[gmx-users] Re: MPI

[Justin A. Lemkul]

simon sham wrote:
> Hi,
> Our system system administrator configured the software with --MPI 
> option according to the installation instructions. He told me there were 
> lots of warning messages when he did it. My related question is that 
> since there is LAM/MPI out there as well, can GROMACS 4.5 be configured 
> with OPENMPI?
>  

Yes, in fact OpenMPI has made LAM obsolete.

-Justin

> Thanks in advance,
>  
> Simon Sham 
> 
> 
>     Hi,
>     This is the same email that I'd sent yesterday but has not got posted?
>     Hi,
>     I have some questions about installation of GROMACS with MPI.
>     Our system administrator
>     has installed the 4.5.1 version in our system. When I tested the
>     software with "openmpi/openmpi-1.3.3-gcc/bin/mpirun" with mdrun_mpi, I
>     got all sorts of error message such as
>     "opt/apps/gromacs/gromacs-4.5.1-mpi/bin/mdrun_mpi: Symbol
>     `ompi_mpi_byte' has different size in shared object, consider
>     re-linking", and many others.
> 
>     My questions:
>     1. When I prepare my system for simulations, do I have to
>     consistently use *_mpi functions in gromacs_4.5.1_mpi directory?
>     2. Are there any details that I need to pay attention when I
>     configure the software with openmpi?
> 
>     Thanks for your help in advance.
> 
>     Best,
> 
>     Simon Sham
> 
> 
> 
> 
>          
>     -------------- next part --------------
>     An HTML attachment was scrubbed...
>     URL:
>     http://lists.gromacs.org/pipermail/gmx-users/attachments/20100924/2a671694/attachment-0001.html
> 
>     ------------------------------
> 
>     Message: 3
>     Date: Fri, 24 Sep 2010 15:18:27 -0400
>     From: "Justin A. Lemkul" <jalemkul at vt.edu
>     <http://us.mc367.mail.yahoo.com/mc/compose?to=jalemkul@vt.edu>>
>     Subject: Re: [gmx-users] MPI
>     To: Discussion list for GROMACS users <gmx-users at gromacs.org
>     <http://us.mc367.mail.yahoo.com/mc/compose?to=gmx-users@gromacs.org>>
>     Message-ID: <4C9CF983.7010006 at vt.edu
>     <http://us.mc367.mail.yahoo.com/mc/compose?to=4C9CF983.7010006@vt.edu>>
>     Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
> 
>     Guess my reply never hit the list, either, but it's in the archive:
> 
>     http://lists.gromacs.org/pipermail/gmx-users/2010-September/054259.html
> 
>     -Justin
> 
>     simon sham wrote:
>      > Hi,
>      > This is the same email that I'd sent yesterday but has not got
>     posted?
>      > Hi,
>      > I have some questions about installation of GROMACS with MPI.
>      > Our system administrator has installed the 4.5.1 version in our
>     system.
>      > When I tested the software with
>     "openmpi/openmpi-1.3.3-gcc/bin/mpirun"
>      > with mdrun_mpi, I got all sorts of error message such as
>      > "opt/apps/gromacs/gromacs-4.5.1-mpi/bin/mdrun_mpi: Symbol
>      > `ompi_mpi_byte' has different size in shared object, consider
>      > re-linking", and many others.
>      >
>      > My questions:
>      > 1. When I prepare my system for simulations, do I have to
>     consistently
>      > use *_mpi functions in gromacs_4.5.1_mpi directory?
>      > 2. Are there any details that I need to pay attention when I
>     configure
>      > the software with openmpi?
>      >
>      > Thanks for your help in advance.
>      >
>      > Best,
>      >
>      > Simon Sham
>      >
>      >
> 
>     -- 
>     ========================================
> 
>     Justin A. Lemkul
>     Ph.D. Candidate
>     ICTAS Doctoral Scholar
>     MILES-IGERT Trainee
>     Department of Biochemistry
>     Virginia Tech
>     Blacksburg, VA
>     jalemkul[at]vt.edu | (540) 231-9080
>     http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
> 
>     ========================================
> 
> 
> 

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================