[gmx-users] mdrun_d not running in parallel

[Mark Abraham]

----- Original Message -----
From: "Justin A. Lemkul" <jalemkul at vt.edu>
Date: Monday, March 8, 2010 23:33
Subject: Re: [gmx-users] mdrun_d not running in parallel
To: Discussion list for GROMACS users <gmx-users at gromacs.org>

> 
> 
> sarbani chattopadhyay wrote:
> >Hi ,
> >I had installed gromacs4.0.7 in double preicision in 64 bit Mac 
> 10.6.1 computer with 8
> >dual core processors.
> >
> >1) I installed the fftw-3.0.1 in the following way
> >
> >( in the directory)
> >./configure --enable float
> >sudo make sudo make install
> >make distclean
> >./configure --enable-threads
> >make
> >make install
> >
> >2) I installed gromacs as
> >./configure --disable-float
> >sudo make
> >sudo make install
> >
> >3) I installed lam-7.1.4 as
> >./configure --prefix=/usr/local/lam --without-fc
> >sudo make
> >sudo make install
> >
> >After that I went back to the gromacs directory
> >and did
> >make clean
> >./configure --enable-mpi --program-suffix=_mpi
> >make mdrun
> >sudo make install -mdrun
> >
> >In this way I couldn't get mdrun_mpi_d. However I read from the 
> revision notes that grompp
> >and mdrun can run in parallel by themselves.( "-np" option is 
> not there in grompp_d)

./configure --enable-double --enable-mpi --program-suffix=_mpi_d

is required for double precision floating-point and MPI named how you would like.

Mark

> The use of -np is entirely unrelated to precision.  Why 
> couldn't you get the MPI-enabled mdrun?  I'm assuming there 
> were some errors.
> 
> If you're referring to the following statement: "Parallel 
> simulations can be started with grompp and mpirun -np mdrun 
> without further options," it only refers to the fact that grompp 
> no longer requires -np.  The only MPI-enabled program is 
> mdrun, and must be configured against MPI libraries; it doesn't 
> just magically work by itself.
> 
> >
> >
> >However when I give the commands
> >grompp_d -f em.mdp -c solvated.gro -o em.tpr -zero
> >lamboot
> >nohup mpirun -np 8 mdrun_d -v -s em.tpr -deffnm em &
> >
> >the md.log file says nnodes=1. the simulation is not running in 
> parallel.>
> 
> Clearly.  The mdrun_d executable is a result of your double-
> precision, non-MPI installation (step 2 above).  If you 
> want both MPI-enabled and double-precision, you need to --
> disable-float and --enable-mpi in the same configuration step.
> 
> -Justin
> 
> >Where have I gone wrong?
> >Any suggestion will be of great help.
> >Thanks in advance.
> >Sarbani
> >
> ><http://sigads.rediff.com/RealMedia/ads/click_nx.ads/www.rediffmail.com/signatureline.htm@Middle?>
> >
> 
> -- 
> ========================================
> 
> 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
> 
> ========================================
> -- 
> 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