I am searching for an example how to create a compute shaders so that the results of each thread group can be analyzed and compared after execution.
I am unsure how to set this up. In my use case I need to iterate over an memory area of 800x600 created by procedural 2D SDFs to find the point of the largest distance into the shape (no textures used). How to split this up over several thread groups ? I have it running on one thread but it’s way to slow.