[ODE] Ragdoll performance with OpenDE

[Henrik Karlsson]

Is there anyone on this list who has written joints that suits for
simulating dead bodies?.. I guess you can combine engine- and
ball-/hinge-joints, but wouldn't it be better to have one joint that is
optimized for this purpose? (When I think about it the engine joint doesn't
constraint same dof's as the ball joint at all? so the performance
difference would be neglectible? At least for the standard dWorldStep
version)


Best regards,
Henrik




-----Original Message-----
From: Mohsin Hasan [mailto:mohsin.hasan at trivor.com]
Sent: den 13 januari 2004 08:11
To: ode at q12.org
Subject: [ODE] Ragdoll performance with OpenDE


Hi,

I'm new to OpenDE. I've done some work with Tokamak but I really like the
feature set offered by ODE so I thought of switching to ODE. I tried to
implement a Ragdoll demo with ODE with 10 Bones (9 joints). The demo seems
to work fine. The only issue I have is performance. I mean I create 5
Ragdolls and the performance of the simulation starts to stall. I haven't
done anything fancy yet, just added simple hinge joints. ODE is built in
single precision mode, MAX_CONTACT points for collision is only 3 and I am
using dWorldStepFast1 method with only 8 iterations to get extra bit of
speed but nothing seems to help. On the other hand my Tokamak demo can
easily manage 40 Ragdolls (same configuration) without any major hiccups. 

My question is. Is ODE not suitable to implement Ragdoll or commercial
quality game physics yet due to speed constraints? Is there anything being
done to increase the performance? I know there are a few experimental
functions to auto disable bodies like in Tokamak but this will only help
once a body is in a resting state. If most of the objects (Ragdolls) are
moving (falling) speed is just unbearably slow.

Regards,
Mohsin

_______________________________________________
ODE mailing list
ODE at q12.org
http://q12.org/mailman/listinfo/ode