Re: Replica exchange on single physical node

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Sat Jul 06 2013 - 10:37:43 CDT

Following a google search, it seems to me that the quickest solution
to running remd on a single physical node, is to change to GROMACS
4.6, which, by adopting CUDA, can now run also Kepler GPUs, and where
MPI does only distinguish processes, not nodes.
fp

On Sat, Jul 6, 2013 at 4:19 PM, Francesco Pietra <chiendarret_at_gmail.com> wrote:
> Did you find a way to run REMD on a single physical node? I have the
> same problem (When I run the replica exchange example for alanin, it
> does a hard-check in the patched charm++ code to make sure the number
> of nodes is greater than the number of replicas and then stops (since
> 1node<8replicas), even if I specify multiple processes for mpirun)
>
> Thanks
> francesco pietra
>

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:23:26 CST