[ODE] Ragdoll performance with OpenDE

[Mohsin Hasan]

Ok. My bad. What was I thinking :( 

Anyway, it was the physics. I set the iterations to 2. It seems ok and is
now very fast (still not as fast as Tokamak though but I am working on it).
I'll upload the demos online soon.

One more question. My simulations don't seem to be as stable as I want. I
mean when I created the box stacking demo, the boxes continually moved a
bit. It seemed like they were always shaking a bit. Why is that? Any ideas?
Most of the time the stack of boxes fell because of this shaking. Sometimes
the boxes just started sliding on the floor. I fixed this by reducing the
time step but this still happens if I leave the simulation running for a
long time.

Thanks,
Mohsin

-----Original Message-----
From: ode-bounces at q12.org [mailto:ode-bounces at q12.org] On Behalf Of Aras
Pranckevicius
Sent: Tuesday, January 13, 2004 12:34 PM
To: ode at q12.org
Subject: Re: [ODE] Ragdoll performance with OpenDE


> 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.

Two main bottlenecks could be: a) "physics", b) collision.

> dWorldStepFast1 method with only 8 iterations

Why that much? :) If the bottleneck is "physics", first try to reduce 
iterations (say, to 2) and see if that helps. Also, maybe you could go with
a 
larger timestep (and tune ERP/CFM parameters a lot to get good simulation).

If the bottleneck is collision, then make sure you don't collide ragdoll 
parts with itself; and use some space to keep the ragdolls (which one - 
depends on, umm, lots of factors :)).

In any case, the first thing would be to profile it and see what's
happening.


Aras Pranckevicius aka NeARAZ
http://www.gim.ktu.lt/nesnausk/nearaz/

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