[Haskell-cafe] [GSoC] A data parallel physics engine

[Bryan O'Sullivan]

Roman Cheplyaka wrote:

> I have not very deep knowledge about both NDP and physics engines.

I've done some physics simulation work for games.  One could certainly
learn enough to be able to write a straightforward implementation in
that time.  Broadphase collision detection is easy; narrowphase somewhat
more difficult; integration and interpenetration resolution would
consume the bulk of your time.

I very strongly doubt that one might both write a physics engine and
adapt it to compute anything interesting using NDP in the SoC time
frame.  That's more like a meaty topic for a Masters thesis.

	<b