[gmx-users] Heterogeneous GPU cluster question?

[Szilárd Páll]

Hi,

You can use multiple types of GPUs in a single run, but it won't be ideal.
Also, with Volta GPUs you'll probably be better off also offloading PME
which won't scale to more than 2-3 GPUs, so probably you'll not want to use
more than 2 GPUs in run with Volta.

--
Szilárd


On Tue, Aug 28, 2018 at 4:31 PM Smith, Micholas D. <smithmd at ornl.gov> wrote:

> Dear users (and developers),
>
> My sys. admin. is exploring purchasing some more GPU nodes for our cluster
> and was curious if we could get away with running GROMACS on a cluster that
> has two different types of GPUs, i.e. 70% of the nodes have 2 p100 gpus
> while 30% have volta (both have 2 gpu's per node).?
>
> I've been looking through the mailing list, but haven't bumped into this
> situation before.
>
> My inclination was to just purchase the "older" p100 since that is what
> all of the nodes are currently sporting, but thought I would ask to see if
> anyone has tried this kind of set-up before.
>
> Thanks,
>
> -Micholas
>
> ===================
> Micholas Dean Smith, PhD. MRSC
> Post-doctoral Research Associate
> University of Tennessee/Oak Ridge National Laboratory
> Center for Molecular Biophysics
>
> ________________________________________
> From: gromacs.org_gmx-users-bounces at maillist.sys.kth.se <
> gromacs.org_gmx-users-bounces at maillist.sys.kth.se> on behalf of Justin
> Lemkul <jalemkul at vt.edu>
> Sent: Tuesday, August 28, 2018 10:08 AM
> To: gmx-users at gromacs.org
> Subject: Re: [gmx-users] rerun doesn't calculate Coulomb & LJ interactions
> between groups
>
> On 8/28/18 10:04 AM, Irem Altan wrote:
> > Hi,
> >
> > I ran some umbrella sampling simulations on GPUs. I want to get the
> Coulomb and LJ energy between certain groups (chain A - chain B, chain A -
> solvent, etc.) , so I ran rerun:
> > gmx mdrun -rerun umbrella1.xtc -nb cpu -v -deffnm umbrella1
> > However, when I run gmx energy on the output .edr file:
> > gmx energy -f umbrella1.edr -o umbrella1_ener.xvg
> > I only get 50 terms listed, as opposed to ~90 that contains things like
> "LJ-SR:SOL-Chain_A". What am I missing?
> > The executable is created with the following command:
> > gmx grompp -f umbrella.mdp -c conf1.gro -p topol.top -n index.ndx
> -maxwarn 1 -o umbrella1.tpr
> > where chain A, chain B, and SOL are defined in index.ndx.
>
> The contents of index.ndx are irrelevant unless you've actually specific
> which energygrps you want to define (in the .mdp file). Also, why are
> you using -maxwarn 1? This should almost never be done.
>
> -Justin
>
> --
> ==================================================
>
> Justin A. Lemkul, Ph.D.
> Assistant Professor
> Virginia Tech Department of Biochemistry
>
> 303 Engel Hall
> 340 West Campus Dr.
> Blacksburg, VA 24061
>
> jalemkul at vt.edu | (540) 231-3129
> http://www.thelemkullab.com
>
> ==================================================
>
> --
> 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.
>