds.fadd.v2bf16(ptr addrspace(3) nocapture, <2 x i16>)
define float @upgrade_amdgcn_ds_fadd_f32(ptr addrspace(3) %ptr, float %val) {
; CHECK: atomicrmw fadd ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result0 = call float @llvm.amdgcn.ds.fadd.f32(ptr addrspace(3) %ptr, float %val, i32 0, i32 0, i1 false)
; CHECK: = atomicrmw volatile fadd ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result1 = call float @llvm.amdgcn.ds.fadd.f32(ptr addrspace(3) %ptr, float %val, i32 0, i32 0, i1 true)
; CHECK: = atomicrmw fadd ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result2 = call float @llvm.amdgcn.ds.fadd.f32(ptr addrspace(3) %ptr, float %val, i32 43, i32 3, i1 false)
; CHECK: = atomicrmw fadd ptr addrspace(3) %ptr, float %val syncscope("agent") acquire, align 4
%result3 = call float @llvm.amdgcn.ds.fadd.f32(ptr addrspace(3) %ptr, float %val, i32 4, i32 2, i1 false)
ret float %result3
}
; Handle missing type suffix
declare float @llvm.amdgcn.ds.fadd(ptr addrspace(3) nocapture, float, i32 immarg, i32 immarg, i1 immarg)
define float @upgrade_amdgcn_ds_fadd_f32_no_suffix(ptr addrspace(3) %ptr, float %val) {
; CHECK: atomicrmw fadd ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result0 = call float @llvm.amdgcn.ds.fadd(ptr addrspace(3) %ptr, float %val, i32 0, i32 0, i1 false)
ret float %result0
}
define void @immarg_violations_ds_fadd_f32(ptr addrspace(3) %ptr, float %fval, i32 %val32, i1 %val1) {
; CHECK: = atomicrmw volatile fadd ptr addrspace(3) %ptr, float %fval syncscope("agent") seq_cst, align 4
%result0 = call float @llvm.amdgcn.ds.fadd.f32(ptr addrspace(3) %ptr, float %fval, i32 %val32, i32 %val32, i1 %val1)
ret void
}
declare float @llvm.amdgcn.ds.fadd.f32broken0(i32, float, i32 immarg, i32 immarg, i1 immarg)
; This will just delete the invalid call, which isn't ideal, but these
; cases were never emitted.
; CHECK-LABEL: define void @ds_fadd_f32_invalid_not_ptr(
; CHECK-NEXT: ret void
define void @ds_fadd_f32_invalid_not_ptr(i32 %ptr, float %fval) {
%result0 = call float @llvm.amdgcn.ds.fadd.f32broken0(i32 %ptr, float %fval, i32 0, i32 0, i1 false)
ret void
}
declare float @llvm.amdgcn.ds.fadd.f32broken1(ptr addrspace(3), double, i32 immarg, i32 immarg, i1 immarg)
; CHECK-LABEL: define void @ds_fadd_f32_invalid_misatch(
; CHECK-NEXT: ret void
define void @ds_fadd_f32_invalid_misatch(ptr addrspace(3) %ptr, double %fval) {
%result0 = call float @llvm.amdgcn.ds.fadd.f32broken1(ptr addrspace(3) %ptr, double %fval, i32 0, i32 0, i1 false)
ret void
}
define double @upgrade_amdgcn_ds_fadd_f64(ptr addrspace(3) %ptr, double %val) {
; CHECK: atomicrmw fadd ptr addrspace(3) %ptr, double %val syncscope("agent") seq_cst, align 8
%result0 = call double @llvm.amdgcn.ds.fadd.f64(ptr addrspace(3) %ptr, double %val, i32 0, i32 0, i1 false)
; CHECK: = atomicrmw volatile fadd ptr addrspace(3) %ptr, double %val syncscope("agent") seq_cst, align 8
%result1 = call double @llvm.amdgcn.ds.fadd.f64(ptr addrspace(3) %ptr, double %val, i32 0, i32 0, i1 true)
; CHECK: = atomicrmw fadd ptr addrspace(3) %ptr, double %val syncscope("agent") seq_cst, align 8
%result2 = call double @llvm.amdgcn.ds.fadd.f64(ptr addrspace(3) %ptr, double %val, i32 43, i32 3, i1 false)
; CHECK: = atomicrmw fadd ptr addrspace(3) %ptr, double %val syncscope("agent") acquire, align 8
%result3 = call double @llvm.amdgcn.ds.fadd.f64(ptr addrspace(3) %ptr, double %val, i32 4, i32 2, i1 false)
ret double %result3
}
; CHECK-LABEL: @immarg_violations_ds_fadd_f64(
define void @immarg_violations_ds_fadd_f64(ptr addrspace(3) %ptr, double %fval, i32 %val32, i1 %val1) {
; CHECK: = atomicrmw volatile fadd ptr addrspace(3) %ptr, double %fval syncscope("agent") seq_cst, align 8
%result0 = call double @llvm.amdgcn.ds.fadd.f64(ptr addrspace(3) %ptr, double %fval, i32 %val32, i32 %val32, i1 %val1)
ret void
}
define <2 x half> @upgrade_amdgcn_ds_fadd_v2f16(ptr addrspace(3) %ptr, <2 x half> %val) {
; CHECK: atomicrmw fadd ptr addrspace(3) %ptr, <2 x half> %val syncscope("agent") seq_cst, align 4
%result0 = call <2 x half> @llvm.amdgcn.ds.fadd.v2f16(ptr addrspace(3) %ptr, <2 x half> %val, i32 0, i32 0, i1 false)
; CHECK: = atomicrmw volatile fadd ptr addrspace(3) %ptr, <2 x half> %val syncscope("agent") seq_cst, align 4
%result1 = call <2 x half> @llvm.amdgcn.ds.fadd.v2f16(ptr addrspace(3) %ptr, <2 x half> %val, i32 0, i32 0, i1 true)
; CHECK: = atomicrmw fadd ptr addrspace(3) %ptr, <2 x half> %val syncscope("agent") seq_cst, align 4
%result2 = call <2 x half> @llvm.amdgcn.ds.fadd.v2f16(ptr addrspace(3) %ptr, <2 x half> %val, i32 43, i32 3, i1 false)
; CHECK: = atomicrmw fadd ptr addrspace(3) %ptr, <2 x half> %val syncscope("agent") acquire, align 4
%result3 = call <2 x half> @llvm.amdgcn.ds.fadd.v2f16(ptr addrspace(3) %ptr, <2 x half> %val, i32 4, i32 2, i1 false)
ret <2 x half> %result3
}
define void @immarg_violations_ds_fadd_v2f16(ptr addrspace(3) %ptr, <2 x half> %fval, i32 %val32, i1 %val1) {
; CHECK: = atomicrmw volatile fadd ptr addrspace(3) %ptr, <2 x half> %fval syncscope("agent") seq_cst, align 4
%result0 = call <2 x half> @llvm.amdgcn.ds.fadd.v2f16(ptr addrspace(3) %ptr, <2 x half> %fval, i32 %val32, i32 %val32, i1 %val1)
ret void
}
define <2 x i16> @upgrade_amdgcn_ds_fadd_v2bf16__as_i16(ptr addrspace(3) %ptr, <2 x i16> %val) {
; CHECK: [[BC0:%[0-9]+]] = bitcast <2 x i16> %val to <2 x bfloat>
; CHECK-NEXT: [[RMW0:%[0-9]+]] = atomicrmw fadd ptr addrspace(3) %ptr, <2 x bfloat> [[BC0]] syncscope("agent") seq_cst, align 4
; CHECK-NEXT: = bitcast <2 x bfloat> [[RMW0]] to <2 x i16>
%result0 = call <2 x i16> @llvm.amdgcn.ds.fadd.v2bf16(ptr addrspace(3) %ptr, <2 x i16> %val, i32 0, i32 0, i1 false)
; CHECK: [[BC1:%[0-9]+]] = bitcast <2 x i16> %val to <2 x bfloat>
; CHECK-NEXT: [[RMW1:%[0-9]+]] = atomicrmw volatile fadd ptr addrspace(3) %ptr, <2 x bfloat> [[BC1]] syncscope("agent") seq_cst, align 4
; CHECK-NEXT: = bitcast <2 x bfloat> [[RMW1]] to <2 x i16>
%result1 = call <2 x i16> @llvm.amdgcn.ds.fadd.v2bf16(ptr addrspace(3) %ptr, <2 x i16> %val, i32 0, i32 0, i1 true)
; CHECK: [[BC2:%[0-9]+]] = bitcast <2 x i16> %val to <2 x bfloat>
; CHECK-NEXT: [[RMW2:%[0-9]+]] = atomicrmw fadd ptr addrspace(3) %ptr, <2 x bfloat> [[BC2]] syncscope("agent") seq_cst, align 4
; CHECK-NEXT: = bitcast <2 x bfloat> [[RMW2]] to <2 x i16>
%result2 = call <2 x i16> @llvm.amdgcn.ds.fadd.v2bf16(ptr addrspace(3) %ptr, <2 x i16> %val, i32 43, i32 3, i1 false)
; CHECK: [[BC3:%[0-9]+]] = bitcast <2 x i16> %val to <2 x bfloat>
; CHECK-NEXT: [[RMW3:%[0-9]+]] = atomicrmw fadd ptr addrspace(3) %ptr, <2 x bfloat> [[BC3]] syncscope("agent") acquire, align 4
; CHECK-NEXT: = bitcast <2 x bfloat> [[RMW3]] to <2 x i16>
%result3 = call <2 x i16> @llvm.amdgcn.ds.fadd.v2bf16(ptr addrspace(3) %ptr, <2 x i16> %val, i32 4, i32 2, i1 false)
ret <2 x i16> %result3
}
; Somehow the bf16 version was defined as a separate intrinsic with missing arguments.
define <2 x i16> @upgrade_amdgcn_ds_fadd_v2bf16__missing_args_as_i16(ptr addrspace(3) %ptr, <2 x i16> %val) {
; CHECK: [[BC0:%[0-9]+]] = bitcast <2 x i16> %val to <2 x bfloat>
; CHECK-NEXT: [[RMW0:%[0-9]+]] = atomicrmw fadd ptr addrspace(3) %ptr, <2 x bfloat> [[BC0]] syncscope("agent") seq_cst, align 4
; CHECK-NEXT: [[BC1:%[0-9]+]] = bitcast <2 x bfloat> [[RMW0]] to <2 x i16>
; CHECK-NEXT: ret <2 x i16> [[BC1]]
%result0 = call <2 x i16> @llvm.amdgcn.ds.fadd.v2bf16(ptr addrspace(3) %ptr, <2 x i16> %val)
ret <2 x i16> %result0
}
declare float @llvm.amdgcn.ds.fmin.f32(ptr addrspace(3) nocapture, float, i32 immarg, i32 immarg, i1 immarg)
declare double @llvm.amdgcn.ds.fmin.f64(ptr addrspace(3) nocapture, double, i32 immarg, i32 immarg, i1 immarg)
define float @upgrade_amdgcn_ds_fmin_f32(ptr addrspace(3) %ptr, float %val) {
; CHECK: atomicrmw fmin ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result0 = call float @llvm.amdgcn.ds.fmin.f32(ptr addrspace(3) %ptr, float %val, i32 0, i32 0, i1 false)
; CHECK: = atomicrmw volatile fmin ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result1 = call float @llvm.amdgcn.ds.fmin.f32(ptr addrspace(3) %ptr, float %val, i32 0, i32 0, i1 true)
; CHECK: = atomicrmw fmin ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result2 = call float @llvm.amdgcn.ds.fmin.f32(ptr addrspace(3) %ptr, float %val, i32 43, i32 3, i1 false)
; CHECK: = atomicrmw fmin ptr addrspace(3) %ptr, float %val syncscope("agent") acquire, align 4
%result3 = call float @llvm.amdgcn.ds.fmin.f32(ptr addrspace(3) %ptr, float %val, i32 4, i32 2, i1 false)
ret float %result3
}
define double @upgrade_amdgcn_ds_fmin_f64(ptr addrspace(3) %ptr, double %val) {
; CHECK: atomicrmw fmin ptr addrspace(3) %ptr, double %val syncscope("agent") seq_cst, align 8
%result0 = call double @llvm.amdgcn.ds.fmin.f64(ptr addrspace(3) %ptr, double %val, i32 0, i32 0, i1 false)
; CHECK: = atomicrmw volatile fmin ptr addrspace(3) %ptr, double %val syncscope("agent") seq_cst, align 8
%result1 = call double @llvm.amdgcn.ds.fmin.f64(ptr addrspace(3) %ptr, double %val, i32 0, i32 0, i1 true)
; CHECK: = atomicrmw fmin ptr addrspace(3) %ptr, double %val syncscope("agent") seq_cst, align 8
%result2 = call double @llvm.amdgcn.ds.fmin.f64(ptr addrspace(3) %ptr, double %val, i32 43, i32 3, i1 false)
; CHECK: = atomicrmw fmin ptr addrspace(3) %ptr, double %val syncscope("agent") acquire, align 8
%result3 = call double @llvm.amdgcn.ds.fmin.f64(ptr addrspace(3) %ptr, double %val, i32 4, i32 2, i1 false)
ret double %result3
}
declare float @llvm.amdgcn.ds.fmin(ptr addrspace(3) nocapture, float, i32 immarg, i32 immarg, i1 immarg)
define float @upgrade_amdgcn_ds_fmin_f32_no_suffix(ptr addrspace(3) %ptr, float %val) {
; CHECK: = atomicrmw fmin ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result0 = call float @llvm.amdgcn.ds.fmin(ptr addrspace(3) %ptr, float %val, i32 0, i32 0, i1 false)
ret float %result0
}
declare float @llvm.amdgcn.ds.fmax(ptr addrspace(3) nocapture, float, i32 immarg, i32 immarg, i1 immarg)
define float @upgrade_amdgcn_ds_fmax_f32_no_suffix(ptr addrspace(3) %ptr, float %val) {
; CHECK: = atomicrmw fmax ptr addrspace(3) %ptr, float %val syncscope("agent") seq_cst, align 4
%result0 = call float @llvm.amdgcn.ds.fmax(ptr addrspace(3) %ptr, float %val, i32 0, i32 0, i1 false)
ret float %result0
}
attributes #0 = { argmemonly nounwind willreturn }
|
