[gmx-developers] The development of a new implementation of Rahman Parrinello

[Michael Shirts]

Hi, all-

Sorry I haven't responded sooner, I've been on the road.

> I have heard that a new implementation of the Rahman Parrinello code
> is under development. All I know is that it uses a verlet velocity
> integrator.

Yes, I've been working with a lot of this locally.  I have velocity
verlet implemented, along with Andersen version of pressure control -
which takes care of the isotropic case.  It has been extended to
non-isotropic cases --see  the references in the the Moldy manual,
which has a good basic outline at
http://www.earth.ox.ac.uk/~keithr/moldy-manual/node14.html. I have to
fiddle around with the different implementations of anisotropic
pressure control to figure out which one is going to work the best. 
If you have any suggestions.  I have not yet committed changes to the
main CVS yet.

> Is this implementation based in any reference in the bibliography?

Again, check out the references listed here:
http://www.earth.ox.ac.uk/~keithr/moldy-manual/node14.html

> Would also include a reversible move for the Nose Hoover thermostat?

This is definitely in the plan, but I haven't quite got around to this
yet.  If you have any particular implementations you think are
particularly good, let me know.   Because of some committments here,
it may take a little longer to get these changes in, but I estimate
that they will almost certainly be in by the end of the year.

Going over this a little more, I think the best idea is to essentially
have two tracks -- one that leads to the current default path
(Berensen temperature and temperature control), and a separate track
that has velocity verlet, shake + rattle, exact pressure control
(which requires velocity verlet to be rigorously correct, and so forth
-- I don't think it's going to be necessary to mix and match all the
options, necessarily.  Please let me know if anyone has any other
thoughts on this -- I'm soliciting ideas to make sure I don't miss
anything important.

Best,
Michael



>
> Message: 2
> Date: Fri, 07 Oct 2005 08:34:24 +0200
> From: Berk Hess <hessb at mpip-mainz.mpg.de>
> Subject: Re: [gmx-developers] The development of a new implementation
>         of Rahman       Parrinello.
> To: Discussion list for GROMACS development
>         <gmx-developers at gromacs.org>
> Message-ID: <434616F0.30704 at mpip-mainz.mpg.de>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> ramon at juguete.quim.ucm.es wrote:
>
> >Hello,
> >
> >I have heard that a new implementation of the Rahman Parrinello code
> >is under development. All I know is that it uses a verlet velocity
> >integrator.
> >
> >Is this implementation based in any reference in the bibliography?
> >
> >
> >
> This will come, the exact algorithm has not been decided yet.
>
> >Would also include a reversible move for the Nose Hoover thermostat?
> >
> >
> Gromacs 3.3 contains a reversible Nose Hoover leap-frog integrator.
>
> Berk.
>
>
>
> ------------------------------
>
> _______________________________________________
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-developers
>
>
> End of gmx-developers Digest, Vol 18, Issue 3
> *********************************************
>