[gmx-users] direction_periodic

[Berk Hess]

Correct.
Make sure than your section A spreads over 90% (or 95%) of the x-range.
Then you are sure that the distance in x between the pbc-atom and every
other atom in A is not more than 0.45*boxlength plus a bit of fluctation.

Berk

> Date: Fri, 2 Oct 2009 11:49:37 +0200
> From: alexander.herz at mytum.de
> To: gmx-users at gromacs.org
> Subject: Re: [gmx-users] direction_periodic
> 
> Hi,
> 
> thx for suggesting the workaround, just to make sure I understand the
> workaround correctly:
> 
> I devide each slab in 2 sections, one section contains 90% of the atoms
> (section A), the other the rest (section B).
> Now I set the pbc_atom to the center atom of section A and pull it using
> constraints?
> 
> Alex
> 
> Berk Hess schrieb:
> > Hi,
> >
> > This is due to the periodicity issue of the COM calculation that I
> > have mentioned before.
> > The problem is that some atoms are at close to half a box length from
> > the pbc_atom
> > and those can jump back and forth between two periodic images in the
> > distance calculations.
> > This "noise" will make the constraint code crash, since it tries to
> > correct for it immediately.
> > The umbrella potential smooths this effect out somewhat and can
> > therefore run stably,
> > but you would still see the noise in the results.
> >
> > Fixing this is difficult, since it would require storing the periodic
> > image of each atom
> > in the pull group at the start of the simulation and determining the
> > image for the COM
> > calculation using the initial image. Also the checkpoint file should
> > store the images.
> > I would not like to put such code in the standard version.
> >
> > A workaround could be to pull on 90% of the slabs, leaving a part of
> > 10% along x out.
> > Then define the pbc_atom in the middle of the 90%. The 10% will just
> > move along.
> > This should solve your problem and should not produce artifacts.
> >
> > Berk
> >
> > > Date: Fri, 2 Oct 2009 10:25:21 +0200
> > > From: alexander.herz at mytum.de
> > > To: gmx-users at gromacs.org
> > > Subject: Re: [gmx-users] direction_periodic
> > >
> > > You did suggest that, but it crashes at step 0, as said :)
> > >
> > > Alex
> > >
> > > Berk Hess schrieb:
> > > > Hi,
> > > >
> > > > Then you should use constraint pulling instead of an umbrella
> > potential.
> > > > I have been wondering all the time why you are using a potential
> > > > and not a constraint. I think I had suggested using a constraint some
> > > > time ago, but I am not sure that I did.
> > > >
> > > > Berk
> > > >
> > > > > Date: Fri, 2 Oct 2009 09:45:42 +0200
> > > > > From: alexander.herz at mytum.de
> > > > > To: gmx-users at gromacs.org
> > > > > Subject: Re: [gmx-users] direction_periodic
> > > > >
> > > > > Hey,
> > > > >
> > > > > I want to pull the diamonds with more or less exactly the speed
> > I give
> > > > > for the pulling without the massive noise
> > > > > (in gromacs 3.3 this worked using afm type pulling, after some
> > > > > equilibration the slabs settled to an almost constant
> > > > > pull speed without soo much fluctuation using the same
> > parameters I use
> > > > > noew for umbrella).
> > > > > So I tried using a constraint rather than umbrella, which alwas
> > crashes
> > > > > at the first step (even if I did some equilibration
> > > > > using umbrella pulling beforehands).
> > > > >
> > > > > So how can I achieve a pull speed which is about constant 0.01
> > nm/ps and
> > > > > with fluctuations less than 10% of this speed?
> > > > >
> > > > > Thx,
> > > > > Alex
> > > > > _______________________________________________
> > > > > gmx-users mailing list gmx-users at gromacs.org
> > > > > http://lists.gromacs.org/mailman/listinfo/gmx-users
> > > > > Please search the archive at http://www.gromacs.org/search before
> > > > posting!
> > > > > Please don't post (un)subscribe requests to the list. Use the
> > > > > www interface or send it to gmx-users-request at gromacs.org.
> > > > > Can't post? Read http://www.gromacs.org/mailing_lists/users.php
> > > >
> > > >
> > ------------------------------------------------------------------------
> > > > See all the ways you can stay connected to friends and family
> > > > <http://www.microsoft.com/windows/windowslive/default.aspx>
> > > >
> > ------------------------------------------------------------------------
> > > >
> > > > _______________________________________________
> > > > gmx-users mailing list gmx-users at gromacs.org
> > > > http://lists.gromacs.org/mailman/listinfo/gmx-users
> > > > Please search the archive at http://www.gromacs.org/search before
> > posting!
> > > > Please don't post (un)subscribe requests to the list. Use the
> > > > www interface or send it to gmx-users-request at gromacs.org.
> > > > Can't post? Read http://www.gromacs.org/mailing_lists/users.php
> > >
> > > _______________________________________________
> > > gmx-users mailing list gmx-users at gromacs.org
> > > http://lists.gromacs.org/mailman/listinfo/gmx-users
> > > Please search the archive at http://www.gromacs.org/search before
> > posting!
> > > Please don't post (un)subscribe requests to the list. Use the
> > > www interface or send it to gmx-users-request at gromacs.org.
> > > Can't post? Read http://www.gromacs.org/mailing_lists/users.php
> >
> > ------------------------------------------------------------------------
> > What can you do with the new Windows Live? Find out
> > <http://www.microsoft.com/windows/windowslive/default.aspx>
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > gmx-users mailing list    gmx-users at gromacs.org
> > http://lists.gromacs.org/mailman/listinfo/gmx-users
> > Please search the archive at http://www.gromacs.org/search before posting!
> > Please don't post (un)subscribe requests to the list. Use the 
> > www interface or send it to gmx-users-request at gromacs.org.
> > Can't post? Read http://www.gromacs.org/mailing_lists/users.php
> 
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
 		 	   		  
_________________________________________________________________
See all the ways you can stay connected to friends and family
http://www.microsoft.com/windows/windowslive/default.aspx
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20091002/012af88e/attachment.html>