Skip to content

G
glslang
Unverified Commit 7923b3e3 by John Kessenich Committed by GitHub
Options

Merge pull request #2353 from vkushwaha-nv/SPV_EXT_shader_atomic_float

Add changes for SPV_EXT_shader_atomic_float_add
parents 9165cbb4 79b93923
Showing with 1111 additions and 5 deletions
SPIRV/GLSL.ext.EXT.h
......@@ -35,5 +35,6 @@ static const char* const E_SPV_EXT_shader_viewport_index_layer = "SPV_EXT_shade
static const char* const E_SPV_EXT_fragment_fully_covered = "SPV_EXT_fragment_fully_covered";
static const char* const E_SPV_EXT_fragment_invocation_density = "SPV_EXT_fragment_invocation_density";
static const char* const E_SPV_EXT_demote_to_helper_invocation = "SPV_EXT_demote_to_helper_invocation";
static const char* const E_SPV_EXT_shader_atomic_float_add = "SPV_EXT_shader_atomic_float_add";
#endif // #ifndef GLSLextEXT_H
SPIRV/GlslangToSpv.cpp
......@@ -6682,6 +6682,14 @@ spv::Id TGlslangToSpvTraverser::createAtomicOperation(glslang::TOperator op, spv
case glslang::EOpImageAtomicAdd:
case glslang::EOpAtomicCounterAdd:
opCode = spv::OpAtomicIAdd;
if (typeProxy == glslang::EbtFloat || typeProxy == glslang::EbtDouble) {
opCode = spv::OpAtomicFAddEXT;
builder.addExtension(spv::E_SPV_EXT_shader_atomic_float_add);
if (typeProxy == glslang::EbtFloat)
builder.addCapability(spv::CapabilityAtomicFloat32AddEXT);
else
builder.addCapability(spv::CapabilityAtomicFloat64AddEXT);
}
break;
case glslang::EOpAtomicCounterSubtract:
opCode = spv::OpAtomicISub;
......
SPIRV/doc.cpp
......@@ -938,6 +938,9 @@ const char* CapabilityString(int info)
case CapabilityIntegerFunctions2INTEL: return "CapabilityIntegerFunctions2INTEL";
case CapabilityAtomicFloat32AddEXT: return "AtomicFloat32AddEXT";
case CapabilityAtomicFloat64AddEXT: return "AtomicFloat64AddEXT";
default: return "Bad";
}
}
......@@ -1313,6 +1316,8 @@ const char* OpcodeString(int op)
case 4430: return "OpSubgroupAllEqualKHR";
case 4432: return "OpSubgroupReadInvocationKHR";
case OpAtomicFAddEXT: return "OpAtomicFAddEXT";
case 5000: return "OpGroupIAddNonUniformAMD";
case 5001: return "OpGroupFAddNonUniformAMD";
case 5002: return "OpGroupFMinNonUniformAMD";
......@@ -2260,6 +2265,11 @@ void Parameterize()
InstructionDesc[OpAtomicIAdd].operands.push(OperandMemorySemantics, "'Semantics'");
InstructionDesc[OpAtomicIAdd].operands.push(OperandId, "'Value'");
InstructionDesc[OpAtomicFAddEXT].operands.push(OperandId, "'Pointer'");
InstructionDesc[OpAtomicFAddEXT].operands.push(OperandScope, "'Scope'");
InstructionDesc[OpAtomicFAddEXT].operands.push(OperandMemorySemantics, "'Semantics'");
InstructionDesc[OpAtomicFAddEXT].operands.push(OperandId, "'Value'");
InstructionDesc[OpAtomicISub].operands.push(OperandId, "'Pointer'");
InstructionDesc[OpAtomicISub].operands.push(OperandScope, "'Scope'");
InstructionDesc[OpAtomicISub].operands.push(OperandMemorySemantics, "'Semantics'");
......
Test/baseResults/spv.atomicFloat.comp.out 0 → 100755
spv.atomicFloat.comp
// Module Version 10000
// Generated by (magic number): 8000a
// Id's are bound by 470
Capability Shader
Capability Float64
Capability ImageCubeArray
Capability ImageRect
Capability Image1D
Capability VulkanMemoryModelKHR
Capability VulkanMemoryModelDeviceScopeKHR
Capability AtomicFloat32AddEXT
Capability AtomicFloat64AddEXT
Extension "SPV_EXT_shader_atomic_float_add"
Extension "SPV_KHR_vulkan_memory_model"
1: ExtInstImport "GLSL.std.450"
MemoryModel Logical VulkanKHR
EntryPoint GLCompute 4 "main"
ExecutionMode 4 LocalSize 16 16 1
Source GLSL 450
SourceExtension "GL_EXT_shader_atomic_float"
SourceExtension "GL_KHR_memory_scope_semantics"
Name 4 "main"
Name 8 "resultf"
Name 11 "atomf"
Name 25 "Buffer"
MemberName 25(Buffer) 0 "dataf"
MemberName 25(Buffer) 1 "datad"
Name 27 "buf"
Name 34 "resultd"
Name 37 "atomd"
Name 143 "fimage1D"
Name 189 "fimage1DArray"
Name 232 "fimage2D"
Name 270 "fimage2DRect"
Name 308 "fimage2DArray"
Name 352 "fimageCube"
Name 392 "fimageCubeArray"
Name 430 "fimage3D"
MemberDecorate 25(Buffer) 0 Offset 0
MemberDecorate 25(Buffer) 1 Offset 8
Decorate 25(Buffer) BufferBlock
Decorate 27(buf) DescriptorSet 0
Decorate 27(buf) Binding 0
Decorate 143(fimage1D) Location 0
Decorate 143(fimage1D) DescriptorSet 0
Decorate 143(fimage1D) Binding 0
Decorate 189(fimage1DArray) Location 1
Decorate 189(fimage1DArray) DescriptorSet 0
Decorate 189(fimage1DArray) Binding 1
Decorate 232(fimage2D) Location 2
Decorate 232(fimage2D) DescriptorSet 0
Decorate 232(fimage2D) Binding 2
Decorate 270(fimage2DRect) Location 4
Decorate 270(fimage2DRect) DescriptorSet 0
Decorate 270(fimage2DRect) Binding 4
Decorate 308(fimage2DArray) Location 3
Decorate 308(fimage2DArray) DescriptorSet 0
Decorate 308(fimage2DArray) Binding 3
Decorate 352(fimageCube) Location 5
Decorate 352(fimageCube) DescriptorSet 0
Decorate 352(fimageCube) Binding 5
Decorate 392(fimageCubeArray) Location 6
Decorate 392(fimageCubeArray) DescriptorSet 0
Decorate 392(fimageCubeArray) Binding 6
Decorate 430(fimage3D) Location 7
Decorate 430(fimage3D) DescriptorSet 0
Decorate 430(fimage3D) Binding 9
Decorate 469 BuiltIn WorkgroupSize
2: TypeVoid
3: TypeFunction 2
6: TypeFloat 32
7: TypePointer Function 6(float)
9: 6(float) Constant 0
10: TypePointer Workgroup 6(float)
11(atomf): 10(ptr) Variable Workgroup
12: 6(float) Constant 1077936128
13: TypeInt 32 0
14: 13(int) Constant 5
15: 13(int) Constant 0
17: 6(float) Constant 1083179008
18: TypeInt 32 1
19: 18(int) Constant 1
20: 18(int) Constant 64
21: 18(int) Constant 0
22: 13(int) Constant 64
24: TypeFloat 64
25(Buffer): TypeStruct 6(float) 24(float64_t)
26: TypePointer Uniform 25(Buffer)
27(buf): 26(ptr) Variable Uniform
28: TypePointer Uniform 6(float)
33: TypePointer Function 24(float64_t)
35:24(float64_t) Constant 0 0
36: TypePointer Workgroup 24(float64_t)
37(atomd): 36(ptr) Variable Workgroup
38:24(float64_t) Constant 0 1074266112
40:24(float64_t) Constant 0 1074921472
42: TypePointer Uniform 24(float64_t)
57: 18(int) Constant 256
58: 13(int) Constant 256
141: TypeImage 6(float) 1D nonsampled format:R32f
142: TypePointer UniformConstant 141
143(fimage1D): 142(ptr) Variable UniformConstant
144: 6(float) Constant 1073741824
145: TypePointer Image 6(float)
147: 13(int) Constant 32768
149: 13(int) Constant 2
155: 18(int) Constant 2048
157: 13(int) Constant 2048
164: 6(float) Constant 1082130432
179: 18(int) Constant 2
187: TypeImage 6(float) 1D array nonsampled format:R32f
188: TypePointer UniformConstant 187
189(fimage1DArray): 188(ptr) Variable UniformConstant
190: TypeVector 18(int) 2
191: 190(ivec2) ConstantComposite 21 21
199: 190(ivec2) ConstantComposite 19 19
207: 190(ivec2) ConstantComposite 19 21
222: 190(ivec2) ConstantComposite 179 179
230: TypeImage 6(float) 2D nonsampled format:R32f
231: TypePointer UniformConstant 230
232(fimage2D): 231(ptr) Variable UniformConstant
268: TypeImage 6(float) Rect nonsampled format:R32f
269: TypePointer UniformConstant 268
270(fimage2DRect): 269(ptr) Variable UniformConstant
306: TypeImage 6(float) 2D array nonsampled format:R32f
307: TypePointer UniformConstant 306
308(fimage2DArray): 307(ptr) Variable UniformConstant
309: TypeVector 18(int) 3
310: 309(ivec3) ConstantComposite 21 21 21
318: 309(ivec3) ConstantComposite 19 19 21
326: 309(ivec3) ConstantComposite 19 21 19
334: 309(ivec3) ConstantComposite 19 19 19
342: 309(ivec3) ConstantComposite 179 179 21
350: TypeImage 6(float) Cube nonsampled format:R32f
351: TypePointer UniformConstant 350
352(fimageCube): 351(ptr) Variable UniformConstant
367: 309(ivec3) ConstantComposite 19 21 21
382: 309(ivec3) ConstantComposite 179 179 19
390: TypeImage 6(float) Cube array nonsampled format:R32f
391: TypePointer UniformConstant 390
392(fimageCubeArray): 391(ptr) Variable UniformConstant
428: TypeImage 6(float) 3D nonsampled format:R32f
429: TypePointer UniformConstant 428
430(fimage3D): 429(ptr) Variable UniformConstant
466: TypeVector 13(int) 3
467: 13(int) Constant 16
468: 13(int) Constant 1
469: 466(ivec3) ConstantComposite 467 467 468
4(main): 2 Function None 3
5: Label
8(resultf): 7(ptr) Variable Function
34(resultd): 33(ptr) Variable Function
Store 8(resultf) 9
16: 6(float) AtomicFAddEXT 11(atomf) 14 15 12
Store 8(resultf) 16
23: 6(float) AtomicFAddEXT 11(atomf) 19 22 17
Store 8(resultf) 23
29: 28(ptr) AccessChain 27(buf) 21
30: 6(float) AtomicFAddEXT 29 14 15 12
Store 8(resultf) 30
31: 28(ptr) AccessChain 27(buf) 21
32: 6(float) AtomicFAddEXT 31 19 22 17
Store 8(resultf) 32
Store 34(resultd) 35
39:24(float64_t) AtomicFAddEXT 37(atomd) 14 15 38
Store 34(resultd) 39
41:24(float64_t) AtomicFAddEXT 37(atomd) 19 22 40
Store 34(resultd) 41
43: 42(ptr) AccessChain 27(buf) 19
44:24(float64_t) AtomicFAddEXT 43 14 15 38
Store 34(resultd) 44
45: 42(ptr) AccessChain 27(buf) 19
46:24(float64_t) AtomicFAddEXT 45 19 22 40
Store 34(resultd) 46
47: 28(ptr) AccessChain 27(buf) 21
48: 6(float) Load 8(resultf)
49: 6(float) AtomicExchange 47 14 15 48
Store 8(resultf) 49
50: 6(float) Load 8(resultf)
51: 28(ptr) AccessChain 27(buf) 21
52: 6(float) Load 51
53: 6(float) FAdd 52 50
54: 28(ptr) AccessChain 27(buf) 21
Store 54 53
55: 28(ptr) AccessChain 27(buf) 21
56: 6(float) Load 8(resultf)
59: 6(float) AtomicExchange 55 19 58 56
Store 8(resultf) 59
60: 6(float) Load 8(resultf)
61: 28(ptr) AccessChain 27(buf) 21
62: 6(float) Load 61
63: 6(float) FAdd 62 60
64: 28(ptr) AccessChain 27(buf) 21
Store 64 63
65: 6(float) Load 8(resultf)
66: 6(float) AtomicExchange 11(atomf) 14 15 65
Store 8(resultf) 66
67: 6(float) Load 8(resultf)
68: 28(ptr) AccessChain 27(buf) 21
69: 6(float) Load 68
70: 6(float) FAdd 69 67
71: 28(ptr) AccessChain 27(buf) 21
Store 71 70
72: 6(float) Load 8(resultf)
73: 6(float) AtomicExchange 11(atomf) 19 58 72
Store 8(resultf) 73
74: 6(float) Load 8(resultf)
75: 28(ptr) AccessChain 27(buf) 21
76: 6(float) Load 75
77: 6(float) FAdd 76 74
78: 28(ptr) AccessChain 27(buf) 21
Store 78 77
79: 42(ptr) AccessChain 27(buf) 19
80:24(float64_t) Load 34(resultd)
81:24(float64_t) AtomicExchange 79 14 15 80
Store 34(resultd) 81
82:24(float64_t) Load 34(resultd)
83: 42(ptr) AccessChain 27(buf) 19
84:24(float64_t) Load 83
85:24(float64_t) FAdd 84 82
86: 42(ptr) AccessChain 27(buf) 19
Store 86 85
87: 42(ptr) AccessChain 27(buf) 19
88:24(float64_t) Load 34(resultd)
89:24(float64_t) AtomicExchange 87 19 58 88
Store 34(resultd) 89
90:24(float64_t) Load 34(resultd)
91: 42(ptr) AccessChain 27(buf) 19
92:24(float64_t) Load 91
93:24(float64_t) FAdd 92 90
94: 42(ptr) AccessChain 27(buf) 19
Store 94 93
95:24(float64_t) Load 34(resultd)
96:24(float64_t) AtomicExchange 37(atomd) 14 15 95
Store 34(resultd) 96
97:24(float64_t) Load 34(resultd)
98: 42(ptr) AccessChain 27(buf) 19
99:24(float64_t) Load 98
100:24(float64_t) FAdd 99 97
101: 42(ptr) AccessChain 27(buf) 19
Store 101 100
102:24(float64_t) Load 34(resultd)
103:24(float64_t) AtomicExchange 37(atomd) 19 58 102
Store 34(resultd) 103
104:24(float64_t) Load 34(resultd)
105: 42(ptr) AccessChain 27(buf) 19
106:24(float64_t) Load 105
107:24(float64_t) FAdd 106 104
108: 42(ptr) AccessChain 27(buf) 19
Store 108 107
109: 28(ptr) AccessChain 27(buf) 21
110: 6(float) AtomicLoad 109 19 58
Store 8(resultf) 110
111: 28(ptr) AccessChain 27(buf) 21
112: 6(float) Load 8(resultf)
AtomicStore 111 19 58 112
113: 6(float) Load 8(resultf)
114: 28(ptr) AccessChain 27(buf) 21
115: 6(float) Load 114
116: 6(float) FAdd 115 113
117: 28(ptr) AccessChain 27(buf) 21
Store 117 116
118: 6(float) AtomicLoad 11(atomf) 19 58
Store 8(resultf) 118
119: 6(float) Load 8(resultf)
AtomicStore 11(atomf) 19 58 119
120: 6(float) Load 8(resultf)
121: 28(ptr) AccessChain 27(buf) 21
122: 6(float) Load 121
123: 6(float) FAdd 122 120
124: 28(ptr) AccessChain 27(buf) 21
Store 124 123
125: 42(ptr) AccessChain 27(buf) 19
126:24(float64_t) AtomicLoad 125 19 58
Store 34(resultd) 126
127: 42(ptr) AccessChain 27(buf) 19
128:24(float64_t) Load 34(resultd)
AtomicStore 127 19 58 128
129:24(float64_t) Load 34(resultd)
130: 42(ptr) AccessChain 27(buf) 19
131:24(float64_t) Load 130
132:24(float64_t) FAdd 131 129
133: 42(ptr) AccessChain 27(buf) 19
Store 133 132
134:24(float64_t) AtomicLoad 37(atomd) 19 58
Store 34(resultd) 134
135:24(float64_t) Load 34(resultd)
AtomicStore 37(atomd) 19 58 135
136:24(float64_t) Load 34(resultd)
137: 42(ptr) AccessChain 27(buf) 19
138:24(float64_t) Load 137
139:24(float64_t) FAdd 138 136
140: 42(ptr) AccessChain 27(buf) 19
Store 140 139
146: 145(ptr) ImageTexelPointer 143(fimage1D) 21 15
148: 6(float) AtomicFAddEXT 146 14 147 144
Store 11(atomf) 148 MakePointerAvailableKHR NonPrivatePointerKHR 149
150: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
151: 28(ptr) AccessChain 27(buf) 21
152: 6(float) Load 151
153: 6(float) FAdd 152 150
154: 28(ptr) AccessChain 27(buf) 21
Store 154 153
156: 145(ptr) ImageTexelPointer 143(fimage1D) 19 15
158: 6(float) AtomicFAddEXT 156 19 157 12
Store 11(atomf) 158 MakePointerAvailableKHR NonPrivatePointerKHR 149
159: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
160: 28(ptr) AccessChain 27(buf) 21
161: 6(float) Load 160
162: 6(float) FAdd 161 159
163: 28(ptr) AccessChain 27(buf) 21
Store 163 162
165: 145(ptr) ImageTexelPointer 143(fimage1D) 19 15
166: 6(float) AtomicExchange 165 19 157 164
Store 11(atomf) 166 MakePointerAvailableKHR NonPrivatePointerKHR 149
167: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
168: 28(ptr) AccessChain 27(buf) 21
169: 6(float) Load 168
170: 6(float) FAdd 169 167
171: 28(ptr) AccessChain 27(buf) 21
Store 171 170
172: 145(ptr) ImageTexelPointer 143(fimage1D) 19 15
173: 6(float) AtomicLoad 172 19 157
Store 11(atomf) 173 MakePointerAvailableKHR NonPrivatePointerKHR 149
174: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
175: 28(ptr) AccessChain 27(buf) 21
176: 6(float) Load 175
177: 6(float) FAdd 176 174
178: 28(ptr) AccessChain 27(buf) 21
Store 178 177
180: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
181: 145(ptr) ImageTexelPointer 143(fimage1D) 179 15
AtomicStore 181 19 157 180
182: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
183: 28(ptr) AccessChain 27(buf) 21
184: 6(float) Load 183
185: 6(float) FAdd 184 182
186: 28(ptr) AccessChain 27(buf) 21
Store 186 185
192: 145(ptr) ImageTexelPointer 189(fimage1DArray) 191 15
193: 6(float) AtomicFAddEXT 192 14 147 144
Store 11(atomf) 193 MakePointerAvailableKHR NonPrivatePointerKHR 149
194: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
195: 28(ptr) AccessChain 27(buf) 21
196: 6(float) Load 195
197: 6(float) FAdd 196 194
198: 28(ptr) AccessChain 27(buf) 21
Store 198 197
200: 145(ptr) ImageTexelPointer 189(fimage1DArray) 199 15
201: 6(float) AtomicFAddEXT 200 19 157 12
Store 11(atomf) 201 MakePointerAvailableKHR NonPrivatePointerKHR 149
202: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
203: 28(ptr) AccessChain 27(buf) 21
204: 6(float) Load 203
205: 6(float) FAdd 204 202
206: 28(ptr) AccessChain 27(buf) 21
Store 206 205
208: 145(ptr) ImageTexelPointer 189(fimage1DArray) 207 15
209: 6(float) AtomicExchange 208 19 157 164
Store 11(atomf) 209 MakePointerAvailableKHR NonPrivatePointerKHR 149
210: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
211: 28(ptr) AccessChain 27(buf) 21
212: 6(float) Load 211
213: 6(float) FAdd 212 210
214: 28(ptr) AccessChain 27(buf) 21
Store 214 213
215: 145(ptr) ImageTexelPointer 189(fimage1DArray) 199 15
216: 6(float) AtomicLoad 215 19 157
Store 11(atomf) 216 MakePointerAvailableKHR NonPrivatePointerKHR 149
217: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
218: 28(ptr) AccessChain 27(buf) 21
219: 6(float) Load 218
220: 6(float) FAdd 219 217
221: 28(ptr) AccessChain 27(buf) 21
Store 221 220
223: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
224: 145(ptr) ImageTexelPointer 189(fimage1DArray) 222 15
AtomicStore 224 19 157 223
225: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
226: 28(ptr) AccessChain 27(buf) 21
227: 6(float) Load 226
228: 6(float) FAdd 227 225
229: 28(ptr) AccessChain 27(buf) 21
Store 229 228
233: 145(ptr) ImageTexelPointer 232(fimage2D) 191 15
234: 6(float) AtomicFAddEXT 233 14 147 144
Store 11(atomf) 234 MakePointerAvailableKHR NonPrivatePointerKHR 149
235: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
236: 28(ptr) AccessChain 27(buf) 21
237: 6(float) Load 236
238: 6(float) FAdd 237 235
239: 28(ptr) AccessChain 27(buf) 21
Store 239 238
240: 145(ptr) ImageTexelPointer 232(fimage2D) 199 15
241: 6(float) AtomicFAddEXT 240 19 157 12
Store 11(atomf) 241 MakePointerAvailableKHR NonPrivatePointerKHR 149
242: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
243: 28(ptr) AccessChain 27(buf) 21
244: 6(float) Load 243
245: 6(float) FAdd 244 242
246: 28(ptr) AccessChain 27(buf) 21
Store 246 245
247: 145(ptr) ImageTexelPointer 232(fimage2D) 207 15
248: 6(float) AtomicExchange 247 19 157 164
Store 11(atomf) 248 MakePointerAvailableKHR NonPrivatePointerKHR 149
249: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
250: 28(ptr) AccessChain 27(buf) 21
251: 6(float) Load 250
252: 6(float) FAdd 251 249
253: 28(ptr) AccessChain 27(buf) 21
Store 253 252
254: 145(ptr) ImageTexelPointer 232(fimage2D) 199 15
255: 6(float) AtomicLoad 254 19 157
Store 11(atomf) 255 MakePointerAvailableKHR NonPrivatePointerKHR 149
256: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
257: 28(ptr) AccessChain 27(buf) 21
258: 6(float) Load 257
259: 6(float) FAdd 258 256
260: 28(ptr) AccessChain 27(buf) 21
Store 260 259
261: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
262: 145(ptr) ImageTexelPointer 232(fimage2D) 222 15
AtomicStore 262 19 157 261
263: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
264: 28(ptr) AccessChain 27(buf) 21
265: 6(float) Load 264
266: 6(float) FAdd 265 263
267: 28(ptr) AccessChain 27(buf) 21
Store 267 266
271: 145(ptr) ImageTexelPointer 270(fimage2DRect) 191 15
272: 6(float) AtomicFAddEXT 271 14 147 144
Store 11(atomf) 272 MakePointerAvailableKHR NonPrivatePointerKHR 149
273: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
274: 28(ptr) AccessChain 27(buf) 21
275: 6(float) Load 274
276: 6(float) FAdd 275 273
277: 28(ptr) AccessChain 27(buf) 21
Store 277 276
278: 145(ptr) ImageTexelPointer 270(fimage2DRect) 199 15
279: 6(float) AtomicFAddEXT 278 19 157 12
Store 11(atomf) 279 MakePointerAvailableKHR NonPrivatePointerKHR 149
280: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
281: 28(ptr) AccessChain 27(buf) 21
282: 6(float) Load 281
283: 6(float) FAdd 282 280
284: 28(ptr) AccessChain 27(buf) 21
Store 284 283
285: 145(ptr) ImageTexelPointer 270(fimage2DRect) 207 15
286: 6(float) AtomicExchange 285 19 157 164
Store 11(atomf) 286 MakePointerAvailableKHR NonPrivatePointerKHR 149
287: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
288: 28(ptr) AccessChain 27(buf) 21
289: 6(float) Load 288
290: 6(float) FAdd 289 287
291: 28(ptr) AccessChain 27(buf) 21
Store 291 290
292: 145(ptr) ImageTexelPointer 270(fimage2DRect) 199 15
293: 6(float) AtomicLoad 292 19 157
Store 11(atomf) 293 MakePointerAvailableKHR NonPrivatePointerKHR 149
294: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
295: 28(ptr) AccessChain 27(buf) 21
296: 6(float) Load 295
297: 6(float) FAdd 296 294
298: 28(ptr) AccessChain 27(buf) 21
Store 298 297
299: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
300: 145(ptr) ImageTexelPointer 270(fimage2DRect) 222 15
AtomicStore 300 19 157 299
301: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
302: 28(ptr) AccessChain 27(buf) 21
303: 6(float) Load 302
304: 6(float) FAdd 303 301
305: 28(ptr) AccessChain 27(buf) 21
Store 305 304
311: 145(ptr) ImageTexelPointer 308(fimage2DArray) 310 15
312: 6(float) AtomicFAddEXT 311 14 147 144
Store 11(atomf) 312 MakePointerAvailableKHR NonPrivatePointerKHR 149
313: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
314: 28(ptr) AccessChain 27(buf) 21
315: 6(float) Load 314
316: 6(float) FAdd 315 313
317: 28(ptr) AccessChain 27(buf) 21
Store 317 316
319: 145(ptr) ImageTexelPointer 308(fimage2DArray) 318 15
320: 6(float) AtomicFAddEXT 319 19 157 12
Store 11(atomf) 320 MakePointerAvailableKHR NonPrivatePointerKHR 149
321: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
322: 28(ptr) AccessChain 27(buf) 21
323: 6(float) Load 322
324: 6(float) FAdd 323 321
325: 28(ptr) AccessChain 27(buf) 21
Store 325 324
327: 145(ptr) ImageTexelPointer 308(fimage2DArray) 326 15
328: 6(float) AtomicExchange 327 19 157 164
Store 11(atomf) 328 MakePointerAvailableKHR NonPrivatePointerKHR 149
329: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
330: 28(ptr) AccessChain 27(buf) 21
331: 6(float) Load 330
332: 6(float) FAdd 331 329
333: 28(ptr) AccessChain 27(buf) 21
Store 333 332
335: 145(ptr) ImageTexelPointer 308(fimage2DArray) 334 15
336: 6(float) AtomicLoad 335 19 157
Store 11(atomf) 336 MakePointerAvailableKHR NonPrivatePointerKHR 149
337: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
338: 28(ptr) AccessChain 27(buf) 21
339: 6(float) Load 338
340: 6(float) FAdd 339 337
341: 28(ptr) AccessChain 27(buf) 21
Store 341 340
343: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
344: 145(ptr) ImageTexelPointer 308(fimage2DArray) 342 15
AtomicStore 344 19 157 343
345: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
346: 28(ptr) AccessChain 27(buf) 21
347: 6(float) Load 346
348: 6(float) FAdd 347 345
349: 28(ptr) AccessChain 27(buf) 21
Store 349 348
353: 145(ptr) ImageTexelPointer 352(fimageCube) 310 15
354: 6(float) AtomicFAddEXT 353 14 147 144
Store 11(atomf) 354 MakePointerAvailableKHR NonPrivatePointerKHR 149
355: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
356: 28(ptr) AccessChain 27(buf) 21
357: 6(float) Load 356
358: 6(float) FAdd 357 355
359: 28(ptr) AccessChain 27(buf) 21
Store 359 358
360: 145(ptr) ImageTexelPointer 352(fimageCube) 318 15
361: 6(float) AtomicFAddEXT 360 19 157 12
Store 11(atomf) 361 MakePointerAvailableKHR NonPrivatePointerKHR 149
362: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
363: 28(ptr) AccessChain 27(buf) 21
364: 6(float) Load 363
365: 6(float) FAdd 364 362
366: 28(ptr) AccessChain 27(buf) 21
Store 366 365
368: 145(ptr) ImageTexelPointer 352(fimageCube) 367 15
369: 6(float) AtomicExchange 368 19 157 164
Store 11(atomf) 369 MakePointerAvailableKHR NonPrivatePointerKHR 149
370: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
371: 28(ptr) AccessChain 27(buf) 21
372: 6(float) Load 371
373: 6(float) FAdd 372 370
374: 28(ptr) AccessChain 27(buf) 21
Store 374 373
375: 145(ptr) ImageTexelPointer 352(fimageCube) 334 15
376: 6(float) AtomicLoad 375 19 157
Store 11(atomf) 376 MakePointerAvailableKHR NonPrivatePointerKHR 149
377: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
378: 28(ptr) AccessChain 27(buf) 21
379: 6(float) Load 378
380: 6(float) FAdd 379 377
381: 28(ptr) AccessChain 27(buf) 21
Store 381 380
383: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
384: 145(ptr) ImageTexelPointer 352(fimageCube) 382 15
AtomicStore 384 19 157 383
385: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
386: 28(ptr) AccessChain 27(buf) 21
387: 6(float) Load 386
388: 6(float) FAdd 387 385
389: 28(ptr) AccessChain 27(buf) 21
Store 389 388
393: 145(ptr) ImageTexelPointer 392(fimageCubeArray) 310 15
394: 6(float) AtomicFAddEXT 393 14 147 144
Store 11(atomf) 394 MakePointerAvailableKHR NonPrivatePointerKHR 149
395: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
396: 28(ptr) AccessChain 27(buf) 21
397: 6(float) Load 396
398: 6(float) FAdd 397 395
399: 28(ptr) AccessChain 27(buf) 21
Store 399 398
400: 145(ptr) ImageTexelPointer 392(fimageCubeArray) 318 15
401: 6(float) AtomicFAddEXT 400 19 157 12
Store 11(atomf) 401 MakePointerAvailableKHR NonPrivatePointerKHR 149
402: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
403: 28(ptr) AccessChain 27(buf) 21
404: 6(float) Load 403
405: 6(float) FAdd 404 402
406: 28(ptr) AccessChain 27(buf) 21
Store 406 405
407: 145(ptr) ImageTexelPointer 392(fimageCubeArray) 326 15
408: 6(float) AtomicExchange 407 19 157 164
Store 11(atomf) 408 MakePointerAvailableKHR NonPrivatePointerKHR 149
409: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
410: 28(ptr) AccessChain 27(buf) 21
411: 6(float) Load 410
412: 6(float) FAdd 411 409
413: 28(ptr) AccessChain 27(buf) 21
Store 413 412
414: 145(ptr) ImageTexelPointer 392(fimageCubeArray) 334 15
415: 6(float) AtomicLoad 414 19 157
Store 11(atomf) 415 MakePointerAvailableKHR NonPrivatePointerKHR 149
416: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
417: 28(ptr) AccessChain 27(buf) 21
418: 6(float) Load 417
419: 6(float) FAdd 418 416
420: 28(ptr) AccessChain 27(buf) 21
Store 420 419
421: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
422: 145(ptr) ImageTexelPointer 392(fimageCubeArray) 342 15
AtomicStore 422 19 157 421
423: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
424: 28(ptr) AccessChain 27(buf) 21
425: 6(float) Load 424
426: 6(float) FAdd 425 423
427: 28(ptr) AccessChain 27(buf) 21
Store 427 426
431: 145(ptr) ImageTexelPointer 430(fimage3D) 310 15
432: 6(float) AtomicFAddEXT 431 14 147 144
Store 11(atomf) 432 MakePointerAvailableKHR NonPrivatePointerKHR 149
433: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
434: 28(ptr) AccessChain 27(buf) 21
435: 6(float) Load 434
436: 6(float) FAdd 435 433
437: 28(ptr) AccessChain 27(buf) 21
Store 437 436
438: 145(ptr) ImageTexelPointer 430(fimage3D) 318 15
439: 6(float) AtomicFAddEXT 438 19 157 12
Store 11(atomf) 439 MakePointerAvailableKHR NonPrivatePointerKHR 149
440: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
441: 28(ptr) AccessChain 27(buf) 21
442: 6(float) Load 441
443: 6(float) FAdd 442 440
444: 28(ptr) AccessChain 27(buf) 21
Store 444 443
445: 145(ptr) ImageTexelPointer 430(fimage3D) 326 15
446: 6(float) AtomicExchange 445 19 157 164
Store 11(atomf) 446 MakePointerAvailableKHR NonPrivatePointerKHR 149
447: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
448: 28(ptr) AccessChain 27(buf) 21
449: 6(float) Load 448
450: 6(float) FAdd 449 447
451: 28(ptr) AccessChain 27(buf) 21
Store 451 450
452: 145(ptr) ImageTexelPointer 430(fimage3D) 334 15
453: 6(float) AtomicLoad 452 19 157
Store 11(atomf) 453 MakePointerAvailableKHR NonPrivatePointerKHR 149
454: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
455: 28(ptr) AccessChain 27(buf) 21
456: 6(float) Load 455
457: 6(float) FAdd 456 454
458: 28(ptr) AccessChain 27(buf) 21
Store 458 457
459: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
460: 145(ptr) ImageTexelPointer 430(fimage3D) 342 15
AtomicStore 460 19 157 459
461: 6(float) Load 11(atomf) MakePointerVisibleKHR NonPrivatePointerKHR 149
462: 28(ptr) AccessChain 27(buf) 21
463: 6(float) Load 462
464: 6(float) FAdd 463 461
465: 28(ptr) AccessChain 27(buf) 21
Store 465 464
Return
FunctionEnd
Test/baseResults/spv.atomicFloat_Error.comp.out 0 → 100755
spv.atomicFloat_Error.comp
ERROR: 0:25: 'atomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:26: 'atomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:27: 'atomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:28: 'atomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:31: 'atomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:32: 'atomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:33: 'atomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:34: 'atomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:37: 'atomicExchange' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:38: 'atomicExchange' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:39: 'atomicExchange' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:40: 'atomicExchange' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:43: 'atomicExchange' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:44: 'atomicExchange' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:45: 'atomicExchange' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:46: 'atomicExchange' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:50: 'atomicLoad' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:51: 'atomicStore' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:52: 'atomicLoad' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:53: 'atomicStore' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:56: 'atomicLoad' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:57: 'atomicStore' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:58: 'atomicLoad' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:59: 'atomicStore' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:63: 'imageAtomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:64: 'imageAtomicAdd' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:66: 'imageAtomicLoad' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:67: 'imageAtomicStore' : required extension not requested: GL_EXT_shader_atomic_float
ERROR: 0:80: 'atomicAdd' : no matching overloaded function found
ERROR: 0:81: 'atomicAdd' : no matching overloaded function found
ERROR: 0:82: 'atomicAdd' : no matching overloaded function found
ERROR: 0:83: 'atomicAdd' : no matching overloaded function found
ERROR: 0:86: 'atomicExchange' : no matching overloaded function found
ERROR: 0:87: 'atomicExchange' : no matching overloaded function found
ERROR: 0:88: 'atomicExchange' : no matching overloaded function found
ERROR: 0:89: 'atomicExchange' : no matching overloaded function found
ERROR: 0:90: 'atomicExchange' : no matching overloaded function found
ERROR: 0:91: 'atomicExchange' : no matching overloaded function found
ERROR: 0:95: 'atomicLoad' : no matching overloaded function found
ERROR: 0:96: 'atomicStore' : no matching overloaded function found
ERROR: 0:99: 'imageAtomicAdd' : no matching overloaded function found
ERROR: 0:100: 'imageAtomicAdd' : no matching overloaded function found
ERROR: 0:101: 'imageAtomicAdd' : no matching overloaded function found
ERROR: 0:102: 'imageAtomicExchange' : no matching overloaded function found
ERROR: 0:103: 'imageAtomicExchange' : no matching overloaded function found
ERROR: 0:104: 'imageAtomicExchange' : no matching overloaded function found
ERROR: 0:105: 'imageAtomicLoad' : no matching overloaded function found
ERROR: 0:106: 'imageAtomicLoad' : no matching overloaded function found
ERROR: 0:107: 'imageAtomicLoad' : no matching overloaded function found
ERROR: 0:108: 'imageAtomicStore' : no matching overloaded function found
ERROR: 0:109: 'imageAtomicStore' : no matching overloaded function found
ERROR: 0:110: 'imageAtomicStore' : no matching overloaded function found
ERROR: 52 compilation errors. No code generated.
SPIR-V is not generated for failed compile or link
Test/spv.atomicFloat.comp 0 → 100755
#version 450 core
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_EXT_shader_atomic_float: enable
#pragma use_vulkan_memory_model
layout(local_size_x = 16, local_size_y = 16) in;
layout(binding = 0) buffer Buffer
{
float dataf;
double datad;
} buf;
shared float atomf;
shared double atomd;
layout(binding = 0, r32f) volatile coherent uniform image1D fimage1D;
layout(binding = 1, r32f) volatile coherent uniform image1DArray fimage1DArray;
layout(binding = 2, r32f) volatile coherent uniform image2D fimage2D;
layout(binding = 3, r32f) volatile coherent uniform image2DArray fimage2DArray;
layout(binding = 4, r32f) volatile coherent uniform image2DRect fimage2DRect;
layout(binding = 5, r32f) volatile coherent uniform imageCube fimageCube;
layout(binding = 6, r32f) volatile coherent uniform imageCubeArray fimageCubeArray;
layout(binding = 9, r32f) volatile coherent uniform image3D fimage3D;
void main()
{
//atomicAdd
float resultf = 0;
resultf = atomicAdd(atomf, 3.0);
resultf = atomicAdd(atomf, 4.5, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
resultf = atomicAdd(buf.dataf, 3.0);
resultf = atomicAdd(buf.dataf, 4.5, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
double resultd = 0;
resultd = atomicAdd(atomd, 3.0);
resultd = atomicAdd(atomd, 4.5, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
resultd = atomicAdd(buf.datad, 3.0);
resultd = atomicAdd(buf.datad, 4.5, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
//atomicExchange
resultf = atomicExchange(buf.dataf, resultf);
buf.dataf += resultf;
resultf = atomicExchange(buf.dataf, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.dataf += resultf;
resultf = atomicExchange(atomf, resultf);
buf.dataf += resultf;
resultf = atomicExchange(atomf, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.dataf += resultf;
resultd = atomicExchange(buf.datad, resultd);
buf.datad += resultd;
resultd = atomicExchange(buf.datad, resultd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.datad += resultd;
resultd = atomicExchange(atomd, resultd);
buf.datad += resultd;
resultd = atomicExchange(atomd, resultd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.datad += resultd;
//atomic load/store
resultf = atomicLoad(buf.dataf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(buf.dataf, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.dataf += resultf;
resultf = atomicLoad(atomf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(atomf, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.dataf += resultf;
resultd = atomicLoad(buf.datad, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(buf.datad, resultd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.datad += resultd;
resultd = atomicLoad(atomd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(atomd, resultd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.datad += resultd;
// image atomics on 1D:
atomf = imageAtomicAdd(fimage1D, int(0), 2.0);
buf.dataf += atomf;
atomf = imageAtomicAdd(fimage1D, int(1), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicExchange(fimage1D, int(1), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicLoad(fimage1D, int(1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
imageAtomicStore(fimage1D, int(2), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
// image atomics on 1D Array:
atomf = imageAtomicAdd(fimage1DArray, ivec2(0,0), 2.0);
buf.dataf += atomf;
atomf = imageAtomicAdd(fimage1DArray, ivec2(1,1), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicExchange(fimage1DArray, ivec2(1,0), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicLoad(fimage1DArray, ivec2(1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
imageAtomicStore(fimage1DArray, ivec2(2,2), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
// image atomics on 2D:
atomf = imageAtomicAdd(fimage2D, ivec2(0,0), 2.0);
buf.dataf += atomf;
atomf = imageAtomicAdd(fimage2D, ivec2(1,1), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicExchange(fimage2D, ivec2(1,0), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicLoad(fimage2D, ivec2(1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
imageAtomicStore(fimage2D, ivec2(2,2), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
// image atomics on 2D Rect:
atomf = imageAtomicAdd(fimage2DRect, ivec2(0,0), 2.0);
buf.dataf += atomf;
atomf = imageAtomicAdd(fimage2DRect, ivec2(1,1), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicExchange(fimage2DRect, ivec2(1,0), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicLoad(fimage2DRect, ivec2(1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
imageAtomicStore(fimage2DRect, ivec2(2,2), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
// image atomics on 2D Array:
atomf = imageAtomicAdd(fimage2DArray, ivec3(0,0,0), 2.0);
buf.dataf += atomf;
atomf = imageAtomicAdd(fimage2DArray, ivec3(1,1,0), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicExchange(fimage2DArray, ivec3(1,0,1), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicLoad(fimage2DArray, ivec3(1,1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
imageAtomicStore(fimage2DArray, ivec3(2,2,0), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
// image atomics on Cube:
atomf = imageAtomicAdd(fimageCube, ivec3(0,0,0), 2.0);
buf.dataf += atomf;
atomf = imageAtomicAdd(fimageCube, ivec3(1,1,0), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicExchange(fimageCube, ivec3(1,0,0), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicLoad(fimageCube, ivec3(1,1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
imageAtomicStore(fimageCube, ivec3(2,2,1), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
// image atomics on Cube Array:
atomf = imageAtomicAdd(fimageCubeArray, ivec3(0,0,0), 2.0);
buf.dataf += atomf;
atomf = imageAtomicAdd(fimageCubeArray, ivec3(1,1,0), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicExchange(fimageCubeArray, ivec3(1,0,1), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicLoad(fimageCubeArray, ivec3(1,1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
imageAtomicStore(fimageCubeArray, ivec3(2,2,0), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
// image atomics on 3D:
atomf = imageAtomicAdd(fimage3D, ivec3(0,0,0), 2.0);
buf.dataf += atomf;
atomf = imageAtomicAdd(fimage3D, ivec3(1,1,0), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicExchange(fimage3D, ivec3(1,0,1), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
atomf = imageAtomicLoad(fimage3D, ivec3(1,1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
imageAtomicStore(fimage3D, ivec3(2,2,0), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
}
\ No newline at end of file
Test/spv.atomicFloat_Error.comp 0 → 100755
#version 450 core
#extension GL_KHR_memory_scope_semantics : enable
#pragma use_vulkan_memory_model
layout(local_size_x = 16, local_size_y = 16) in;
layout(binding = 0) buffer Buffer
{
int datai;
float dataf;
double datad;
} buf;
shared int atomi;
shared float atomf;
shared double atomd;
layout(binding = 0, r32f) volatile coherent uniform image1D fimage1D;
layout(binding = 1, r32f) volatile coherent uniform image2D fimage2D;
void undeclared_errors()
{
//atomicAdd
float resultf = 0;
resultf = atomicAdd(atomf, 3.0);
resultf = atomicAdd(atomf, 4.5, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
resultf = atomicAdd(buf.dataf, 3.0);
resultf = atomicAdd(buf.dataf, 4.5, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
double resultd = 0;
resultd = atomicAdd(atomd, 3.0);
resultd = atomicAdd(atomd, 4.5, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
resultd = atomicAdd(buf.datad, 3.0);
resultd = atomicAdd(buf.datad, 4.5, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
//atomicExchange
resultf = atomicExchange(buf.dataf, resultf);
resultf = atomicExchange(buf.dataf, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
resultf = atomicExchange(atomf, resultf);
resultf = atomicExchange(atomf, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.dataf += resultf;
resultd = atomicExchange(buf.datad, resultd);
resultd = atomicExchange(buf.datad, resultd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
resultd = atomicExchange(atomd, resultd);
resultd = atomicExchange(atomd, resultd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.datad += resultd;
//atomic load/store
resultf = atomicLoad(buf.dataf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(buf.dataf, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
resultf = atomicLoad(atomf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(atomf, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.dataf += resultf;
resultd = atomicLoad(buf.datad, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(buf.datad, resultd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
resultd = atomicLoad(atomd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(atomd, resultd, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.datad += resultd;
// image atomics:
atomf = imageAtomicAdd(fimage2D, ivec2(0,0), 2.0);
atomf = imageAtomicAdd(fimage2D, ivec2(1,1), 3.0, gl_ScopeDevice, gl_StorageSemanticsImage , gl_SemanticsRelaxed);
atomf = imageAtomicExchange(fimage2D, ivec2(1,0), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
atomf = imageAtomicLoad(fimage2D, ivec2(1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
imageAtomicStore(fimage2D, ivec2(2,2), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
}
#extension GL_EXT_shader_atomic_float: enable
void main()
{
float resultf = 0;
double resultd = 0;
int resulti = 0;
//atomicAdd
resultf = atomicAdd(atomi);
resultf = atomicAdd(atomi, 3.0);
resultf = atomicAdd(atomi, resultf, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
resultf = atomicAdd(atomi, resultf, gl_ScopeDevice, gl_StorageSemanticsBuffer, gl_SemanticsRelaxed);
//atomicExchange
resultf = atomicExchange(buf.datai);
resultf = atomicExchange(buf.datai, resultf);
resultf = atomicExchange(buf.datai, resultf, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
resultf = atomicExchange(buf.datai, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
resultf = atomicExchange(atomi, resultf);
resultf = atomicExchange(atomi, resultf, gl_ScopeDevice, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
buf.dataf += resultf;
//atomic load/store
resultf = atomicLoad(buf.datai, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
atomicStore(buf.datai, resulti, gl_StorageSemanticsShared, gl_SemanticsRelaxed);
// image atomics:
atomf = imageAtomicAdd(fimage1D, ivec2(0,0), 2.0);
atomf = imageAtomicAdd(fimage2D, ivec3(0,0,0), 2.0);
atomf = imageAtomicAdd(fimage2D, 2.0);
atomf = imageAtomicExchange(fimage1D, ivec2(1,0), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
atomf = imageAtomicExchange(fimage2D, 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
atomf = imageAtomicExchange(fimage2D, ivec3(1,0,1), 4.0, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
atomf = imageAtomicLoad(fimage1D, ivec2(1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
atomf = imageAtomicLoad(fimage2D, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
atomf = imageAtomicLoad(fimage2D, ivec3(1,1,1), gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
imageAtomicStore(fimage1D, ivec2(2,2), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
imageAtomicStore(fimage2D, atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
imageAtomicStore(fimage2D, ivec3(2,2,1), atomf, gl_ScopeDevice, gl_StorageSemanticsImage, gl_SemanticsRelaxed);
buf.dataf += atomf;
}
\ No newline at end of file
glslang/MachineIndependent/Initialize.cpp
......@@ -1241,11 +1241,19 @@ void TBuiltIns::initialize(int version, EProfile profile, const SpvVersion& spvV
" int64_t atomicAdd(coherent volatile inout int64_t, int64_t);"
"uint64_t atomicAdd(coherent volatile inout uint64_t, uint64_t, int, int, int);"
" int64_t atomicAdd(coherent volatile inout int64_t, int64_t, int, int, int);"
" float atomicAdd(coherent volatile inout float, float);"
" float atomicAdd(coherent volatile inout float, float, int, int, int);"
" double atomicAdd(coherent volatile inout double, double);"
" double atomicAdd(coherent volatile inout double, double, int, int, int);"
"uint64_t atomicExchange(coherent volatile inout uint64_t, uint64_t);"
" int64_t atomicExchange(coherent volatile inout int64_t, int64_t);"
"uint64_t atomicExchange(coherent volatile inout uint64_t, uint64_t, int, int, int);"
" int64_t atomicExchange(coherent volatile inout int64_t, int64_t, int, int, int);"
" float atomicExchange(coherent volatile inout float, float);"
" float atomicExchange(coherent volatile inout float, float, int, int, int);"
" double atomicExchange(coherent volatile inout double, double);"
" double atomicExchange(coherent volatile inout double, double, int, int, int);"
"uint64_t atomicCompSwap(coherent volatile inout uint64_t, uint64_t, uint64_t);"
" int64_t atomicCompSwap(coherent volatile inout int64_t, int64_t, int64_t);"
......@@ -1254,9 +1262,13 @@ void TBuiltIns::initialize(int version, EProfile profile, const SpvVersion& spvV
"uint64_t atomicLoad(coherent volatile in uint64_t, int, int, int);"
" int64_t atomicLoad(coherent volatile in int64_t, int, int, int);"
" float atomicLoad(coherent volatile in float, int, int, int);"
" double atomicLoad(coherent volatile in double, int, int, int);"
"void atomicStore(coherent volatile out uint64_t, uint64_t, int, int, int);"
"void atomicStore(coherent volatile out int64_t, int64_t, int, int, int);"
"void atomicStore(coherent volatile out float, float, int, int, int);"
"void atomicStore(coherent volatile out double, double, int, int, int);"
"\n");
}
#endif
......@@ -6013,12 +6025,39 @@ void TBuiltIns::addImageFunctions(TSampler sampler, const TString& typeName, int
// not int or uint
// GL_ARB_ES3_1_compatibility
// TODO: spec issue: are there restrictions on the kind of layout() that can be used? what about dropping memory qualifiers?
if ((profile != EEsProfile && version >= 450) ||
(profile == EEsProfile && version >= 310)) {
if (profile == EEsProfile && version >= 310) {
commonBuiltins.append("float imageAtomicExchange(volatile coherent ");
commonBuiltins.append(imageParams);
commonBuiltins.append(", float);\n");
}
if (profile != EEsProfile && version >= 450) {
commonBuiltins.append("float imageAtomicAdd(volatile coherent ");
commonBuiltins.append(imageParams);
commonBuiltins.append(", float);\n");
commonBuiltins.append("float imageAtomicAdd(volatile coherent ");
commonBuiltins.append(imageParams);
commonBuiltins.append(", float");
commonBuiltins.append(", int, int, int);\n");
commonBuiltins.append("float imageAtomicExchange(volatile coherent ");
commonBuiltins.append(imageParams);
commonBuiltins.append(", float);\n");
commonBuiltins.append("float imageAtomicExchange(volatile coherent ");
commonBuiltins.append(imageParams);
commonBuiltins.append(", float");
commonBuiltins.append(", int, int, int);\n");
commonBuiltins.append("float imageAtomicLoad(readonly volatile coherent ");
commonBuiltins.append(imageParams);
commonBuiltins.append(", int, int, int);\n");
commonBuiltins.append("void imageAtomicStore(writeonly volatile coherent ");
commonBuiltins.append(imageParams);
commonBuiltins.append(", float");
commonBuiltins.append(", int, int, int);\n");
}
}
}
......
glslang/MachineIndependent/ParseHelper.cpp
......@@ -2110,7 +2110,14 @@ void TParseContext::builtInOpCheck(const TSourceLoc& loc, const TFunction& fnCan
if (imageType.getQualifier().getFormat() != ElfR32i && imageType.getQualifier().getFormat() != ElfR32ui)
error(loc, "only supported on image with format r32i or r32ui", fnCandidate.getName().c_str(), "");
} else {
if (fnCandidate.getName().compare(0, 19, "imageAtomicExchange") != 0)
bool isImageAtomicOnFloatAllowed = ((fnCandidate.getName().compare(0, 14, "imageAtomicAdd") == 0) ||
(fnCandidate.getName().compare(0, 15, "imageAtomicLoad") == 0) ||
(fnCandidate.getName().compare(0, 16, "imageAtomicStore") == 0) ||
(fnCandidate.getName().compare(0, 19, "imageAtomicExchange") == 0));
if (imageType.getSampler().type == EbtFloat && isImageAtomicOnFloatAllowed &&
(fnCandidate.getName().compare(0, 19, "imageAtomicExchange") != 0)) // imageAtomicExchange doesn't require GL_EXT_shader_atomic_float
requireExtensions(loc, 1, &E_GL_EXT_shader_atomic_float, fnCandidate.getName().c_str());
if (!isImageAtomicOnFloatAllowed)
error(loc, "only supported on integer images", fnCandidate.getName().c_str(), "");
else if (imageType.getQualifier().getFormat() != ElfR32f && isEsProfile())
error(loc, "only supported on image with format r32f", fnCandidate.getName().c_str(), "");
......@@ -2139,10 +2146,18 @@ void TParseContext::builtInOpCheck(const TSourceLoc& loc, const TFunction& fnCan
if (argp->size() > 3) {
requireExtensions(loc, 1, &E_GL_KHR_memory_scope_semantics, fnCandidate.getName().c_str());
memorySemanticsCheck(loc, fnCandidate, callNode);
if ((callNode.getOp() == EOpAtomicAdd || callNode.getOp() == EOpAtomicExchange ||
callNode.getOp() == EOpAtomicLoad || callNode.getOp() == EOpAtomicStore) &&
(arg0->getType().isFloatingDomain())) {
requireExtensions(loc, 1, &E_GL_EXT_shader_atomic_float, fnCandidate.getName().c_str());
}
} else if (arg0->getType().getBasicType() == EbtInt64 || arg0->getType().getBasicType() == EbtUint64) {
const char* const extensions[2] = { E_GL_NV_shader_atomic_int64,
E_GL_EXT_shader_atomic_int64 };
requireExtensions(loc, 2, extensions, fnCandidate.getName().c_str());
} else if ((callNode.getOp() == EOpAtomicAdd || callNode.getOp() == EOpAtomicExchange) &&
(arg0->getType().isFloatingDomain())) {
requireExtensions(loc, 1, &E_GL_EXT_shader_atomic_float, fnCandidate.getName().c_str());
}
break;
}
......
glslang/MachineIndependent/Versions.cpp
......@@ -347,6 +347,7 @@ void TParseVersions::initializeExtensionBehavior()
extensionBehavior[E_GL_EXT_shader_subgroup_extended_types_int16] = EBhDisable;
extensionBehavior[E_GL_EXT_shader_subgroup_extended_types_int64] = EBhDisable;
extensionBehavior[E_GL_EXT_shader_subgroup_extended_types_float16] = EBhDisable;
extensionBehavior[E_GL_EXT_shader_atomic_float] = EBhDisable;
}
#endif // GLSLANG_WEB
......@@ -520,6 +521,8 @@ void TParseVersions::getPreamble(std::string& preamble)
"#define GL_EXT_shader_subgroup_extended_types_int16 1\n"
"#define GL_EXT_shader_subgroup_extended_types_int64 1\n"
"#define GL_EXT_shader_subgroup_extended_types_float16 1\n"
"#define GL_EXT_shader_atomic_float 1\n"
;
if (version >= 150) {
......
glslang/MachineIndependent/Versions.h
......@@ -298,6 +298,8 @@ const char* const E_GL_EXT_shader_subgroup_extended_types_int16 = "GL_EXT_shad
const char* const E_GL_EXT_shader_subgroup_extended_types_int64 = "GL_EXT_shader_subgroup_extended_types_int64";
const char* const E_GL_EXT_shader_subgroup_extended_types_float16 = "GL_EXT_shader_subgroup_extended_types_float16";
const char* const E_GL_EXT_shader_atomic_float = "GL_EXT_shader_atomic_float";
// Arrays of extensions for the above AEP duplications
const char* const AEP_geometry_shader[] = { E_GL_EXT_geometry_shader, E_GL_OES_geometry_shader };
......
gtests/Spv.FromFile.cpp
......@@ -603,6 +603,8 @@ INSTANTIATE_TEST_CASE_P(
"spv.460.vert",
"spv.460.comp",
"spv.atomic.comp",
"spv.atomicFloat.comp",
"spv.atomicFloat_Error.comp",
"spv.glFragColor.frag",
"spv.rankShift.comp",
"spv.specConst.vert",
......
known_good.json
......@@ -5,14 +5,14 @@
"site" : "github",
"subrepo" : "KhronosGroup/SPIRV-Tools",
"subdir" : "External/spirv-tools",
"commit" : "a1fb255a2a6856bba1e5a8486469ac93fca518c2"
"commit" : "e4aebf99fa14f2cd00026a56b21f7a0cf9940bbd"
},
{
"name" : "spirv-tools/external/spirv-headers",
"site" : "github",
"subrepo" : "KhronosGroup/SPIRV-Headers",
"subdir" : "External/spirv-tools/external/spirv-headers",
"commit" : "ac638f1815425403e946d0ab78bac71d2bdbf3be"
"commit" : "979924c8bc839e4cb1b69d03d48398551f369ce7"
}
]
}
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment