Rationale for SAT in GimpactCollisionAlgorithm ?

Post Reply
Karrok
Posts: 65
Joined: Fri May 13, 2011 1:11 pm

Rationale for SAT in GimpactCollisionAlgorithm ?

Post by Karrok »

I was wondering what the rationale was for using SAT in the Gimpact triangle collide over GJK.

From what I know, SAT would be faster for convex objects, but breaks down if there are (m)any concave objects.
From an old thread I gathered Gimpact initially went with GJK, but then swapped to SAT.

This was back in 2007, and other than that old topic I can't find a lot of information on it.

So in short, why was SAT defaulted over GJK? what is the performance penalty payed for GJK (if any?) and what is the difference in accuracy between implemented methods for Gimpact objects?

Cheers,

- Karrok
Post Reply