Re: [AMBER] failed to compile Amber11 with mpi on Mac

From: David Watson <dewatson.olemiss.edu>
Date: Thu, 29 Apr 2010 09:54:08 -0500

In my experience, mpich2 performs well for Amber on Mac.
I am not sure about Amber 11 performance, but for my needs (REMD) mpich2 allowed me to start a job which needed "4" CPUs on a Core Duo or Core 2 Duo without complaining that there weren't enough CPUs.
2 threads were started on each CPU, but the total amount of CPU usage was as close to 100% as the system would allow without a renice.

In one of my previous posts to this list, I describe how to go about building mpich2 with Intel fortran for the mac, as you will probably have to build your own implementation of openmpi that includes fortran wrappers anyway.
Just make sure that you change the prefix in the configuration step to something like "/opt/mpich2" in order to keep the system-provided openmpi from getting convoluted with the wrappers and libraries of mpich2.

On Apr 29, 2010, at 9:41 AM, Alan wrote:

> Hi there,
>
> It's not clear on the manuals which version of openmpi is required. I
> have openmpi-1.3.3 from Fink on my Mac, and I managed to compile pjtraj and
> nab with mpi.
>
> Then I try to install amber11. Serial was smooth, test (although long) went
> all fine. So parallel...
>
> Using the same successful config.h from AmbertTools with mpi, my attempt
> fails with:
>
> [snip]
> mpif90 -c -O3 -mtune=generic -ffree-form -o ips.o _ips.f
> cpp -traditional -P -DBINTRAJ -DMPI sander.f > _sander.f
> mpif90 -c -O0 -ffree-form -o sander.o _sander.f
> _sander.f:95.75:
>
> use qmmm_module, only : qmmm_nml,qmmm_struct, deallocate_qmmm, qmmm_mpi,
> &
> 1
> Error: Symbol 'qmmm_mpi_setup' referenced at (1) not found in module
> 'qmmm_module'
> make[1]: *** [sander.o] Error 1
> make: *** [parallel] Error 2
>
> I see that in configure_openmpi is mentioned openmpi 1.4, but should be
> 1.3.3 OK?
>
> Any ideas about the error above?
>
> Many thanks,
>
> Alan
> --
> Alan Wilter S. da Silva, D.Sc. - CCPN Research Associate
> Department of Biochemistry, University of Cambridge.
> 80 Tennis Court Road, Cambridge CB2 1GA, UK.
>>> http://www.bio.cam.ac.uk/~awd28<<
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Apr 29 2010 - 08:00:08 PDT
Custom Search