[gmx-users] ewald_rtol parameter question

Berk Hess gmx3 at hotmail.com
Wed Feb 18 16:29:54 CET 2009

> Date: Thu, 19 Feb 2009 01:53:47 +1100
> From: Mark.Abraham at anu.edu.au
> To: gmx-users at gromacs.org
> Subject: Re: [gmx-users] ewald_rtol parameter question
> Claus Valka wrote:
> > Hello,
> > 
> > I would like to ask you a (simple) clarification question:
> > Reading the manual and equations 4.158 up to 4.160 there is a 
> > convergence parameter b (with hellenic letters β).
> > I would like to ask you whether this parameter is the same as the 
> > ewald_rtol that I come accross in the manual.
> No.
> > If it is not, is there any 
> > function between those two quantities?
> Yes. It's something like erfc(R)/R <= ewald_rtol for R = beta * 
> rcoulomb. That is, beta is controlled by the value of the two parameters 
> so that the contributions to the direct-space sum from charges outside 
> rcoulomb are negligibly small. Check the Darden papers referenced from 
> the GROMACS manual for details. GROMACS reports the value of 1/beta in 
> the .log file, IIRC.
> Mark

ewald_rtol sets the relative magnitude of the direct space eletrostatics part
at the cut-off (as described briefly in the mdp options manual pages), i.e.:
erfc(rcoulomb)/rcoulomb = ewald_rtol

We should add a few lines on this to the manual.


Express yourself instantly with MSN Messenger! Download today it's FREE!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20090218/c9cb6e9c/attachment.html>

More information about the gromacs.org_gmx-users mailing list