[ODE] Large physics worlds - let autodisable handle it, or create a sliding window of physical activity?

Mathias Weber mweb at gmx.ch
Sun Apr 24 12:36:03 MST 2005


Aras Pranckevicius wrote:
>>>Should I package the patch again, and maybe submit/upload it somewhere?
>>>
>>
>>I know I for one would love to try it out...  Are there any major
>>issues with it?


I would like to try it out too. So it would be nice if you could provied
a patch.

I did implement your idea with having all static objects in a quadtree
space and not using categroy/collide mask. And it works great. I don't
have the exact time that it takes within collision detection but when i
took a look at the cpu usage i get promissing results.

categroy/collide mask:
about 100 static objects:   8% cpu usage
about 1000 static objects   27% cpu usage

quadtree space for static objects:
about 100 static objects:   1.7% cpu usage
about 1000 static objects   1.7% cpu usage

I'm aware this values arn't very exact but they show that the first
version i used isn't usable. but i need to make some more tests with the
 second version but i have to change my network code first, so that i
can add more objects than 1000 static ones. And sure there will be more
to calculate if i got more dynamic objects but since never all objects
will be active this won't be a problem.

thanks for your help and i hope to see a patch

Mathias


More information about the ODE mailing list