[ODE] Auto-Disabling?

Marc Rochel marc.rochel at udicom.de
Sat Apr 17 19:02:28 MST 2004


Hi !

I was wondering about 2 comments that were posted lately. They both advice
to use Auto-Disabling of bodies, escpecially the first one, the second one
only says, that using autodis makes his simulation act a lot better.
But looking in the current ODE-doc from cvs it says under 8.4 Experimental
Utilities included with StepFast1: "Note that the functions pertaining to
auto-disabling are not yet implemented!" This lets me think that autodis is
not implemented yet. What are the authors of these 2 posts talking about?
Looking in objects.h only these functions are said to be not implemented
yet:

/* These functions are not yet implemented by ODE. */
/*
dReal dBodyGetAutoDisableThresholdSF1(dBodyID);
void dBodySetAutoDisableStepsSF1(dBodyID, int AutoDisableSteps);
int dBodyGetAutoDisableStepsSF1(dBodyID);
void dBodySetAutoDisableSF1(dBodyID, int doAutoDisable);
int dBodyGetAutoDisableSF1(dBodyID);
*/

but these maybe working (???)

void dWorldSetAutoEnableDepthSF1(dWorldID, int autoEnableDepth);
int dWorldGetAutoEnableDepthSF1(dWorldID);
void dBodySetAutoDisableThresholdSF1(dBodyID, dReal autoDisableThreshold);

So what's the status of autodis? Is it implemented or not? Perhaps it is
implemented but it doesn't work correctly so we shouldn't use it? Perhaps
it's only implemented a little bit? Can someone clarify this to me?

ciao
       Marc


Here are the 2 posts:

----- Original Message ----- 
From: "Nguyen Binh" <ngbinh at glassegg.com>
To: "W.J. van der Laan" <wj at goliath.darktech.org>
Cc: <ODE at q12.org>
Sent: Wednesday, April 14, 2004 5:20 AM
Subject: Re: [ODE] Box stacking and ODE


>
>         Use Auto-Disable!
>
>         Try to search the list for more. I had produced a very
>         stable boxes stacking with StepFast1 before.
>
> WJvdL> I keep having problems stacking boxes in ODE, stacks of boxes just
> WJvdL> collapse after a while even if no force is applied to them, and the
> WJvdL> surfaces have infinite friction (so there should be no "shifting".
The
> WJvdL> number of contact points doesn't seem to matter either.
>
> WJvdL> Even the crash demo (from CVS) currently experiences this. The
boxes keep
> WJvdL> pushing each other, until the whole heap collapses.
>
> WJvdL> Can something be done against this?

----- Original Message ----- 
From: "Jani Laakso" <jani.laakso at itmill.com>
To: <ode at q12.org>
Sent: Wednesday, April 14, 2004 9:58 AM
Subject: [ODE] Getting more speed (auto enable + auto disable)



> PS. My simulation consists mostly of large stack of boxes. Auto
> disabling is something that makes my simulation act a lot better:
> -forcing extreme friction to certain bodies only
> -stable stacks
> -no jittering boxes on the ground
> -noticeable speed gains when you got lot's of boxes
>



More information about the ODE mailing list