pyode-user

[Pyode-user] Threads and ODE?

[Chris S. <chr...@gm...>]

I'm unable to run a small ODE simulation for more than a couple
minutes before ODE crashes with the error:

ODE INTERNAL ERROR 2: invalid operation for locked space in dGeomMoved()

Is this caused by ODE's buggy memory management, or is it something
I'm doing? I'm manipulating ODE objects in a separate thread from the
one running the ODE simulation. Could that be causing the error? What
would be the best way to update a simulation from multiple threads
without having to pause the simulation?

Thanks,
Chris

Re: [Pyode-user] Threads and ODE?

[Chris B. <chr...@gm...>]

On 17/06/07, Chris Spencer <chr...@gm...> wrote:
> I'm unable to run a small ODE simulation for more than a couple
> minutes before ODE crashes with the error:
>
> ODE INTERNAL ERROR 2: invalid operation for locked space in dGeomMoved()
>
> Is this caused by ODE's buggy memory management, or is it something
> I'm doing? I'm manipulating ODE objects in a separate thread from the
> one running the ODE simulation. Could that be causing the error? What
> would be the best way to update a simulation from multiple threads
> without having to pause the simulation?

You probably need to serialise all ODE operations in a single thread.
Send messages to that thread to update the ODE objects, and call
World.step() from that thread when your message queue is empty.
Without examining the source in details it's impossible to tell the
effect of changing objects and using objects in parallel.

Re: [Pyode-user] Threads and ODE?

[Chris S. <chr...@gm...>]

Yep, that's exactly what I had to do to fix the problem. Thanks.

Chris

On 6/18/07, Chris Bainbridge <chr...@gm...> wrote:
> On 17/06/07, Chris Spencer <chr...@gm...> wrote:
> > I'm unable to run a small ODE simulation for more than a couple
> > minutes before ODE crashes with the error:
> >
> > ODE INTERNAL ERROR 2: invalid operation for locked space in dGeomMoved()
> >
> > Is this caused by ODE's buggy memory management, or is it something
> > I'm doing? I'm manipulating ODE objects in a separate thread from the
> > one running the ODE simulation. Could that be causing the error? What
> > would be the best way to update a simulation from multiple threads
> > without having to pause the simulation?
>
> You probably need to serialise all ODE operations in a single thread.
> Send messages to that thread to update the ODE objects, and call
> World.step() from that thread when your message queue is empty.
> Without examining the source in details it's impossible to tell the
> effect of changing objects and using objects in parallel.
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Pyode-user mailing list
> Pyo...@li...
> https://lists.sourceforge.net/lists/listinfo/pyode-user
>

Re: [Pyode-user] [ODE] Threads and ODE?

[Jon W. (ODE) <hpl...@mi...>]

ODE is not thread safe. You can only manipulate objects when you know 
that the world collision/simulation isn't running.
ODE does not have any known bugs in memory management.
If you need to manipulate objects, then you should double-buffer the 
state, and pick up any requested changes before you call collide/step 
the next time in the simulation thread.

Cheers,

          / h+


Chris Spencer wrote:
> I'm unable to run a small ODE simulation for more than a couple
> minutes before ODE crashes with the error:
>
> ODE INTERNAL ERROR 2: invalid operation for locked space in dGeomMoved()
>
> Is this caused by ODE's buggy memory management, or is it something
> I'm doing? I'm manipulating ODE objects in a separate thread from the
> one running the ODE simulation. Could that be causing the error? What
> would be the best way to update a simulation from multiple threads
> without having to pause the simulation?
>
>   

Re: [Pyode-user] [ODE] Threads and ODE?

[Chris S. <chr...@gm...>]

On 6/18/07, Jon Watte (ODE) <hpl...@mi...> wrote:
> ODE is not thread safe. You can only manipulate objects when you know
> that the world collision/simulation isn't running.
> ODE does not have any known bugs in memory management.

The bug I'm referring to is the one reported at
https://sourceforge.net/tracker/?func=detail&atid=382799&aid=1566350&group_id=24884

Even for thread-safe code, ODE tends to crash quite regularly when I
try to add or remove large numbers of bodies/geoms from a simulation.

Regards,
Chris

Re: [Pyode-user] [ODE] Threads and ODE?

[Chris B. <chr...@gm...>]

On 18/06/07, Chris Spencer <chr...@gm...> wrote:
> On 6/18/07, Jon Watte (ODE) <hpl...@mi...> wrote:
> > ODE is not thread safe. You can only manipulate objects when you know
> > that the world collision/simulation isn't running.
> > ODE does not have any known bugs in memory management.
>
> The bug I'm referring to is the one reported at
> https://sourceforge.net/tracker/?func=detail&atid=382799&aid=1566350&group_id=24884
>
> Even for thread-safe code, ODE tends to crash quite regularly when I
> try to add or remove large numbers of bodies/geoms from a simulation.

One common problem with large number of bodies is running out of stack
space. You can increase the stack space available to your program by
running 'ulimit -s unlimited' from your shell, before you start your
ODE program. But if you run python under gdb or similar you'll be able
to see exactly which function is crashing (the stack problem is
dSolveLCP).