Re: [AMBER] Parallel installation of MMPBSA.py.MPI failed

From: Jason Swails <jason.swails.gmail.com>
Date: Thu, 8 Dec 2016 11:35:42 -0500

On Thu, Dec 8, 2016 at 5:03 AM, MOHD HOMAIDUR RAHMAN <rahmanhpu.gmail.com>
wrote:

> Dear Amber Users
>
> I am trying to install Amber 16 and AmberTools 16 on my lab workstation. I
> followed similar protocol as mention on Jason Swails wikidot.
> After installation, MMPBSA.py.MPI is missing from bin directory. So I look
> in mpi4py installation log file and found this error.
>
>
>
> */home/dynamix/sware/mpiDIR/lib/libmpich.a: could not read symbols: Bad
> valuecollect2: error: ld returned 1 exit statuserror: command
> '/home/dynamix/sware/mpiDIR/bin/mpicc' failed with exit status 1*
>

​It looks like whatever MPI installation you are using is not built
correctly for mpi4py. In particular, mpi4py should never try to link to a
sta​tic library (i.e., a file ending with .a).

Since we don't know how this MPI was installed, it's hard to help much.

Note, if you use the "configure_mpich" or "configure_openmpi" scripts
included in AmberTools/src, those *will* build a compatible MPI that should
work for MMPBSA.py.MPI.

HTH,
Jason

-- 
Jason M. Swails
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Dec 08 2016 - 09:00:02 PST
Custom Search