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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -march=amdgcn -mcpu=verde -verify-machineinstrs | FileCheck %s -check-prefix=SI
; RUN: llc < %s -march=amdgcn -mcpu=hawaii -verify-machineinstrs | FileCheck %s -check-prefix=GFX7
; RUN: llc < %s -march=amdgcn -mcpu=gfx1010 -verify-machineinstrs | FileCheck %s -check-prefix=GFX10
; RUN: llc < %s -march=amdgcn -mcpu=gfx1030 -verify-machineinstrs | FileCheck %s -check-prefix=GFX1030
; RUN: not --crash llc < %s -march=amdgcn -mcpu=gfx1100 -verify-machineinstrs 2>&1 | FileCheck %s -check-prefix=GFX11-ERR
; RUN: llc < %s -global-isel -march=amdgcn -mcpu=verde -verify-machineinstrs | FileCheck %s -check-prefix=G_SI
; RUN: llc < %s -global-isel -march=amdgcn -mcpu=hawaii -verify-machineinstrs | FileCheck %s -check-prefix=G_GFX7
; RUN: llc < %s -global-isel -march=amdgcn -mcpu=gfx1010 -verify-machineinstrs | FileCheck %s -check-prefix=G_GFX10
; RUN: llc < %s -global-isel -march=amdgcn -mcpu=gfx1030 -verify-machineinstrs | FileCheck %s -check-prefix=G_GFX1030
; RUN: not --crash llc < %s -global-isel -march=amdgcn -mcpu=gfx1100 -verify-machineinstrs 2>&1 | FileCheck %s -check-prefix=G_GFX11-ERR
; GFX11-ERR: LLVM ERROR: Cannot select: intrinsic %llvm.amdgcn.image.atomic.f
; G_GFX11-ERR: LLVM ERROR: cannot select: {{.*}} = G_AMDGPU_INTRIN_IMAGE_LOAD intrinsic(@llvm.amdgcn.image.atomic.f
declare float @llvm.amdgcn.image.atomic.fmin.1d.f32.f32(float, i32, <8 x i32>, i32, i32)
declare float @llvm.amdgcn.image.atomic.fmax.1d.f32.f32(float, i32, <8 x i32>, i32, i32)
define amdgpu_ps float @atomic_fmin_1d(<8 x i32> inreg %rsrc, float %data, i32 %s) {
; SI-LABEL: atomic_fmin_1d:
; SI: ; %bb.0: ; %main_body
; SI-NEXT: image_atomic_fmin v0, v1, s[0:7] dmask:0x1 unorm glc
; SI-NEXT: s_waitcnt vmcnt(0) expcnt(0)
; SI-NEXT: ; return to shader part epilog
;
; GFX7-LABEL: atomic_fmin_1d:
; GFX7: ; %bb.0: ; %main_body
; GFX7-NEXT: image_atomic_fmin v0, v1, s[0:7] dmask:0x1 unorm glc
; GFX7-NEXT: s_waitcnt vmcnt(0)
; GFX7-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: atomic_fmin_1d:
; GFX10: ; %bb.0: ; %main_body
; GFX10-NEXT: image_atomic_fmin v0, v1, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: ; return to shader part epilog
;
; GFX1030-LABEL: atomic_fmin_1d:
; GFX1030: ; %bb.0: ; %main_body
; GFX1030-NEXT: image_atomic_fmin v0, v1, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc
; GFX1030-NEXT: s_waitcnt vmcnt(0)
; GFX1030-NEXT: ; return to shader part epilog
;
; G_SI-LABEL: atomic_fmin_1d:
; G_SI: ; %bb.0: ; %main_body
; G_SI-NEXT: image_atomic_fmin v0, v1, s[0:7] dmask:0x1 unorm glc
; G_SI-NEXT: s_waitcnt vmcnt(0) expcnt(0)
; G_SI-NEXT: ; return to shader part epilog
;
; G_GFX7-LABEL: atomic_fmin_1d:
; G_GFX7: ; %bb.0: ; %main_body
; G_GFX7-NEXT: image_atomic_fmin v0, v1, s[0:7] dmask:0x1 unorm glc
; G_GFX7-NEXT: s_waitcnt vmcnt(0)
; G_GFX7-NEXT: ; return to shader part epilog
;
; G_GFX10-LABEL: atomic_fmin_1d:
; G_GFX10: ; %bb.0: ; %main_body
; G_GFX10-NEXT: image_atomic_fmin v0, v1, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc
; G_GFX10-NEXT: s_waitcnt vmcnt(0)
; G_GFX10-NEXT: ; return to shader part epilog
;
; G_GFX1030-LABEL: atomic_fmin_1d:
; G_GFX1030: ; %bb.0: ; %main_body
; G_GFX1030-NEXT: image_atomic_fmin v0, v1, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc
; G_GFX1030-NEXT: s_waitcnt vmcnt(0)
; G_GFX1030-NEXT: ; return to shader part epilog
main_body:
%v = call float @llvm.amdgcn.image.atomic.fmin.1d.f32.f32(float %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
ret float %v
}
define amdgpu_ps float @atomic_fmax_1d(<8 x i32> inreg %rsrc, float %data, i32 %s) {
; SI-LABEL: atomic_fmax_1d:
; SI: ; %bb.0: ; %main_body
; SI-NEXT: image_atomic_fmax v0, v1, s[0:7] dmask:0x1 unorm glc
; SI-NEXT: s_waitcnt vmcnt(0) expcnt(0)
; SI-NEXT: ; return to shader part epilog
;
; GFX7-LABEL: atomic_fmax_1d:
; GFX7: ; %bb.0: ; %main_body
; GFX7-NEXT: image_atomic_fmax v0, v1, s[0:7] dmask:0x1 unorm glc
; GFX7-NEXT: s_waitcnt vmcnt(0)
; GFX7-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: atomic_fmax_1d:
; GFX10: ; %bb.0: ; %main_body
; GFX10-NEXT: image_atomic_fmax v0, v1, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: ; return to shader part epilog
;
; GFX1030-LABEL: atomic_fmax_1d:
; GFX1030: ; %bb.0: ; %main_body
; GFX1030-NEXT: image_atomic_fmax v0, v1, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc
; GFX1030-NEXT: s_waitcnt vmcnt(0)
; GFX1030-NEXT: ; return to shader part epilog
;
; G_SI-LABEL: atomic_fmax_1d:
; G_SI: ; %bb.0: ; %main_body
; G_SI-NEXT: image_atomic_fmax v0, v1, s[0:7] dmask:0x1 unorm glc
; G_SI-NEXT: s_waitcnt vmcnt(0) expcnt(0)
; G_SI-NEXT: ; return to shader part epilog
;
; G_GFX7-LABEL: atomic_fmax_1d:
; G_GFX7: ; %bb.0: ; %main_body
; G_GFX7-NEXT: image_atomic_fmax v0, v1, s[0:7] dmask:0x1 unorm glc
; G_GFX7-NEXT: s_waitcnt vmcnt(0)
; G_GFX7-NEXT: ; return to shader part epilog
;
; G_GFX10-LABEL: atomic_fmax_1d:
; G_GFX10: ; %bb.0: ; %main_body
; G_GFX10-NEXT: image_atomic_fmax v0, v1, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc
; G_GFX10-NEXT: s_waitcnt vmcnt(0)
; G_GFX10-NEXT: ; return to shader part epilog
;
; G_GFX1030-LABEL: atomic_fmax_1d:
; G_GFX1030: ; %bb.0: ; %main_body
; G_GFX1030-NEXT: image_atomic_fmax v0, v1, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm glc
; G_GFX1030-NEXT: s_waitcnt vmcnt(0)
; G_GFX1030-NEXT: ; return to shader part epilog
main_body:
%v = call float @llvm.amdgcn.image.atomic.fmax.1d.f32.f32(float %data, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
ret float %v
}
|
