[ODE] ODE Physics Benchmark Setup

[Jon Watte (ODE)]

Adrian Boeing wrote:
> a) The best accuracy (in terms of following the ideal physical models)
>   
Double precision, dWorldStep(), make sure to filter/merge contact points 
that are coincident, make sure to add some slight dampening (especially 
rotational dampening).

> b) The best performance (in terms of computational efficiency).
>   

Double precision, dWorldQuickStep(), make sure to filter/merge contact 
points that are coincident, make sure to add a noticeable amount of 
dampening (especially rotational dampening).

In both cases you probably want to use dContactApprox for friction, as 
the non-approx function is too simplistic to be of much use.

Cheers,

          / h+