[AMBER] Problem running pmemd.cuda.mpi REMD with different number of replicas

From: Thiago Lipinski Paes <thiagopaes.gmail.com>
Date: Wed, 10 Sep 2014 13:21:40 -0300

Hello amber users and developers,

I'm trying to use AMBER14 and pmemd.cuda.MPI to perform T-REMD simulations.

My workstation description:
Intel® Core™ i7-3970X CPU . 3.50GHz × 12 (hexacore)
GPUS: 2x GeForce GTX 680

I'm using export CUDA_VISIBLE_DEVICES="0,1" and following the AMBER A7
Tutorial (http://ambermd.org/tutorials/advanced/tutorial7/)

I'm having issues when trying to use more than 12 replicas.

"mpirun -np 12 $AMBERHOME/bin/pmemd.cuda.MPI -ng 12 -groupfile
remd.groupfile" runs smoothly, but if I try to use more replicas like "-np
13" and "-ng 13", I have no luck.

Am I doing something really wrong?

Thanks in advance.


Thiago Lipinski Paes
Computer Engineer - Computer Science MSc - Computer Science PhD Candidate
Laboratório de Bioinformática, Modelagem e Simulação de Biossistemas (LABIO)
Pontifícia Universidade Católica do Rio Grande do Sul (PUCRS)
Av. Ipiranga, 6681 - Partenon - Porto Alegre/RS - CEP: 90619-900
Tel: +55 51 33203611 ext.8602
AMBER mailing list
Received on Wed Sep 10 2014 - 09:30:02 PDT
Custom Search