Oregon State University Computer Graphics
OpenGL Compute Shaders
Mike Bailey mjb@cs.oregonstate.edu
Oregon State University
compute.shader.pptx
Oregon State University Computer Graphics
A Shader Program, with only a Compute Shader in it
Application Invokes the Compute Shader to Modify the OpenGL Buffer Data
Application Invokes OpenGL Rendering which Reads the Buffer Data
Another Shader Program, with pipeline rendering in it
mjb – January1, 2019
Why Not Just Use OpenCL Instead? 4
OpenGL Compute Shader – the Basic Idea
Paraphrased from the ARB_compute_shader spec:
mjb – January1, 2019
3
Recent graphics hardware has become extremely powerful. A strong desire to harness this power for work that does not fit the traditional graphics pipeline has emerged. To address this, Compute Shaders are a new single-stage program. They are launched in a manner that is essentially stateless. This allows arbitrary workloads to be sent to the graphics hardware with minimal disturbance to the GL state machine.
In most respects, a Compute Shader is identical to all other OpenGL shaders, with similar status, uniforms, and other such properties. It has access to many of the same data as all other shader types, such as textures, image textures, atomic counters, and so on. However, the Compute Shader has no predefined inputs, nor any fixed-function outputs. It cannot be part of a rendering pipeline and its visible side effects are through its actions on shader storage buffers, image textures, and atomic counters.
Oregon State University Computer Graphics
•
• • • •
OpenCL requires installing a separate driver and separate libraries. While this is not a huge deal, it does take time and effort. When everyone catches up to OpenGL 4.3, Compute Shaders will just “be there” as part of core OpenGL.
Compute Shaders use the GLSL language, something that all OpenGL programmers should already be familiar with (or will be soon).
Compute shaders use the same context as does the OpenGL rendering pipeline. There is no need to acquire and release the context as OpenGL+OpenCL must do.
I’m assuming that calls to OpenGL compute shaders are more lightweight than calls to OpenCL kernels are. (true?) This should result in better performance. (true? how much?)
Using OpenCL is somewhat cumbersome. It requires a lot of setup (queries, platforms, devices, queues, kernels, etc.). Compute Shaders look to be more convenient. They just kind of flow in with the graphics.
The bottom line is that I will continue to use OpenCL for the big, bad stuff. But, for lighter-weight data-parallel computing that interacts with graphics, I will use the Compute Shaders.
I suspect that a good example of a lighter-weight data-parallel graphics-related application is a pOarretgiocnleStastye sUtneivmers.ityThis will be shown here in the rest of these notes. I hope I’m right.
Computer Graphics
mjb – January1, 2019
Passing Data to the Compute Shader Happens with a Cool 6 New Buffer Type – the Shader Storage Buffer Object
The tricky part is getting data into and out of the Compute Shader. The trickiness comes from the specification phrase: “In most respects, a Compute Shader is identical to all other OpenGL shaders, with similar status, uniforms, and other such properties. It has access to many of the same data as all other shader types, such as textures, image textures, atomic counters, and so on.”
OpenCL programs have access to general arrays of data, and also access to OpenGL arrays of data in the form of buffer objects. Compute Shaders, looking like other shaders, haven’t had direct access to general arrays of data (hacked access, yes; direct access, no). But, because Compute Shaders represent opportunities for massive data-parallel computations, that is exactly what you want them to use.
Thus, OpenGL 4.3 introduced the Shader Storage Buffer Object. This is very cool, and has been needed for a long time!
If I Know GLSL,
What Do I Need to Do Differently to Write a Compute Shader?
Not much:
1. A Compute Shader is created just like any other GLSL shader, except that its type is GL_COMPUTE_SHADER (duh…). You compile it and link it just like any other GLSL shader program.
2. A Compute Shader must be in a shader program all by itself. There cannot be vertex, fragment, etc. shaders in there with it. (why?)
3. A Compute Shader has access to uniform variables and buffer objects, but cannot access any pipeline variables such as attributes or variables from other stages. It stands alone.
4. A Compute Shader needs to declare the number of work-items in each of its work-groups in a special GLSL layout statement.
5
Oregon State University Computer Graphics
More information on items 3 and 4 are coming up . . .
Shader Storage Buffer Object
Oregon State University Computer Graphics
Arbitrary data, including Arrays of Structures
Shader Storage Buffer Objects are created with arbitrary data (same as other buffer objects), but what is new is that the shaders can read and write them in the same C-like way as they were created, including treating parts of the buffer as an array of structures – perfect for data- parallel computing!
mjb – January1, 2019
mjb – January1, 2019
mjb – January1, 2019
1
OpenGL Compute Shader – the Basic Idea 2
OpenCL is great! It does a super job of using the GPU for general-purpose data-parallel computing. And, OpenCL is more feature-rich than OpenGL compute shaders. So, why use Compute Shaders ever if you’ve got OpenCL? Here’s what I think:
1
Passing Data to the Compute Shader Happens with a Cool New Buffer Type – the Shader Storage Buffer Object
7
The Example We Are Going to Use Here is a Particle System 8
Shader Storage Buffer Object
Oregon State University Computer Graphics
And, like other OpenGL buffer types, Shader Storage Buffer Objects can be bound to indexed binding points, making them easy to access from inside the Compute Shaders.
The Compute Shader Moves the Particles by Recomputing the Position and Velocity Buffers
The OpenGL Rendering Draws the Particles by Reading the Position Buffer
Oregon State University Computer Graphics
Texture0 Texture1 Texture2 Texture3
OpenGL Context
Buffer0 Buffer1 Buffer2. Buffer3.
Display Dest.
Setting up the Shader Storage Buffer Objects in Your C Program
9
Setting up the Shader Storage Buffer Objects in Your C Program 10 glGenBuffers( 1, &posSSbo);
glBindBuffer( GL_SHADER_STORAGE_BUFFER, posSSbo );
glBufferData( GL_SHADER_STORAGE_BUFFER, NUM_PARTICLES * sizeof(struct pos), NULL, GL_STATIC_DRAW );
GLint bufMask = GL_MAP_WRITE_BIT | GL_MAP_INVALIDATE_BUFFER_BIT ; // the invalidate makes a big difference when re-writing
struct pos *points = (struct pos *) glMapBufferRange( GL_SHADER_STORAGE_BUFFER, 0, NUM_PARTICLES * sizeof(struct pos), bufMask ); for( int i = 0; i < NUM_PARTICLES; i++ )
{
points[ i ].x = Ranf( XMIN, XMAX ); points[ i ].y = Ranf( YMIN, YMAX ); points[ i ].z = Ranf( ZMIN, ZMAX ); points[ i ].w = 1.;
}
glUnmapBuffer( GL_SHADER_STORAGE_BUFFER );
glGenBuffers( 1, &velSSbo);
glBindBuffer( GL_SHADER_STORAGE_BUFFER, velSSbo );
glBufferData( GL_SHADER_STORAGE_BUFFER, NUM_PARTICLES * sizeof(struct vel), NULL, GL_STATIC_DRAW );
struct vel *vels = (struct vel *) glMapBufferRange( GL_SHADER_STORAGE_BUFFER, 0, NUM_PARTICLES * sizeof(struct vel), bufMask ); for( int i = 0; i < NUM_PARTICLES; i++ )
{
vels[ i ].vx = Ranf( VXMIN, VXMAX ); vels[ i ].vy = Ranf( VYMIN, VYMAX ); vels[ i ].vz = Ranf( VZMIN, VZMAX ); vels[ i ].vw = 0.;
}
glUnmapBuffer( GL_SHADER_STORAGE_BUFFER );
Oregon State UnTivheersistyame would possibly need to be done for the color shader storage buffer object Computer Graphics
mjb – January1, 2019
The Data Needs to be Divided into Large Quantities call Work-Groups, each of 12 which is further Divided into Smaller Units Called Work-Items
#define NUM_PARTICLES #define WORK_GROUP_SIZE
1024*1024 128
// positions
// velocities
// colors
// total number of particles to move // # work-items per work-group
struct pos {
};
struct vel {
};
struct color {
};
float x, y, z, w;
float vx, vy, vz, vw;
float r, g, b, a;
// need to do the following for both position, velocity, and colors of the particles:
GLuint posSSbo; GLuint velSSbo GLuint colSSbo;
Note that .w and .vw are not actually needed. But, by making these structure sizes a multiple of 4
floats, it doesn’t matter if they are declared with the std140 or the std430 qualifier. I think this is a
Oregon State University
good thing.
Computer Graphics
mjb – January1, 2019
The Data Needs to be Divided into Large Quantities call Work-Groups, each of 11 which is further Divided into Smaller Units Called Work-Items
20 total items to compute:
5 Work Groups
#WorkGroups GlobalInvocationSize WorkGroupSize
The Invocation Space can be 1D, 2D, or 3D. This one is 1D.
20x12 (=240) total items to compute:
5 Work-Groups
#WorkGroups GlobalInvocationSize WorkGroupSize
The Invocation Space can be 1D, 2D, or 3D. This one is 2D.
Oregon State University Computer Graphics
Oregon State University Computer Graphics
5x4 20 4
5x4 20x12 4x3
4 Work-Items
(Any resemblance this diagram has to a mother sow is accidental, but not entirely inaccurate...)
4 Work-Items
mjb – January1, 2019
mjb – January1, 2019
3 Work-Items
4 Work-Groups
mjb – January1, 2019
mjb – January1, 2019
2
in in in in
uvec3 uvec3 uvec3 uint
};
Running the Compute Shader from the Application
void glDispatchCompute( num_groups_x, num_groups_y, num_groups_z );
13
A Mechanical Equivalent...
14
“Streaming Multiprocessor”
“CUDA Cores” “Data”
mjb – January1, 2019
16
num_groups_y
Oregon State University Computer Graphics
Oregon State University Computer Graphics
http://news.cision.com
≤ gl_WorkGroupID
≤ gl_LocalInvocationID
0
0 gl_GlobalInvocationID gl_LocalInvocationIndex
Oregon State University Computer Graphics
≤
gl_NumWorkGroups – 1
num_groups_x
If the problem is 2D, then num_groups_z = 1
If the problem is 1D, then num_groups_y = 1 and num_groups_z = 1
Invoking the Compute Shader in Your C Program
glBindBufferBase( GL_SHADER_STORAGE_BUFFER, 4, posSSbo ); glBindBufferBase( GL_SHADER_STORAGE_BUFFER, 5, velSSbo ); glBindBufferBase( GL_SHADER_STORAGE_BUFFER, 6, colSSbo );
...
glUseProgram( MyComputeShaderProgram );
glDispatchCompute( NUM_PARTICLES / WORK_GROUP_SIZE, 1, 1 ); glMemoryBarrier( GL_SHADER_STORAGE_BARRIER_BIT );
...
glUseProgram( MyRenderingShaderProgram ); glBindBuffer( GL_ARRAY_BUFFER, posSSbo ); glVertexPointer( 4, GL_FLOAT, 0, (void *)0 ); glEnableClientState( GL_VERTEX_ARRAY ); glDrawArrays( GL_POINTS, 0, NUM_PARTICLES ); glDisableClientState( GL_VERTEX_ARRAY ); glBindBuffer( GL_ARRAY_BUFFER, 0 );
Oregon State University Computer Graphics
mjb – January1, 2019
15
Writing a C++ Class to Handle Everything is Fairly Straightforward
in uvec3 const uvec3
gl_NumWorkGroups ; gl_WorkGroupSize ; gl_WorkGroupID ; gl_LocalInvocationID ; gl_GlobalInvocationID ; gl_LocalInvocationIndex ;
Same numbers as in the glDispatchCompute call Same numbers as in the layout local_size_* Which workgroup this thread is in
Where this thread is in the current workgroup Where this thread is in all the work items
1D representation of the gl_LocalInvocationID (used for indexing into a shared array)
#version 430 compatibility
#extension GL_ARB_compute_shader :
#extension GL_ARB_shader_storage_buffer_object :
layout( std140, binding=4 ) buffer Pos {
enable enable;
Special Pre-set Variables in the Compute Shader
17
The Particle System Compute Shader -- Setup 18
≤
= gl_WorkGroupID * gl_WorkGroupSize +
gl_WorkGroupSize – 1
= gl_LocalInvocationID.z * gl_WorkGroupSize.y * gl_WorkGroupSize.x + gl_LocalInvocationID.y * gl_WorkGroupSize.x + gl_LocalInvocationID.x
gl_LocalInvocationID
mjb – January1, 2019
mjb – January1, 2019
mjb – January1, 2019
Setup:
GLSLProgram *Particles = new GLSLProgram( ); bool valid = Particles>Create( “particles.cs” );
if( ! valid ) { . . . }
Using:
Particles->Use( );
Particles->DispatchCompute( NUM_PARTICLES / WORK_GROUP_SIZE, 1, 1 );
Render->Use( ); …
Oregon State University Computer Graphics
vec4 Positions[ ];
// array of structures
// array of structures
// array of structures
You can use the empty brackets, but only on the last element of the buffer. The actual dimension will be determined for you when OpenGL examines the size of this buffer’s data store.
mjb – January1, 2019
layout( std140, binding=5 ) buffer {
vec4 Velocities[ ];
};
layout( std140, binding=6 ) buffer Col {
vec4 Colors[ ];
};
layout( local_size_x = 128, local_size_y = 1, local_size_z = 1 ) in;
Oregon State University Computer Graphics
Vel
3
const vec3 G = vec3( 0., -9.8, 0. ); const float DT = 0.1;
…
uint gid = gl_GlobalInvocationID.x;
vec3 p = Positions[ gid ].xyz; vec3 v = Velocities[ gid ].xyz;
vec3 pp = p + v*DT + .5*DT*DT*G; vec3 vp = v + G*DT;
Positions[ gid ].xyz = pp; Velocities[ gid ].xyz = vp;
// the .y and .z are both 1 in this case
p’ pvt1Gt2 2
v’vGt
The Particle System Compute Shader – How About Introducing a Bounce?
20
const vec4 Sphere = vec4( -100., -800., 0., 600. ); // x, y, z, r
// (could also have passed this in)
vec3
Bounce( vec3 vin, vec3 n ) {n
vec3 vout = reflect( vin, n ); in out
return vout; }
vec3
BounceSphere( vec3 p, vec3 v, vec4 s ) {
vec3 n = normalize( p – s.xyz );
return Bounce( v, n ); }
bool
IsInsideSphere( vec3 p, vec4 s ) {
float r = length( p – s.xyz );
return ( r < s.w ); }
Oregon State University Computer Graphics
mjb – January1, 2019
The Particle System Compute Shader – The Physics
Oregon State University Computer Graphics
19
mjb – January1, 2019
The Particle System Compute Shader – How About Introducing a Bounce?
21
uint gid = gl_GlobalInvocationID.x;
vec3 p = Positions[ gid ].xyz; vec3 v = Velocities[ gid ].xyz;
vec3 pp = p + v*DT + .5*DT*DT*G; vec3 vp = v + G*DT;
if( IsInsideSphere( pp, Sphere ) ) {
vp = BounceSphere( p, v, Sphere );
pp = p + vp*DT + .5*DT*DT*G; }
Positions[ gid ].xyz = pp; Velocities[ gid ].xyz = vp;
// the .y and .z are both 1 in this case
p' pvt1Gt2 2
v'vGt
Graphics Trick Alert: Making the bounce happen from the surface of the sphere is time-consuming. Instead, bounce from the previous position in space. If DT is small enough, nobody will ever know...
Oregon State University Computer Graphics
mjb – January1, 2019
The Bouncing Particle System Compute Shader – What Does It Look Like?
22
Oregon State University Computer Graphics
mjb – January1, 2019
Other Useful Stuff –
Copying Global Data to a Local Array Shared by the Entire Work-Group
There are some applications, such as image convolution, where threads within a work- group need to operate on each other’s input or output data. In those cases, it is usually a good idea to create a local shared array that all of the threads in the work-group can access. You do it like this:
23
layout( std140, binding=6 ) buffer Col {
vec4 Colors[ ];
};
layout( shared ) vec4 rgba[ gl_WorkGroupSize.x ];
uint gid = gl_GlobalInvocationID.x; uint lid = gl_LocalInvocationID.x;
rgba[ lid ] = Colors[ gid ]; memory_barrier_shared( );
<< operate on the rgba array elements >>
Colors[ gid ] = rgba[ lid ];
Oregon State Computer Graphics
University
mjb – January1, 2019
4