[ODE] How many contacts for dCollide () ?

[Jon Watte (ODE)]

But you're not actually going to get 128 contacts back from dCollide() 
except in very rare circumstances.

Cheers,

			/ h+

Vianney Lecroart wrote:
> The real problem with 128 is that you have to setup the 128 contact
> with correct values for each collision. So if you have 10 collisions
> and you run the simulation every 1 ms, you have 128*10*1000  = 1 280
> 000 contact setup per second.
> 
> On 17/09/2007, Jon Watte (ODE) <hplus-ode at mindcontrol.org> wrote:
>> 128 is fine, if perhaps a little much. The only time you'll actually get
>> 128 contacts back is when you're doing trimesh->trimesh, or when you
>> have a large geom embedded into a trimesh with lots of small triangles.
>>
>> Cheers,
>>
>>           / h+