[ODE] new trimesh vs trimesh code

[Jon Watte (ODE)]

Oleh Derevenko wrote:
> Well, I'd predef doing this way rather than examining the neighbours. First 
> of all, examining all the neighbours may spoil all the timing benefits from 
> the hashing. Another thing that curently, contact points are groupped by 
>   

I don't think so -- doing the n-squared linear scan is a lot slower when 
n gets big. When n is 10, though, you're probably better off putting 
them all in a linear array and doing brute force, as it'll cache well. 
However, the timing difference will likely not matter for such small a 
contact set.

> rounding the position coordinates to 1e-4 precision - that makes neighbour 
> buckets very small (a cube with edge of 1e-4). However including several 
> close contacts usually makes final dynamics calculation equation system 
> worse, as far as I understand.
>   

Yes, you typically want to merge with 1e-2 precision if you're using meters.

Merging contacts with differing normals is necessary to get smooth 
results from a ball rolling on a trimesh, because the normal of a sphere 
intersecting with a triangle edge goes away from the edge toward the 
sphere, not orthogonal to the triangle surface.


So, if you don't like it, the code's all yours :-)