Thanks Jason,
That is very helpful information. For now, I just needed sander.MPI, so
commenting out the calls did the job!
Thanks,
Marc
On 11 April 2013 15:15, Jason Swails <jason.swails.gmail.com> wrote:
> On Thu, Apr 11, 2013 at 6:19 AM, Marc van der Kamp <
> marcvanderkamp.gmail.com
> > wrote:
>
> > Dear all,
> >
> > I tried to compile AMBER12 in parallel, by first configuring:
> > ./configure -mpi gnu
> >
> > And then compiling with 'make install'.
> > After MMPBSA.py.MPI is compiled successfully, I get the following errors
> > (to do with coupling to Terachem as external QM code it seems):
> >
> > qm2_extern_tc_module.o: In function
> > `__qm2_extern_tc_module_MOD_connect_to_terachem':
> > qm2_extern_tc_module.F90:(.text+0x4a53): undefined reference to
> > `mpi_lookup_name_'
> > qm2_extern_tc_module.F90:(.text+0x4dba): undefined reference to
> > `mpi_comm_connect_'
> > collect2: ld returned 1 exit status
> >
>
> This indicates that the MPI you're using does not support the MPI-2
> standard used by the QM/MM terachem interface. You can just look for
> "MPI_COMM_CONNECT" and "MPI_COMM_LOOKUP" and comment-out those two calls
> (this will of course kill the terachem-amber MPI interface).
>
> Another way, of course, is to simply use an MPI that does support MPI-2
> (this is probably the preferred way, especially if you plan on using
> pmemd.cuda.MPI, which also requires MPI-2 features).
>
> HTH,
> Jason
>
> --
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Candidate
> 352-392-4032
> _______________________________________________
> 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 11 2013 - 14:00:03 PDT
