Smoothing geometry on the GPU using OpenCL.

In VEX it has to be done in 2 steps where the same thing can be achieved in OpenCL in one step using barriers [khronos.org].

For a 822K polygon model, it's over 500 times faster than VEX on GTX 970 for 1000 iterations, and over 600 times faster for 10K iterations.

Timings are recorded separately as there seems to be some overhead when running them one after another.

Special thanks to SESI.

Sorry for the necro, but why does VEX need two steps? I am trying to write my own smooth in VEX right now and it is moving towards the average distance of its neighbors… but it is inside a SOP Solver. What would the two steps be in that case?

Because it's using Taubin's smooth paper which has 2 steps (positive and negative = lambda and mu).