Does anyone know of an implementation in CUDA of a 3D Voronoi Diagram? If its a Delaunay diagram I will be happy to. As always, I do not want to reinvent the wheel (that is not funny at all ;) ).
I’ve been looking through some papers in the area and I found some, very nice, approaches but I am not sure they can be implemented properly -maximum performance- over CUDA (to many cases to be able to assure that every thread in one warp is executing the same instruction).
If anyone has any information or reference it will be actually very appreciated, it’s estrange that if so paper or program exists is not already into the Nvidia database since this is a very popular algorithm.