[ODE] geoms having the same position

Jorrit Tyberghein jorrit.tyberghein at gmail.com
Tue Oct 4 16:00:15 MST 2005


On 10/4/05, Aras Pranckevicius <nearaz at gmail.com> wrote:
> >  I've 1223 geoms having the same position (all are in one hash space,
> >  and NearCallback contents are commented so ther shouldn't be any
> >  collision checking for that geoms) and it all runs VERY SLOW (frame per minute or so...).
>
> In ODE hashspace implementation, the hash space is rebuilt every step
> (frame). So while in the end the NearCallback does nothing, it still
> goes through each geom, hashes it and inserts into internal hashspace
> structures, then goes through all nearby pairs (in your case it's over
> million pairs - because each geom potentially collides with each
> other) just to call the empty callback.
>
> There's not that much you can do about that. Quadtree space does not
> rebuild the tree each frame, but would still check these million
> pairs, so it should not magically help.
>
> So, the question is... why do you need a thousand geoms in one place?

I'm the original person with the problem. The 1223 geoms are not in
the same place. They just connect with each other. i.e. the 1223 geoms
represent walls and floors for a big maze like level.

This should be a very common situation. I don't understand how ODE
cannot cope with this.

Greetings,

--
Project Manager of Crystal Space (http://www.crystalspace3d.org)
and CEL (http://cel.crystalspace3d.org)
Support Crystal Space. Donate at
https://sourceforge.net/donate/index.php?group_id=649



More information about the ODE mailing list