[gmx-users] PME problems

[Carsten Kutzner]

Hi Gloria,

there is nothing wrong with your parameters, the note just refers to
the direct:reciprocal workload being not optimal from a performance
point of view. With a small cutoff and a fine PME grid you will have
a high workload for the reciprocal space calculations - which is ok
and will yield correct results. However, it will not scale very well
to many cores. If you choose a larger value for rvdw, rlist and rcoulomb
and at the same time a larger value for fourierspacing, the scaling
will be better. You can manually multiply the cutoff radii and the
fourierspacing by a (small) factor and reach similar accuracy
while at the same time enabling better scaling.

You can also use the tool g_tune_pme for that, a version for
gmx 4.0.x is available here:

www.mpibpc.mpg.de/home/grubmueller/projects/MethodAdvancements/Gromacs/index.html

Carsten


On Aug 31, 2009, at 1:51 PM, 秦光荣 wrote:

> hello:
>        I am now using Gromacs4 to deal with a membrane protein and I  
> am confused on one NOTE.
>        When I use the parameters "rvdw = 0.9  ,  rlist = 0.9,   
> rcoulomb = 0.9, fourierspacing = 0.12" at the same time, a NOTE will  
> come out, which is "The optimal PME mesh load for parallel  
> simulation is below 0.5 and for highly parallel simulations between  
> 0.25 and 0.33, for higher performance, increase the cut-off and the  
> PME grid spacing".
>        If I change the fourierspacing to 0.3, this note will dismiss.
>        Also, if  I alter the other parameters to 1, the note will  
> dismiss too.
>        It's true that I want to do a parallel simulation, but I  
> don't which is the better way to solve this problem? Will the change  
> of these parameters cause something wrong? Or is there anyting wrong  
> with my .gro file?
>       Thanks for your attention.
>                                                                                                            Gloria
>
>
> _______________________________________________
> 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


--
Dr. Carsten Kutzner
Max Planck Institute for Biophysical Chemistry
Theoretical and Computational Biophysics
Am Fassberg 11, 37077 Goettingen, Germany
Tel. +49-551-2012313, Fax: +49-551-2012302
http://www.mpibpc.mpg.de/home/grubmueller/ihp/ckutzne