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
|
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=kaveri -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,CI %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GFX9 %s
; GCN-LABEL: {{^}}atomic_load_monotonic_i8:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_u8 v0, v0{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i8 @atomic_load_monotonic_i8(i8 addrspace(3)* %ptr) {
%load = load atomic i8, i8 addrspace(3)* %ptr monotonic, align 1
ret i8 %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_i8_offset:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_u8 v0, v0 offset:16{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i8 @atomic_load_monotonic_i8_offset(i8 addrspace(3)* %ptr) {
%gep = getelementptr inbounds i8, i8 addrspace(3)* %ptr, i8 16
%load = load atomic i8, i8 addrspace(3)* %gep monotonic, align 1
ret i8 %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_i16:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_u16 v0, v0{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i16 @atomic_load_monotonic_i16(i16 addrspace(3)* %ptr) {
%load = load atomic i16, i16 addrspace(3)* %ptr monotonic, align 2
ret i16 %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_i16_offset:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_u16 v0, v0 offset:32{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i16 @atomic_load_monotonic_i16_offset(i16 addrspace(3)* %ptr) {
%gep = getelementptr inbounds i16, i16 addrspace(3)* %ptr, i16 16
%load = load atomic i16, i16 addrspace(3)* %gep monotonic, align 2
ret i16 %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_i32:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_b32 v0, v0{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i32 @atomic_load_monotonic_i32(i32 addrspace(3)* %ptr) {
%load = load atomic i32, i32 addrspace(3)* %ptr monotonic, align 4
ret i32 %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_i32_offset:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_b32 v0, v0 offset:64{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i32 @atomic_load_monotonic_i32_offset(i32 addrspace(3)* %ptr) {
%gep = getelementptr inbounds i32, i32 addrspace(3)* %ptr, i32 16
%load = load atomic i32, i32 addrspace(3)* %gep monotonic, align 4
ret i32 %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_i64:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_b64 v[0:1], v0{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i64 @atomic_load_monotonic_i64(i64 addrspace(3)* %ptr) {
%load = load atomic i64, i64 addrspace(3)* %ptr monotonic, align 8
ret i64 %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_i64_offset:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_b64 v[0:1], v0 offset:128{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i64 @atomic_load_monotonic_i64_offset(i64 addrspace(3)* %ptr) {
%gep = getelementptr inbounds i64, i64 addrspace(3)* %ptr, i32 16
%load = load atomic i64, i64 addrspace(3)* %gep monotonic, align 8
ret i64 %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_f32_offset:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_b32 v0, v0 offset:64{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define float @atomic_load_monotonic_f32_offset(float addrspace(3)* %ptr) {
%gep = getelementptr inbounds float, float addrspace(3)* %ptr, i32 16
%load = load atomic float, float addrspace(3)* %gep monotonic, align 4
ret float %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_f64_offset:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_b64 v[0:1], v0 offset:128{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define double @atomic_load_monotonic_f64_offset(double addrspace(3)* %ptr) {
%gep = getelementptr inbounds double, double addrspace(3)* %ptr, i32 16
%load = load atomic double, double addrspace(3)* %gep monotonic, align 8
ret double %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_p0i8_offset:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_b64 v[0:1], v0 offset:128{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i8* @atomic_load_monotonic_p0i8_offset(i8* addrspace(3)* %ptr) {
%gep = getelementptr inbounds i8*, i8* addrspace(3)* %ptr, i32 16
%load = load atomic i8*, i8* addrspace(3)* %gep monotonic, align 8
ret i8* %load
}
; GCN-LABEL: {{^}}atomic_load_monotonic_p3i8_offset:
; GCN: s_waitcnt
; GFX9-NOT: s_mov_b32 m0
; CI-NEXT: s_mov_b32 m0
; GCN-NEXT: ds_read_b32 v0, v0 offset:64{{$}}
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_setpc_b64
define i8 addrspace(3)* @atomic_load_monotonic_p3i8_offset(i8 addrspace(3)* addrspace(3)* %ptr) {
%gep = getelementptr inbounds i8 addrspace(3)*, i8 addrspace(3)* addrspace(3)* %ptr, i32 16
%load = load atomic i8 addrspace(3)*, i8 addrspace(3)* addrspace(3)* %gep monotonic, align 4
ret i8 addrspace(3)* %load
}
|
