Hi Raimon,
What you are trying should work in principle - however, in practice communication via MPI-2 client/server model is in my experience quite finicky. I got it to work with certain MPICH versions in the past but for instance never OpenMPI.
I suggest to try getting your HREMD simulations to work with file based data exchange (set mpi = 0 in the &tc namelist). If this works and you are adventurous, we can try to get communication via MPI working.
All the best,
Andy
—
Dr. Andreas W. Goetz
Assistant Research Scientist
San Diego Supercomputer Center
Tel: +1-858-822-4771
Email: agoetz.sdsc.edu<mailto:agoetz.sdsc.edu>
Web: www.awgoetz.de<http://www.awgoetz.de>
On Apr 17, 2018, at 8:12 AM, Raimon Fabregat <raimon.fabregat.epfl.ch<mailto:raimon.fabregat.epfl.ch>> wrote:
Dear all,
I am trying to launch a simulation with hamiltonian replica exchange
where TeraChem (in gpu) should be in charge to do the QM part, but the
simulation crashes at the start with this error:
Running multisander version of sander Amber16
Total processors = 2
Number of groups = 2
Fatal error in MPI_Lookup_name: Invalid service name (see
MPI_Publish_name), error stack:
MPI_Lookup_name(161): MPI_Lookup_name(service="terachem_port.001",
MPI_INFO_NULL, port=0x17f91d00) failed
MPI_Lookup_name(136): fail failed
MPID_NS_Lookup(73)..: Lookup failed for service name terachem_port.001
First of all, do you know if it is possible to do what I want to do?
If so I will detail more the problem to see what I did wrong.
Thanks a lot!
Best,
Raimon
_______________________________________________
AMBER mailing list
AMBER.ambermd.org<mailto: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 Wed Apr 18 2018 - 14:00:05 PDT
