yade-dev team mailing list archive

[boon chiaweng <booncw@xxxxxxxxxxx>]

Yes, I've just checked.  At the first timestep, all the stiffnesses (rotational and non-rotational) are zero. So, the if-else gives dt = Mathr::MAX_REAL;  Maybe setting defaultDt = -1 is a bad idea, and could let things go havoc at the start of the simulation.  

> Note you can assign GSTimeStepper::defaultDt, which will replace the 
> computed value when computed > default.
According to the code, it is that use min(computed,default).  In this case defaultDt is the maximum allowable timestep.

Thanks Bruno for the advice!

Cheers,

CWBoon

> Date: Mon, 11 Jan 2010 15:10:13 +0100
> From: bruno.chareyre@xxxxxxxxxxx
> To: yade-dev@xxxxxxxxxxxxxxxxxxx
> Subject: Re: [Yade-dev] [Bug 505523] Re: Velocity Bin
> 
> Sometimes, StiffnessTimestepper can give irrelevant value in the first 
> steps before conatcts are created (I don't know if it is the problem here).

> In order to identify the problem, you could check what is computed in 
> GSTimeStepper for each body (mass, stiffness, dt), using your prefered 
> debugging method.
> 
> Bruno
> 
> boon chiaweng a écrit :
> > Hi,
> >
> > I've enabled LOG_INFO.  The computed timestep is ridiculously large.  
> > Can somebody advise me how do i progressively track down the cause?
> >
> > 36697 [0x7f1eecb81950] INFO yade.GlobalStiffnessTimeStepper null - 
> > /home/chiab/yadeNew/Branch3/pkg/dem/Engine/GlobalEngine/GlobalStiffnessTimeStepper.cpp:170 
> > computeTimeStep: computed timestep 7.56436e+153, BUT timestep is 7.59375.
> >
> >
> > Yours,
> >
> > CWBoon
> >
> >
> > ------------------------------------------------------------------------
> > From: booncw@xxxxxxxxxxx
> > To: yade-dev@xxxxxxxxxxxxxxxxxxx
> > Date: Mon, 11 Jan 2010 19:52:20 +0800
> > Subject: Re: [Yade-dev] [Bug 505523] Re: Velocity Bin
> >
> > Thanks Bruno. (Sorry, I forgot to clicked Reply All the first time).   
> > I'm using a relatively new version right before 'bex' is changed to 
> > 'fex'.  I have looped through in python:
> >
> > for b in O.bodies:
> >     if b.shape.name == 'Sphere':
> >         if b.shape['radius']<0.0001:
> >              print b.shape[radius]
> >
> > Since nothing appears, all my radius are positive.  I've also checked 
> > the smallest radii range at the start of the simulation before 
> > expansion.  The radius of the smallest particles are between 0.005 and 
> > 0.006.  The weird thing is that the step size at the start of the 
> > simulation is as large as 200000.  I've included a screenshot
> >  
> >
> > Yours,
> >
> > Boon
> >
> > > Date: Mon, 11 Jan 2010 11:53:04 +0100
> > > From: bruno.chareyre@xxxxxxxxxxx
> > > To: yade-dev@xxxxxxxxxxxxxxxxxxx
> > > Subject: Re: [Yade-dev] [Bug 505523] Re: Velocity Bin
> > >
> > > Hi Boon,
> > >
> > > Are you sure your distribution is not giving negative radii or
> > > something? The timestep is really (should be) the dt obtained on the
> > > smallest particle. Or there is a bug in this old release.
> > >
> > > Bruno
> > >
> > > > Some particles are really small. But everytime i run a simulation, it
> > > > gives me a different timestep. And sometimes it is very large, because
> > > > I have very large particles in the simulation. But the calculation of
> > > > the timestep should be based on the smallest particle.
> > > >
> > >
> > >
> > > _______________________________________________
> > > Mailing list: https://launchpad.net/~yade-dev
> > > Post to : yade-dev@xxxxxxxxxxxxxxxxxxx
> > > Unsubscribe : https://launchpad.net/~yade-dev
> > > More help : https://help.launchpad.net/ListHelp
> >
> > ------------------------------------------------------------------------
> > New Windows 7: Find the right PC for you. Learn more. 
> > <http://windows.microsoft.com/shop>
> > ------------------------------------------------------------------------
> > New Windows 7: Find the right PC for you. Learn more. 
> > <http://windows.microsoft.com/shop>
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Mailing list: https://launchpad.net/~yade-dev
> > Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx
> > Unsubscribe : https://launchpad.net/~yade-dev
> > More help   : https://help.launchpad.net/ListHelp
> >   
> 
> 
> -- 
> _______________
> Bruno Chareyre
> Associate Professor
> Grenoble INP
> Lab. 3SR
> BP 53 - 38041, Grenoble cedex 9 - France
> Tél : 33 4 56 52 86 21
> Fax : 33 4 76 82 70 43
> ________________
> 
> 
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-dev
> Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~yade-dev
> More help   : https://help.launchpad.net/ListHelp
 		 	   		  
_________________________________________________________________
New Windows 7: Find the right PC for you. Learn more.
http://windows.microsoft.com/shop