[gmx-users] PME parameter in GROMACS

Florian Dommert dommert at icp.uni-stuttgart.de
Sun Jan 30 15:11:07 CET 2011

Hash: SHA1

On 01/30/2011 01:18 PM, aldi asmadi wrote:
> Dear all,
> I would like to know if there is a way in GROMACS to find or specify the
> value of kappa (or Ewald coupling/splitting parameter) for the PME
> technique that we use in our simulation. In my case, I would like to
> perform simulations with a specified kappa value and a grid size of A x
> A x A for the PME parameter.  If I understand it correctly, the grid can
> be specified by the following in GROMACS: fourier_nx = A, fourier_ny =
> 6, and fourier_nz = A.  For the kappa, I could not find any information
> about it so far from the manual or our mailing list. 
> Many thanks.
> Best regards,
> Aldi 

Unfortunately, Gromacs does not allow to input the splitting parameter
directly, however just a small change in src/gmxlib/ewald_util.c is
necessary to interpret a negative ewald_rtol in the mdp file as the
splitting parameter. Therefore add after the declaration of the
variables in calc_ewaldcoeff (line 53):

if (dtol<0.0){
	return -dtol;

If you are furthermore interested in an estimate of the error introduced
by your SPME parameters, try the tool g_pme_error, which also tells you
in the output file the corresponding splitting parameter \beta for a
certain cut-off and ewald_rtol. Together with g_tune_pme I have realized
that a quite remarkable gain in performance can be achieved, if the
tools are used hand in hand.


- -- 
Florian Dommert

Institute for Computational Physics

University Stuttgart

Pfaffenwaldring 27
70569 Stuttgart

Phone: +49(0)711/685-6-3613
Fax:   +49-(0)711/685-6-3658

EMail: dommert at icp.uni-stuttgart.de
Home: http://www.icp.uni-stuttgart.de/~icp/Florian_Dommert
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/


More information about the gromacs.org_gmx-users mailing list