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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn -mcpu=gfx1200 -verify-machineinstrs < %s | FileCheck %s -check-prefix=GFX12
define float @raw_buffer_atomic_cond_sub_return(<4 x i32> inreg %rsrc, i32 inreg %data) #0 {
; GFX12-LABEL: raw_buffer_atomic_cond_sub_return:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v0, s16
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v0, off, s[0:3], null th:TH_ATOMIC_RETURN
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%orig = call i32 @llvm.amdgcn.raw.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 0)
%r = bitcast i32 %orig to float
ret float %r
}
define void @raw_buffer_atomic_cond_sub_no_return(<4 x i32> inreg %rsrc, i32 inreg %data) #0 {
; GFX12-LABEL: raw_buffer_atomic_cond_sub_no_return:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v0, s16
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v0, off, s[0:3], null th:TH_ATOMIC_RETURN
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%unused = call i32 @llvm.amdgcn.raw.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 0)
ret void
}
define void @raw_buffer_atomic_cond_sub_no_return_forced(<4 x i32> inreg %rsrc, i32 inreg %data) #1 {
; GFX12-LABEL: raw_buffer_atomic_cond_sub_no_return_forced:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v0, s16
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v0, off, s[0:3], null
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%unused = call i32 @llvm.amdgcn.raw.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 0)
ret void
}
define float @raw_buffer_atomic_cond_sub_imm_soff_return(<4 x i32> inreg %rsrc, i32 inreg %data) #0 {
; GFX12-LABEL: raw_buffer_atomic_cond_sub_imm_soff_return:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v0, s16
; GFX12-NEXT: s_mov_b32 s4, 4
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v0, off, s[0:3], s4 th:TH_ATOMIC_RETURN
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_wait_alu 0xfffe
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%orig = call i32 @llvm.amdgcn.raw.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 4, i32 0)
%r = bitcast i32 %orig to float
ret float %r
}
define void @raw_buffer_atomic_cond_sub_imm_soff_no_return(<4 x i32> inreg %rsrc, i32 inreg %data) #0 {
; GFX12-LABEL: raw_buffer_atomic_cond_sub_imm_soff_no_return:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v0, s16
; GFX12-NEXT: s_mov_b32 s4, 4
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v0, off, s[0:3], s4 th:TH_ATOMIC_RETURN
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_wait_alu 0xfffe
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%unused = call i32 @llvm.amdgcn.raw.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 4, i32 0)
ret void
}
define void @raw_buffer_atomic_cond_sub_imm_soff_no_return_forced(<4 x i32> inreg %rsrc, i32 inreg %data) #1 {
; GFX12-LABEL: raw_buffer_atomic_cond_sub_imm_soff_no_return_forced:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v0, s16
; GFX12-NEXT: s_mov_b32 s4, 4
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v0, off, s[0:3], s4
; GFX12-NEXT: s_wait_alu 0xfffe
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%unused = call i32 @llvm.amdgcn.raw.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 4, i32 0)
ret void
}
define float @struct_buffer_atomic_cond_sub_return(<4 x i32> inreg %rsrc, i32 inreg %data) #0 {
; GFX12-LABEL: struct_buffer_atomic_cond_sub_return:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v0, s16
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v0, v1, s[0:3], null idxen th:TH_ATOMIC_RETURN
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%orig = call i32 @llvm.amdgcn.struct.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 0, i32 0)
%r = bitcast i32 %orig to float
ret float %r
}
define void @struct_buffer_atomic_cond_sub_no_return(<4 x i32> inreg %rsrc, i32 inreg %data) #0 {
; GFX12-LABEL: struct_buffer_atomic_cond_sub_no_return:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s16
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v1, v0, s[0:3], null idxen th:TH_ATOMIC_RETURN
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%unused = call i32 @llvm.amdgcn.struct.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 0, i32 0)
ret void
}
define void @struct_buffer_atomic_cond_sub_no_return_forced(<4 x i32> inreg %rsrc, i32 inreg %data) #1 {
; GFX12-LABEL: struct_buffer_atomic_cond_sub_no_return_forced:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s16
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v1, v0, s[0:3], null idxen
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%unused = call i32 @llvm.amdgcn.struct.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 0, i32 0)
ret void
}
define float @struct_buffer_atomic_cond_sub_imm_soff_return(<4 x i32> inreg %rsrc, i32 inreg %data) #0 {
; GFX12-LABEL: struct_buffer_atomic_cond_sub_imm_soff_return:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_mov_b32 v0, s16
; GFX12-NEXT: s_mov_b32 s4, 4
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v0, v1, s[0:3], s4 idxen th:TH_ATOMIC_RETURN
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_wait_alu 0xfffe
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%orig = call i32 @llvm.amdgcn.struct.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 4, i32 0)
%r = bitcast i32 %orig to float
ret float %r
}
define void @struct_buffer_atomic_cond_sub_imm_soff_no_return(<4 x i32> inreg %rsrc, i32 inreg %data) #0 {
; GFX12-LABEL: struct_buffer_atomic_cond_sub_imm_soff_no_return:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s16
; GFX12-NEXT: s_mov_b32 s4, 4
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v1, v0, s[0:3], s4 idxen th:TH_ATOMIC_RETURN
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_wait_alu 0xfffe
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%unused = call i32 @llvm.amdgcn.struct.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 4, i32 0)
ret void
}
define void @struct_buffer_atomic_cond_sub_imm_soff_no_return_forced(<4 x i32> inreg %rsrc, i32 inreg %data) #1 {
; GFX12-LABEL: struct_buffer_atomic_cond_sub_imm_soff_no_return_forced:
; GFX12: ; %bb.0: ; %main_body
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, s16
; GFX12-NEXT: s_mov_b32 s4, 4
; GFX12-NEXT: buffer_atomic_cond_sub_u32 v1, v0, s[0:3], s4 idxen
; GFX12-NEXT: s_wait_alu 0xfffe
; GFX12-NEXT: s_setpc_b64 s[30:31]
main_body:
%unused = call i32 @llvm.amdgcn.struct.buffer.atomic.cond.sub.u32.i32(i32 %data, <4 x i32> %rsrc, i32 0, i32 0, i32 4, i32 0)
ret void
}
declare i32 @llvm.amdgcn.raw.buffer.atomic.cond.sub.u32.i32(i32, <4 x i32>, i32, i32, i32) #0
declare i32 @llvm.amdgcn.struct.buffer.atomic.cond.sub.u32.i32(i32, <4 x i32>, i32, i32, i32, i32) #0
attributes #0 = { nounwind }
attributes #1 = { nounwind "target-features"="+atomic-csub-no-rtn-insts" }
|
