[gmx-users] question re; building Gromacs 4.6

[Roland Schulz]

On Tue, Jan 29, 2013 at 11:07 AM, Susan Chacko <susanc at helix.nih.gov> wrote:

> Thanks for the info! Our cluster is somewhat heterogenous, with
> some 32-core GigE-connected nodes, some older 8-core Infiniband-connected
> nodes, and some GPU nodes. So we need pretty much every variation
> of mdrun :-).
>

There is no disadvantage in having GPU or MPI compiled in. Including for
tools (the binaries other than mdrun). The tools don't use MPI or GPU, so
they work the same with or without it compiled in. As long as the cuda
librararies are available also on the nodes which don't have a GPU, the GPU
version works everywhere. Also if you use OpenMPI (not sure about MPICH)
and shared libraries you don't need different binaries for GigE and
Infiniband (only different version of OpenMPI). Only if your CPUs are
different (if only some support SSE4.1/AVX) you need different binaries. So
you might only need to compile Gromacs once.

Roland


> On Jan 29, 2013, at 11:00 AM, Mark Abraham wrote:
>
> > On Tue, Jan 29, 2013 at 4:39 PM, Susan Chacko <susanc at helix.nih.gov>
> wrote:
> >
> >>
> >> Sorry for a newbie question -- I've built several versions of Gromacs in
> >> the
> >> past but am not very familiar with the new cmake build system.
> >>
> >> In older versions, the procedure was:
> >> - build the single-threaded version
> >> - then build the MPI version of mdrun only. No need to build the other
> >> executables with MPI.
> >>
> >> Is this still how it should be done, or should one just build everything
> >> once with MPI?
> >>
> >
> > You can still follow this workflow if you need mdrun with real MPI to run
> > on your hardware (i.e. multiple physical nodes with network connections
> > between them).
> >
> >
> >> Likewise, if I want a separate GPU version (only a few nodes on our
> >> cluster have GPUs), do I build the whole tree separately with
> -DGMX_GPU=ON,
> >> or just a GPU-enabled version of mdrun?
> >>
> >
> > Only mdrun is GPU-aware, so that's all you'd need/want. I'll update the
> > installation instructions accordingly. Thanks!
> >
> > Mark
> > --
> > 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/Support/Mailing_Lists/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/Support/Mailing_Lists
>
> Susan Chacko
> Helix/Biowulf Staff
>
>
>
>
> --
> 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/Support/Mailing_Lists/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/Support/Mailing_Lists
>
>
>
>
>


-- 
ORNL/UT Center for Molecular Biophysics cmb.ornl.gov
865-241-1537, ORNL PO BOX 2008 MS6309