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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -global-isel -mtriple=amdgcn-mesa-mesa3d -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -check-prefix=GFX6 %s
; RUN: llc -global-isel -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1010 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX10PLUS,GFX10 %s
; RUN: llc -global-isel -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1100 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX10PLUS,GFX11 %s
define amdgpu_ps <4 x float> @load_3d_v4f32_xyzw(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %r) {
; GFX6-LABEL: load_3d_v4f32_xyzw:
; GFX6: ; %bb.0:
; GFX6-NEXT: s_mov_b32 s0, s2
; GFX6-NEXT: s_mov_b32 s1, s3
; GFX6-NEXT: s_mov_b32 s2, s4
; GFX6-NEXT: s_mov_b32 s3, s5
; GFX6-NEXT: s_mov_b32 s4, s6
; GFX6-NEXT: s_mov_b32 s5, s7
; GFX6-NEXT: s_mov_b32 s6, s8
; GFX6-NEXT: s_mov_b32 s7, s9
; GFX6-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
; GFX6-NEXT: s_waitcnt vmcnt(0)
; GFX6-NEXT: ; return to shader part epilog
;
; GFX10PLUS-LABEL: load_3d_v4f32_xyzw:
; GFX10PLUS: ; %bb.0:
; GFX10PLUS-NEXT: s_mov_b32 s0, s2
; GFX10PLUS-NEXT: s_mov_b32 s1, s3
; GFX10PLUS-NEXT: s_mov_b32 s2, s4
; GFX10PLUS-NEXT: s_mov_b32 s3, s5
; GFX10PLUS-NEXT: s_mov_b32 s4, s6
; GFX10PLUS-NEXT: s_mov_b32 s5, s7
; GFX10PLUS-NEXT: s_mov_b32 s6, s8
; GFX10PLUS-NEXT: s_mov_b32 s7, s9
; GFX10PLUS-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm
; GFX10PLUS-NEXT: s_waitcnt vmcnt(0)
; GFX10PLUS-NEXT: ; return to shader part epilog
%v = call <4 x float> @llvm.amdgcn.image.load.3d.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %r, <8 x i32> %rsrc, i32 0, i32 0)
ret <4 x float> %v
}
define amdgpu_ps <4 x float> @load_3d_v4f32_xyzw_tfe(<8 x i32> inreg %rsrc, ptr addrspace(1) inreg %out, i32 %s, i32 %t, i32 %r) {
; GFX6-LABEL: load_3d_v4f32_xyzw_tfe:
; GFX6: ; %bb.0:
; GFX6-NEXT: v_mov_b32_e32 v5, v0
; GFX6-NEXT: v_mov_b32_e32 v0, 0
; GFX6-NEXT: s_mov_b32 s0, s2
; GFX6-NEXT: s_mov_b32 s1, s3
; GFX6-NEXT: s_mov_b32 s2, s4
; GFX6-NEXT: s_mov_b32 s3, s5
; GFX6-NEXT: s_mov_b32 s4, s6
; GFX6-NEXT: s_mov_b32 s5, s7
; GFX6-NEXT: s_mov_b32 s6, s8
; GFX6-NEXT: s_mov_b32 s7, s9
; GFX6-NEXT: v_mov_b32_e32 v6, v1
; GFX6-NEXT: v_mov_b32_e32 v7, v2
; GFX6-NEXT: v_mov_b32_e32 v1, v0
; GFX6-NEXT: v_mov_b32_e32 v2, v0
; GFX6-NEXT: v_mov_b32_e32 v3, v0
; GFX6-NEXT: v_mov_b32_e32 v4, v0
; GFX6-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe
; GFX6-NEXT: s_mov_b32 s8, s10
; GFX6-NEXT: s_mov_b32 s9, s11
; GFX6-NEXT: s_mov_b32 s10, -1
; GFX6-NEXT: s_mov_b32 s11, 0xf000
; GFX6-NEXT: s_waitcnt vmcnt(0)
; GFX6-NEXT: buffer_store_dword v4, off, s[8:11], 0
; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0)
; GFX6-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: load_3d_v4f32_xyzw_tfe:
; GFX10: ; %bb.0:
; GFX10-NEXT: v_mov_b32_e32 v8, 0
; GFX10-NEXT: v_mov_b32_e32 v5, v0
; GFX10-NEXT: v_mov_b32_e32 v6, v1
; GFX10-NEXT: v_mov_b32_e32 v7, v2
; GFX10-NEXT: s_mov_b32 s0, s2
; GFX10-NEXT: v_mov_b32_e32 v9, v8
; GFX10-NEXT: v_mov_b32_e32 v10, v8
; GFX10-NEXT: v_mov_b32_e32 v11, v8
; GFX10-NEXT: v_mov_b32_e32 v12, v8
; GFX10-NEXT: s_mov_b32 s1, s3
; GFX10-NEXT: s_mov_b32 s2, s4
; GFX10-NEXT: s_mov_b32 s3, s5
; GFX10-NEXT: s_mov_b32 s4, s6
; GFX10-NEXT: s_mov_b32 s5, s7
; GFX10-NEXT: s_mov_b32 s6, s8
; GFX10-NEXT: s_mov_b32 s7, s9
; GFX10-NEXT: v_mov_b32_e32 v0, v8
; GFX10-NEXT: v_mov_b32_e32 v1, v9
; GFX10-NEXT: v_mov_b32_e32 v2, v10
; GFX10-NEXT: v_mov_b32_e32 v3, v11
; GFX10-NEXT: v_mov_b32_e32 v4, v12
; GFX10-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm tfe
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: global_store_dword v8, v4, s[10:11]
; GFX10-NEXT: ; return to shader part epilog
;
; GFX11-LABEL: load_3d_v4f32_xyzw_tfe:
; GFX11: ; %bb.0:
; GFX11-NEXT: v_dual_mov_b32 v7, v2 :: v_dual_mov_b32 v8, 0
; GFX11-NEXT: v_dual_mov_b32 v5, v0 :: v_dual_mov_b32 v6, v1
; GFX11-NEXT: s_mov_b32 s0, s2
; GFX11-NEXT: s_mov_b32 s1, s3
; GFX11-NEXT: s_delay_alu instid0(VALU_DEP_2)
; GFX11-NEXT: v_mov_b32_e32 v9, v8
; GFX11-NEXT: v_mov_b32_e32 v10, v8
; GFX11-NEXT: v_mov_b32_e32 v11, v8
; GFX11-NEXT: v_mov_b32_e32 v12, v8
; GFX11-NEXT: s_mov_b32 s2, s4
; GFX11-NEXT: s_mov_b32 s3, s5
; GFX11-NEXT: s_mov_b32 s4, s6
; GFX11-NEXT: s_mov_b32 s5, s7
; GFX11-NEXT: s_mov_b32 s6, s8
; GFX11-NEXT: s_mov_b32 s7, s9
; GFX11-NEXT: v_dual_mov_b32 v0, v8 :: v_dual_mov_b32 v1, v9
; GFX11-NEXT: v_dual_mov_b32 v2, v10 :: v_dual_mov_b32 v3, v11
; GFX11-NEXT: v_mov_b32_e32 v4, v12
; GFX11-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm tfe
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: global_store_b32 v8, v4, s[10:11]
; GFX11-NEXT: ; return to shader part epilog
%v = call { <4 x float>, i32 } @llvm.amdgcn.image.load.3d.sl_v4f32i32s.i32(i32 15, i32 %s, i32 %t, i32 %r, <8 x i32> %rsrc, i32 1, i32 0)
%v.vec = extractvalue { <4 x float>, i32 } %v, 0
%v.err = extractvalue { <4 x float>, i32 } %v, 1
store i32 %v.err, ptr addrspace(1) %out, align 4
ret <4 x float> %v.vec
}
define amdgpu_ps <4 x float> @load_3d_v4f32_xyzw_tfe_lwe(<8 x i32> inreg %rsrc, ptr addrspace(1) inreg %out, i32 %s, i32 %t, i32 %r) {
; GFX6-LABEL: load_3d_v4f32_xyzw_tfe_lwe:
; GFX6: ; %bb.0:
; GFX6-NEXT: v_mov_b32_e32 v5, v0
; GFX6-NEXT: v_mov_b32_e32 v0, 0
; GFX6-NEXT: s_mov_b32 s0, s2
; GFX6-NEXT: s_mov_b32 s1, s3
; GFX6-NEXT: s_mov_b32 s2, s4
; GFX6-NEXT: s_mov_b32 s3, s5
; GFX6-NEXT: s_mov_b32 s4, s6
; GFX6-NEXT: s_mov_b32 s5, s7
; GFX6-NEXT: s_mov_b32 s6, s8
; GFX6-NEXT: s_mov_b32 s7, s9
; GFX6-NEXT: v_mov_b32_e32 v6, v1
; GFX6-NEXT: v_mov_b32_e32 v7, v2
; GFX6-NEXT: v_mov_b32_e32 v1, v0
; GFX6-NEXT: v_mov_b32_e32 v2, v0
; GFX6-NEXT: v_mov_b32_e32 v3, v0
; GFX6-NEXT: v_mov_b32_e32 v4, v0
; GFX6-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe lwe
; GFX6-NEXT: s_mov_b32 s8, s10
; GFX6-NEXT: s_mov_b32 s9, s11
; GFX6-NEXT: s_mov_b32 s10, -1
; GFX6-NEXT: s_mov_b32 s11, 0xf000
; GFX6-NEXT: s_waitcnt vmcnt(0)
; GFX6-NEXT: buffer_store_dword v4, off, s[8:11], 0
; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0)
; GFX6-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: load_3d_v4f32_xyzw_tfe_lwe:
; GFX10: ; %bb.0:
; GFX10-NEXT: v_mov_b32_e32 v8, 0
; GFX10-NEXT: v_mov_b32_e32 v5, v0
; GFX10-NEXT: v_mov_b32_e32 v6, v1
; GFX10-NEXT: v_mov_b32_e32 v7, v2
; GFX10-NEXT: s_mov_b32 s0, s2
; GFX10-NEXT: v_mov_b32_e32 v9, v8
; GFX10-NEXT: v_mov_b32_e32 v10, v8
; GFX10-NEXT: v_mov_b32_e32 v11, v8
; GFX10-NEXT: v_mov_b32_e32 v12, v8
; GFX10-NEXT: s_mov_b32 s1, s3
; GFX10-NEXT: s_mov_b32 s2, s4
; GFX10-NEXT: s_mov_b32 s3, s5
; GFX10-NEXT: s_mov_b32 s4, s6
; GFX10-NEXT: s_mov_b32 s5, s7
; GFX10-NEXT: s_mov_b32 s6, s8
; GFX10-NEXT: s_mov_b32 s7, s9
; GFX10-NEXT: v_mov_b32_e32 v0, v8
; GFX10-NEXT: v_mov_b32_e32 v1, v9
; GFX10-NEXT: v_mov_b32_e32 v2, v10
; GFX10-NEXT: v_mov_b32_e32 v3, v11
; GFX10-NEXT: v_mov_b32_e32 v4, v12
; GFX10-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm tfe lwe
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: global_store_dword v8, v4, s[10:11]
; GFX10-NEXT: ; return to shader part epilog
;
; GFX11-LABEL: load_3d_v4f32_xyzw_tfe_lwe:
; GFX11: ; %bb.0:
; GFX11-NEXT: v_dual_mov_b32 v7, v2 :: v_dual_mov_b32 v8, 0
; GFX11-NEXT: v_dual_mov_b32 v5, v0 :: v_dual_mov_b32 v6, v1
; GFX11-NEXT: s_mov_b32 s0, s2
; GFX11-NEXT: s_mov_b32 s1, s3
; GFX11-NEXT: s_delay_alu instid0(VALU_DEP_2)
; GFX11-NEXT: v_mov_b32_e32 v9, v8
; GFX11-NEXT: v_mov_b32_e32 v10, v8
; GFX11-NEXT: v_mov_b32_e32 v11, v8
; GFX11-NEXT: v_mov_b32_e32 v12, v8
; GFX11-NEXT: s_mov_b32 s2, s4
; GFX11-NEXT: s_mov_b32 s3, s5
; GFX11-NEXT: s_mov_b32 s4, s6
; GFX11-NEXT: s_mov_b32 s5, s7
; GFX11-NEXT: s_mov_b32 s6, s8
; GFX11-NEXT: s_mov_b32 s7, s9
; GFX11-NEXT: v_dual_mov_b32 v0, v8 :: v_dual_mov_b32 v1, v9
; GFX11-NEXT: v_dual_mov_b32 v2, v10 :: v_dual_mov_b32 v3, v11
; GFX11-NEXT: v_mov_b32_e32 v4, v12
; GFX11-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm tfe lwe
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: global_store_b32 v8, v4, s[10:11]
; GFX11-NEXT: ; return to shader part epilog
%v = call { <4 x float>, i32 } @llvm.amdgcn.image.load.3d.sl_v4f32i32s.i32(i32 15, i32 %s, i32 %t, i32 %r, <8 x i32> %rsrc, i32 3, i32 0)
%v.vec = extractvalue { <4 x float>, i32 } %v, 0
%v.err = extractvalue { <4 x float>, i32 } %v, 1
store i32 %v.err, ptr addrspace(1) %out, align 4
ret <4 x float> %v.vec
}
declare <4 x float> @llvm.amdgcn.image.load.3d.v4f32.i32(i32 immarg, i32, i32, i32, <8 x i32>, i32 immarg, i32 immarg) #0
declare { <4 x float>, i32 } @llvm.amdgcn.image.load.3d.sl_v4f32i32s.i32(i32 immarg, i32, i32, i32, <8 x i32>, i32 immarg, i32 immarg) #0
attributes #0 = { nounwind readonly }
|
