I just wondered if anyone had done a comparison of Bullet's own btCompoundShape, vs. the GIMPACT btGImpactCompoundShape/btGImpactConvexDecompositionShape - in particular, for convex decomposition. They both seem to serve the same purpose... so is one faster? More stable? Simpler to implement?
I'm trying to work out which one to use and am feeling too lazy right now to try to implement both

Thanks!
Kate