ERROR on proc 15: Atom count changed in fix neb

LAMMPS LAMMPS Mailing List Mirror
1

Dear LAMMPS users,

I was wondering if someone could tell me what situations could cause the error “Atom count changed in fix neb”. I did not get this error when I ran 10 replicas but increasing to 20 and 40 replicas I am getting this error with the proc number changing.

Thanks,
Mihir.

2

Dear LAMMPS users,

I was wondering if someone could tell me what situations could cause
the error "Atom count changed in fix neb". I did not get this error
when I ran 10 replicas but increasing to 20 and 40 replicas I am
getting this error with the proc number changing.

i am guessing here, but this sounds a lot like you have
the NEB version of the infamous "lost atoms" condition.

does it happen right away, or after a while?
have you tried increasing the neighborlist skin?
do you use shrinkwrap boundary conditions somewhere?

cheers,
   axel.

3

More precisely, it means one (or more) of your replicas
lost an atom during its simulation. NEB cannot continue
in that mode, so it checks and stops. You can probably
look at the log file for the offending replica and see something
bad happening, either a LOST message or bad thermodyanamics.

Fix the problem and NEB should continue.

Steve