Metaprogramming GPUs with Sh

This book is a high-level overview of Sh and its relationship to other realtime shading and Graphics processing unit programming languages. It is a reference manual and language specification and methodically and exhaustively presents details of the various features of Sh.

Author Notes

McCool , Michael

List of Tables	p. xi
List of Listings	p. xv
Preface	p. xvii
I Introduction	p. 1
1 Overview	p. 3
1.1 Programmable GPUs	p. 8
1.2 Shading Languages	p. 10
1.3 Stream Processing	p. 13
1.4 Metaprogramming	p. 16
1.5 The Sh Metaprogramming Toolkit	p. 19
1.5.1 Immediate Mode	p. 20
1.5.2 Stream Programming	p. 20
1.5.3 Shader Programming	p. 24
1.6 Conventions and Notation	p. 27
1.6.1 Typesetting and Stylistic Conventions	p. 27
1.6.2 Coordinate Systems	p. 28
1.6.3 Standard Vectors and Variable Names	p. 29
1.6.4 Global Parameters	p. 32
1.7 License, History, and Access	p. 35
1.8 Status and Future Development	p. 36
1.8.1 Beta Release Features	p. 37
1.8.2 Year-End Features	p. 38
1.8.3 Future Features	p. 40
2 Tuples, Matrices, Operators, and Functions	p. 43
2.1 Tuples	p. 43
2.2 Operators and Library Functions	p. 44
2.3 Matrices	p. 46
2.4 Functions	p. 47
2.4.1 Example: Glass Support Functions	p. 47
2.4.2 Example: Reflectance Models	p. 50
3 Programs, Parameters, and Attributes	p. 63
3.1 Programs	p. 63
3.2 Example: Blinn-Phong Point-Source Lighting	p. 64
3.2.1 Vertex Shader	p. 64
3.2.2 Fragment Shader	p. 67
3.2.3 Modularity	p. 68
3.2.4 Shader Framework	p. 71
3.3 Shader Algebra	p. 72
3.4 Example: Surfaces, Lights, and Postprocessing	p. 74
3.5 Example: Metaprogrammed CSG Text	p. 77
3.6 Control Constructs	p. 83
3.7 Example: Julia Set	p. 84
4 Arrays and Textures	p. 89
4.1 Texture Types	p. 89
4.2 Example: Texture Mapped Blinn-Phong	p. 91
4.3 Example: Encapsulated BRDF Representation	p. 93
4.4 Example: Shiny Bump Map	p. 96
4.4.1 Vertex Shader	p. 97
4.4.2 Fragment Shader	p. 98
4.5 Example: Glass	p. 98
4.6 Example: Lucite	p. 100
4.7 Example: Wood and Marble	p. 102
4.7.1 Parameters	p. 103
4.7.2 Vertex Shader	p. 104
4.7.3 Fragment Shader	p. 104
4.8 Example: Worley Noise Shaders	p. 105
4.9 Example: Wood from Phong	p. 106
5 Streams and Channels	p. 119
5.1 Channels	p. 119
5.2 Combining Channels into Streams	p. 119
5.3 Applying Programs to Streams	p. 120
5.4 Example: Particle Systems	p. 121
II Reference	p. 127
6 Reference Overview	p. 129
6.1 Setup	p. 129
6.2 Sample OpenGL GLUT Application	p. 130
6.3 The shrike Testbed	p. 134
6.4 API Documentation	p. 135
7 Types	p. 137
7.1 Programs	p. 137
7.2 Tuples	p. 138
7.2.1 Semantic Types	p. 138
7.2.2 Storage Types	p. 139
7.2.3 Binding Types	p. 141
7.2.4 Template Declaration	p. 142
7.3 Matrices	p. 144
7.4 Arrays, Tables, and Textures	p. 145
7.5 Contexts	p. 146
7.6 Exceptions and Error Handling	p. 146
7.7 Metadata	p. 146
7.7.1 Object Names	p. 148
8 Standard Operators and Functions	p. 151
8.1 Swizzling and Write Masking	p. 153
8.2 Arithmetic	p. 156
8.3 Derivatives	p. 159
8.4 Noise	p. 161
8.5 Trigonometric and Exponential Functions	p. 163
8.6 Interpolation and Approximation	p. 165
8.7 Geometry	p. 166
8.8 Linear Algebra	p. 167
8.9 Logical and Comparison Functions	p. 171
8.10 Discontinuities	p. 173
8.11 Miscellaneous Functions	p. 173
9 Arrays, Tables and Textures	p. 177
9.1 Texture Formats	p. 178
9.2 Trait Modifiers	p. 180
9.3 Texture Memory and Storage	p. 180
9.4 Texture and Array Access	p. 181
10 Programs and Streams	p. 183
10.1 Defining Programs	p. 183
10.1.1 Compiling Programs	p. 184
10.2 Binding Programs and Data	p. 185
10.2.1 Binding Programs	p. 185
10.2.2 Binding and Updating Data	p. 186
10.2.3 Unbinding	p. 186
10.2.4 Querying Binding State	p. 186
10.3 Introspection	p. 187
10.4 Algebra	p. 189
10.4.1 Connection	p. 190
10.4.2 Combination	p. 190
10.5 Nibbles	p. 191
10.5.1 Interface Adaptation and Specialization	p. 191
10.5.2 Passthrough	p. 192
10.5.3 Texture Access	p. 193
10.5.4 Type and Size Conversion	p. 193
10.5.5 Transformations	p. 194
10.5.6 Basis Conversion	p. 194
10.5.7 Primitive Computations	p. 195
10.6 Manipulators	p. 196
10.6.1 Fixed Manipulators	p. 197
10.6.2 Expandable Manipulators	p. 197
10.7 Kernels	p. 198
10.7.1 Universal Vertex Shader Kernel	p. 199
10.7.2 Surface Shader Kernels	p. 201
10.7.3 Light Shader Kernels	p. 203
10.7.4 Surface Map Kernels	p. 205
10.7.5 Postprocessing Kernels	p. 206
10.8 Streams	p. 206
11 Control Constructs	p. 209
11.1 Selection	p. 210
11.1.1 IF	p. 211
11.1.2 When	p. 212
11.2 Iteration	p. 212
11.2.1 For	p. 213
11.2.2 While	p. 214
11.2.3 Do/Until	p. 214
11.2.4 Break/Continue	p. 214
12 Backends	p. 217
12.1 OpenGL Backends	p. 218
12.1.1 Input and Output Binding	p. 218
12.1.2 Texture Bindings	p. 219
12.1.3 Parameter Bindings	p. 220
12.2 CPU Backends	p. 220
12.3 The Sm Backend	p. 221
III Engineering	p. 223
13 Engineering Overview	p. 225
13.1 Code Organization and Build System	p. 226
13.2 Template Instantiation	p. 227
14 Intermediate Representation	p. 229
14.1 Reference Counting	p. 229
14.2 Basic Blocks	p. 230
14.2.1 Special Statements	p. 234
14.3 The Control-Flow Graph	p. 235
14.4 Parsing	p. 236
14.4.1 Basic Blocks	p. 236
14.4.2 Control Constructs	p. 237
15 Streams and Textures	p. 241
15.1 Stream Representation	p. 241
15.2 Texture Representation	p. 242
15.3 Stream and Texture Memory	p. 242
15.3.1 Memory and Storage Objects	p. 242
15.3.2 Finding Storages	p. 244
15.3.3 Transfer Operators	p. 244
15.3.4 Existing Implementations	p. 244
15.4 Future Work	p. 245
15.4.1 Stream Interface Improvements	p. 245
15.4.2 Index Streams	p. 246
15.4.3 Reductions	p. 246
15.4.4 Stream Ordering	p. 246
15.4.5 Memory Abstraction Improvements	p. 247
16 Optimization	p. 249
16.1 Copy Propagation and Forward Substitution	p. 249
16.2 Dead Code Removal	p. 251
16.2.1 Block Predecessors	p. 252
16.2.2 Reaching Definitions	p. 252
16.2.3 Use-Def Chains	p. 252
16.2.4 Live and Dead Code	p. 253
16.3 Graph Straightening	p. 253
16.4 Future Work	p. 254
16.4.1 Eliminating Redundancies	p. 254
16.4.2 Constant Folding and Uniform Lifting	p. 255
16.4.3 Algebraic Simplification	p. 256
16.4.4 Conditional Assignment	p. 257
16.4.5 Static Single Assignment and Other Forms	p. 257
17 Backend	p. 261
17.1 The Backend Interface	p. 261
17.2 Linear Register Allocation	p. 263
17.3 Transformers	p. 263
17.3.1 Splitting Long Tuples	p. 264
17.3.2 Input and Output Conversion	p. 265
17.3.3 Texture Indexing Conversion	p. 266
17.4 Future Work	p. 266
17.4.1 New Backends	p. 266
17.4.2 Better Transformations	p. 267
17.4.3 Better Reuse	p. 267
17.4.4 Instruction Count Virtualization	p. 268
17.4.5 Data-Dependent Control Flow Virtualization	p. 268
Bibliography	p. 271
Index	p. 285

Available:*

On Order

Summary

Summary

Author Notes

Table of Contents