[gmx-users] Important: Bugs in NEMD calculation

[Berk Hess]

> Date: Wed, 19 Jan 2011 20:52:13 +0100
> From: spoel at xray.bmc.uu.se
> To: gmx-users at gromacs.org
> Subject: Re: [gmx-users] Important: Bugs in NEMD calculation
> 
> On 2011-01-19 20.43, Berk Hess wrote:
> >
> >
> >  > Date: Wed, 19 Jan 2011 19:13:12 +0100
> >  > From: spoel at xray.bmc.uu.se
> >  > To: gmx-users at gromacs.org
> >  > Subject: Re: [gmx-users] Important: Bugs in NEMD calculation
> >  >
> >  > On 2011-01-19 18.36, Xiaohu Li wrote:
> >  > > Hi, All,
> >  > > I've found a bug in the NEMD calculation for viscosity. This has
> >  > > been reviewed in /*Hess's paper at JCP 116 209 2002.*/
> >  > > The version of gromacs I'm using is the development version.
> >  > > Notice that this version correct a previous but of 4.5.3, where you
> > uses
> >  > > NEMD, both the term V(eq. 21 on Hess's paper) and 1/eta(shear viscosity
> >  > > inverse) are supposed to be written to the *.edr file, however,
> >  > > the 4.5.3 versions does not have this. This version can be retrieved at
> >  > >
> > http://repo.or.cz/w/gromacs.git/commit/c83de86d65ce7135be6cef15e9100d7516e6d9a7
> >  > > *However, even this version is buggy since eta=A*rho/(V*k**2)(eq. 20
> >  > > Hess's paper)*. *I have performed simulation and has found out that the
> >  > > V and eta which are written in *edr file does not match according
> > to the
> >  > > formula, a little further check on the source code mdebin.c under the
> >  > > directory src/mdlib shows that it is actually calculating
> >  > > eta=A*Volume/(V*k**2) where density of rho should have been used. (This
> >  > > is at line 755 of mdebin.c ).
> >  > > I hope everyone who is using this can be aware of this, if you ever
> >  > > used this code to produce data, the V is correct from *edr,
> > however, you
> >  > > need to manuelly get your eta using the above formula.
> >  > > For the GMX developers, I hope anyone of you can correct this bug.
> >  > >
> >
> > I fixed this bug recently is the git release-4-5-patches branch.
> > You can get the fix from git and it will be in the 4.5.4 release
> > (no date yet).
> >
> >  > Thanks for pointing that out. There also is a small issue in that the
> >  > volume is computed for a rectangular box
> >  > vol = box[XX][XX]*box[YY][YY]*box[ZZ][ZZ];
> >  > dens = (tmass*AMU)/(vol*NANO*NANO*NANO);
> >  >
> >  > which would be incorrect for a non-rectangular box. You should however
> >  > use a rectangular box for this kind of calculations, although this is
> >  > not enforced by grompp.
> >
> > No.
> > That formula is correct for any triclinic box!
> 
> But the volume of a triclinic box is the determinant of the box (maybe 
> less than the product of the XX YY ZZ).

Gromacs only allows boxes with the upper (or lower depending on how you
order the elements) triangle zero. This means the determinant is equal
to the product of the diagonal elements. This product is (slighlty) more accurate
and computationally less expensive than calling the det function.

Berk

> 
> >
> > Berk
> >
> >  >
> >  > >
> >  > >
> >  > > Xiaohu
> >  > > *
> >  > >
> >  >
> >  >
> >  > --
> >  > David van der Spoel, Ph.D., Professor of Biology
> >  > Dept. of Cell & Molec. Biol., Uppsala University.
> >  > Box 596, 75124 Uppsala, Sweden. Phone: +46184714205.
> >  > spoel at xray.bmc.uu.se http://folding.bmc.uu.se
> >  > --
> >  > 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/Support/Mailing_Lists/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/Support/Mailing_Lists
> >
> 
> 
> -- 
> David van der Spoel, Ph.D., Professor of Biology
> Dept. of Cell & Molec. Biol., Uppsala University.
> Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
> spoel at xray.bmc.uu.se    http://folding.bmc.uu.se
> -- 
> 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/Support/Mailing_Lists/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/Support/Mailing_Lists
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110119/ef5132c6/attachment.html>