I forgot to ask you a question.

The current GPU version only works with one collision object. I want to correct this issue but OpenCL don't manage complex arrays. It is necessary to define all the elements of a collision object independently. So I want to limit the number of collision objects in the program... for GPU and for the CPU. Do you think it would be fine a have this limitation defined to 3. I think this allows to create all the simulations but I prefer to have your opinion.

____________________________________________

Follow me on Twitter : @VWDynamics

Watch demo videos on Youtube