I have since a while thought about creating a game where thinks do not move the usual way but use local interactions(like cellular automata)
or execute transformations on an image(like warping space ^^)
My problem is I do not accidentally want to do stuff on my PC that that the graphic card is doing anyway or save data that could actually remain in the ram.
My knowledge of the workings of a graphic card, which classes in c# or java access those or using a 3D engine or physics engine are pretty slim.
(I am actually not even sure i get the difference between a 3D engine and a physics engine).
So I need pointers where to look for information, maybe some explanation where engines help and where they limit such an procedural approach.
(What data do i get from the engines? What input can i give?)
A little bit of trivia about graphic card function and tips how to access them for procedural graphics would also be helpful.