Hi !
In the current NEB implementation in LAMMPS, only one NEB replica is allowed to run in a processor. I was wondering if this could be upgraded to allow multiple replicas per processor, which would then considerably increase the speed/efficiency of the NEB implementation.
Thank you very much in advance !
Sankar
See section 4.5 of the manual. In particular:
Also note that with MPI installed on a machine (e.g. your desktop),
you can run on more (virtual) processors than you have physical
processors. Thus the above commands could be run on a single-processor
(or few-processor) desktop so that you can run a multi-replica
simulation on more replicas than you have physical processors.
This means you can run 10-replica NEB on a single proc
if you like. But unlike what you say, it will not be faster or
more efficient than running on 10 procs. I think your logic
is mixed up in that respect.
Steve