dude about broadphase-solver process

gdlk
Posts: 62
Joined: Fri Oct 24, 2014 7:01 pm

dude about broadphase-solver process

Post by gdlk »

Hi!

A dude about the process since broadphase stage up to solver stage (thinking in a "default" bullet configuration).

Imagine the next case: you have a bridge make up by 3 boxes in compound. below the bridge there are some physics objects (the objects there arent touching the bridge, but they are in the same altitude that the base bridge boxes). So because the bridge is compound, their aabb touch the objects below it and wake up the objects.

Now the dude. Because the aabb are touching (or one inside the another), the broadphase is detecting this and it will generate the colliding pair. After that, there is another filter that detect the colliding pair are too far ( I suppose that, it has some name this stage? ), so is not necessary send the colliding pair to the solver. Then the solver don't need to do nothing (it never receive colliding pairs in this case). Is this right?

I am asking for this because I had implemented that case. However, in my case the processing time depends on the number of objects I had below the bridge (I would expect a little increase in the processing time because there are more active objects (so broadphase had more work to do), but they are not colliding (expect with the floor), however it see like the solver is processing the colliding pair anyways).

Thanks!!

Regards!