We are very excited for next week's Shading Language Symposium in San Diego!
This event will bring together graphics and computer shader programmers, researchers and technical artists with shading language implementers to explore the landscape of shading languages, their future development, and the role how new techniques such as neural graphics will play in shaping the future of real-time rendering.
Do any GPGPU people happen to know why compute shaders like this freeze my desktop until they complete (Vulkan + Mesa + AMD)? The call to memoryBarrierBuffer is load bearing, without it the freeze doesn't occur
Am I doing something wrong? Is there a better way to structure operations like this?
for (uint i = 0; i < pushConstants.stepCount; i++)
{
// do some work based on neighbors
myData[outputIndex].values[index] = myData[inputIndex].values[nextIndex];
// wait for neighbors to update
memoryBarrierBuffer();
Early Bird pricing for the Shading Language Symposium ends on January 18th. Register now to attend this inaugural, ground-breaking event that will bring together graphics and computer shader programmers, researchers and technical artists with shading language implementers to explore the landscape of shading languages, their future development, and new techniques.
Intentional camera movement: Open source software is used to exaggerate the camera shaking directly during recording. This image was taken on a smartphone in a browser using GLSL.