Brian | 9595d19 | 2007-01-20 13:40:57 -0700 | [diff] [blame^] | 1 | <HTML> |
| 2 | |
| 3 | <TITLE>Shading Language Support</TITLE> |
| 4 | |
| 5 | <link rel="stylesheet" type="text/css" href="mesa.css"></head> |
| 6 | |
| 7 | <BODY> |
| 8 | |
| 9 | <H1>Shading Language Support</H1> |
| 10 | |
| 11 | <p> |
| 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>. |
| 15 | </p> |
| 16 | |
| 17 | <p> |
| 18 | Last updated on 20 Jan 2007. |
| 19 | </p> |
| 20 | |
| 21 | <h2>Unsupported Features</h2> |
| 22 | |
| 23 | <p> |
| 24 | The following features of the shading language are not yet supported |
| 25 | in Mesa: |
| 26 | </p> |
| 27 | |
| 28 | <ul> |
| 29 | <li>Arrays |
| 30 | <li>Structs |
| 31 | <li>Linking of multiple shaders is not supported |
| 32 | <li>Noise functions |
| 33 | <li>Not all built-in OpenGL state variables are supported yet. |
| 34 | Common variables such as gl_ModelViewMatrix and gl_NormalMatrix |
| 35 | are supported. |
| 36 | <li>Integer operations are not fully implemented (most are implemented |
| 37 | as floating point). |
| 38 | </ul> |
| 39 | |
| 40 | <p> |
| 41 | All other major features of the shading language should function. |
| 42 | </p> |
| 43 | |
| 44 | |
| 45 | <h2>Implementation Notes</h2> |
| 46 | |
| 47 | <ul> |
| 48 | <li>Shading language programs are compiled into low-level programs |
| 49 | very similar to those of GL_ARB_vertex/fragment_program. |
| 50 | <li>All float/int/bool and vector types currently occupy full |
| 51 | float[4] registers. |
| 52 | <li>Float constants are packed so that up to four floats can occupy one |
| 53 | program parameter/register. |
| 54 | <li>All function calls are inlined. |
| 55 | <li>Shaders which use too many registers will not compile. |
| 56 | <li>The quality of generated code is pretty good, register usage is fair. |
| 57 | <li>Shader error detection and reporting of errors (InfoLog) is not |
| 58 | very good yet. |
| 59 | <li>There are massive memory leaks in the compiler. |
| 60 | </ul> |
| 61 | |
| 62 | <p> |
| 63 | These issues will be addressed/resolved in the future. |
| 64 | </p> |
| 65 | |
| 66 | |
| 67 | <h2>Programming Hints</h2> |
| 68 | |
| 69 | <ul> |
| 70 | <li>Always declare <em>in</em> function parameters as <em>const</em>. |
| 71 | This improves the efficiency of function inlining. |
| 72 | </li> |
| 73 | <br> |
| 74 | <li>To reduce register usage, declare variables within smaller scopes. |
| 75 | For example, the following code: |
| 76 | <pre> |
| 77 | void main() |
| 78 | { |
| 79 | vec4 a1, a2, b1, b2; |
| 80 | gl_Position = expression using a1, a2. |
| 81 | gl_Color = expression using b1, b2; |
| 82 | } |
| 83 | </pre> |
| 84 | Can be rewritten as follows to use half as many registers: |
| 85 | <pre> |
| 86 | void main() |
| 87 | { |
| 88 | { |
| 89 | vec4 a1, a2; |
| 90 | gl_Position = expression using a1, a2. |
| 91 | } |
| 92 | { |
| 93 | vec4 b1, b2; |
| 94 | gl_Color = expression using b1, b2; |
| 95 | } |
| 96 | } |
| 97 | </pre> |
| 98 | Alternately, rather than using several float variables, use |
| 99 | a vec4 instead. Use swizzling and writemasks to access the |
| 100 | components of the vec4 as floats. |
| 101 | </li> |
| 102 | <br> |
| 103 | <li>Use the built-in library functions whenever possible. |
| 104 | For example, instead of writing this: |
| 105 | <pre> |
| 106 | float x = 1.0 / sqrt(y); |
| 107 | </pre> |
| 108 | Write this: |
| 109 | <pre> |
| 110 | float x = inversesqrt(y); |
| 111 | </pre> |
| 112 | </ul> |
| 113 | |
| 114 | |
| 115 | </BODY> |
| 116 | </HTML> |