[gmx-users] PME

[Mark Abraham]

On Mon, Sep 8, 2014 at 11:39 PM, Szilárd Páll <pall.szilard at gmail.com>
wrote:

> Hi,
>
> By default, there will be no separate PME ranks used with less than
> AFAIR 12 ranks (i.e. the default with small number of ranks is -npme
> 0). Without separate PME ranks (and without GPUs) there is no PP-PME
> load balance to tweak, so the PME load is not very relevant from
> performance optimization point of view.
>

Nitpick - even when the two phases are sequential (no separate PME ranks),
there is still load in both PP and PME parts, and the relative weight needs
to be managed for optimal performance. Shameless self-plug - see red lines
in Figure 1 of http://dx.doi.org/10.1002/jcc.21773

Mark


> Cheers,
> --
> Szilárd
>
>
> On Mon, Sep 8, 2014 at 9:06 PM, kiana moghaddam <ki_moghaddam at yahoo.com>
> wrote:
> > Dear GMX Users
> >
> > I have a question about PME loading When executing mdrun.
> > All my MD simulations (DNA-ligand interaction in triclinic box) are
> computed on in-house Linux 64-bit Intel Core-i7.
> > According to gromacs tutorial in Justin web site (
> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin/gmx-tutorials/lysozyme/08_MD.html),
> "For a cubic box, the optimal setup will have a PME load of 0.25 and for a
> dodecahedral box, the optimal PME load is 0.33".
> > Is this result should be obtained with my computer (with np=8)? or these
> PME load will be obtained only with np>8?
> >
> > Best Regards
> > --
> > Gromacs Users mailing list
> >
> > * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
> >
> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> >
> > * For (un)subscribe requests visit
> > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.
>