3
<TITLE>Shading Language Support</TITLE>
5
<link rel="stylesheet" type="text/css" href="mesa.css"></head>
9
<H1>Shading Language Support</H1>
12
This page describes the features and status of Mesa's support for the
13
<a href="http://opengl.org/documentation/glsl/" target="_parent">
14
OpenGL Shading Language</a>.
18
Last updated on 28 March 2007.
25
<li><a href="#unsup">Unsupported Features</a>
26
<li><a href="#notes">Implementation Notes</a>
27
<li><a href="#hints">Programming Hints</a>
28
<li><a href="#standalone">Stand-alone Compiler</a>
29
<li><a href="#implementation">Compiler Implementation</a>
30
<li><a href="#validation">Compiler Validation</a>
35
<h2>Unsupported Features</h2>
38
The following features of the shading language are not yet supported
43
<li>Dereferencing arrays with non-constant indexes
44
<li>Comparison of user-defined structs
45
<li>Linking of multiple shaders is not supported
47
<li>The derivative functions such as dFdx() are not implemented
48
<li>The inverse trig functions asin(), acos(), and atan() are not implemented
49
<li>The gl_Color and gl_SecondaryColor varying vars are interpolated
50
without perspective correction
51
<li>Floating point literal suffixes 'f' and 'F' aren't allowed.
55
All other major features of the shading language should function.
60
<h2>Implementation Notes</h2>
63
<li>Shading language programs are compiled into low-level programs
64
very similar to those of GL_ARB_vertex/fragment_program.
65
<li>All vector types (vec2, vec3, vec4, bvec2, etc) currently occupy full
67
<li>Float constants and variables are packed so that up to four floats
68
can occupy one program parameter/register.
69
<li>All function calls are inlined.
70
<li>Shaders which use too many registers will not compile.
71
<li>The quality of generated code is pretty good, register usage is fair.
72
<li>Shader error detection and reporting of errors (InfoLog) is not
74
<li>The ftransform() function doesn't necessarily match the results of
75
fixed-function transformation.
79
These issues will be addressed/resolved in the future.
84
<h2>Programming Hints</h2>
87
<li>Declare <em>in</em> function parameters as <em>const</em> whenever possible.
88
This improves the efficiency of function inlining.
91
<li>To reduce register usage, declare variables within smaller scopes.
92
For example, the following code:
97
gl_Position = expression using a1, a2.
98
gl_Color = expression using b1, b2;
101
Can be rewritten as follows to use half as many registers:
107
gl_Position = expression using a1, a2.
111
gl_Color = expression using b1, b2;
115
Alternately, rather than using several float variables, use
116
a vec4 instead. Use swizzling and writemasks to access the
117
components of the vec4 as floats.
120
<li>Use the built-in library functions whenever possible.
121
For example, instead of writing this:
123
float x = 1.0 / sqrt(y);
127
float x = inversesqrt(y);
130
Use ++i when possible as it's more efficient than i++
135
<a name="standalone">
136
<h2>Stand-alone Compiler</h2>
139
A unique stand-alone GLSL compiler driver has been added to Mesa.
143
The stand-alone compiler (like a conventional command-line compiler)
144
is a tool that accepts Shading Language programs and emits low-level
149
This tool is useful for:
152
<li>Inspecting GPU code to gain insight into compilation
153
<li>Generating initial GPU code for subsequent hand-tuning
154
<li>Debugging the GLSL compiler itself
158
To build the glslcompiler program (this will be improved someday):
163
cd drivers/glslcompiler
169
Here's an example of using the compiler to compile a vertex shader and
170
emit GL_ARB_vertex_program-style instructions:
173
glslcompiler --arb --linenumbers --vs vertshader.txt
176
The output may look similar to this:
180
0: MOV result.texcoord[0], vertex.texcoord[0];
181
1: DP4 temp0.x, state.matrix.mvp.row[0], vertex.position;
182
2: DP4 temp0.y, state.matrix.mvp.row[1], vertex.position;
183
3: DP4 temp0.z, state.matrix.mvp.row[2], vertex.position;
184
4: DP4 temp0.w, state.matrix.mvp.row[3], vertex.position;
185
5: MOV result.position, temp0;
190
Note that some shading language constructs (such as uniform and varying
191
variables) aren't expressible in ARB or NV-style programs.
192
Therefore, the resulting output is not always legal by definition of
193
those program languages.
196
Also note that this compiler driver is still under development.
197
Over time, the correctness of the GPU programs, with respect to the ARB
198
and NV languagues, should improve.
203
<a name="implementation">
204
<h2>Compiler Implementation</h2>
207
The source code for Mesa's shading language compiler is in the
208
<code>src/mesa/shader/slang/</code> directory.
212
The compiler follows a fairly standard design and basically works as follows:
215
<li>The input string is tokenized (see grammar.c) and parsed
216
(see slang_compiler_*.c) to produce an Abstract Syntax Tree (AST).
217
The nodes in this tree are slang_operation structures
218
(see slang_compile_operation.h).
219
The nodes are decorated with symbol table, scoping and datatype information.
220
<li>The AST is converted into an Intermediate representation (IR) tree
221
(see the slang_codegen.c file).
222
The IR nodes represent basic GPU instructions, like add, dot product,
224
The IR tree is mostly a binary tree, but a few nodes have three or four
226
In principle, the IR tree could be executed by doing an in-order traversal.
227
<li>The IR tree is traversed in-order to emit code (see slang_emit.c).
228
This is also when registers are allocated to store variables and temps.
229
<li>In the future, a pattern-matching code generator-generator may be
230
used for code generation.
231
Programs such as L-BURG (Bottom-Up Rewrite Generator) and Twig look for
232
patterns in IR trees, compute weights for subtrees and use the weights
233
to select the best instructions to represent the sub-tree.
234
<li>The emitted GPU instructions (see prog_instruction.h) are stored in a
235
gl_program object (see mtypes.h).
236
<li>When a fragment shader and vertex shader are linked (see slang_link.c)
237
the varying vars are matched up, uniforms are merged, and vertex
238
attributes are resolved (rewriting instructions as needed).
242
The final vertex and fragment programs may be interpreted in software
243
(see prog_execute.c) or translated into a specific hardware architecture
244
(see drivers/dri/i915/i915_fragprog.c for example).
247
<h3>Code Generation Options</h3>
250
Internally, there are several options that control the compiler's code
251
generation and instruction selection.
252
These options are seen in the gl_shader_state struct and may be set
253
by the device driver to indicate its preferences:
256
struct gl_shader_state
259
/** Driver-selectable options: */
260
GLboolean EmitHighLevelInstructions;
261
GLboolean EmitCondCodes;
262
GLboolean EmitComments;
267
<li>EmitHighLevelInstructions
269
This option controls instruction selection for loops and conditionals.
270
If the option is set high-level IF/ELSE/ENDIF, LOOP/ENDLOOP, CONT/BRK
271
instructions will be emitted.
272
Otherwise, those constructs will be implemented with BRA instructions.
277
If set, condition codes (ala GL_NV_fragment_program) will be used for
278
branching and looping.
279
Otherwise, ordinary registers will be used (the IF instruction will
280
examine the first operand's X component and do the if-part if non-zero).
281
This option is only relevant if EmitHighLevelInstructions is set.
286
If set, instructions will be annoted with comments to help with debugging.
287
Extra NOP instructions will also be inserted.
293
<a name="validation">
294
<h2>Compiler Validation</h2>
297
A new <a href="http://glean.sf.net" target="_parent">Glean</a> test has
298
been create to exercise the GLSL compiler.
301
The <em>glsl1</em> test runs over 150 sub-tests to check that the language
302
features and built-in functions work properly.
303
This test should be run frequently while working on the compiler to catch
307
The test coverage is reasonably broad and complete but additional tests