Here is a work in progress of my first working CUDA TOP. With inspiration from Jeffers, it’s a Voronoi cell diagram generator. One strange thing I am noticing is that if I run it at 1000x1000 it lags signifigantly, but if I run three copies of the TOP at 500x500 each it runs much faster.
voronoi.zip (13.4 KB)
Very cool. Is there anything about the CUDA interface that you found lacking that I could improve?
It would be helpful to have some examples of using transforms and chop data in the kernel. I still haven’t totally figured that out. Also, I’m wondering if there is an efficient way to pass 3d texture info to a GLSL TOP. For example, if you created a particle system in CUDA thats rendered in GLSL.