Hello Users,
I’m running simulated annealing on a coarse grain polymer and I’m having trouble with the simulation terminating with either bonds, angles or dihedrals missing on processors. Based on the information given in the error messages in the documentation:
“Typically this is because the pairwise cutoff is set too short or the bond has blown apart and an atom is too far away.”
I’ve set the pairwise cutoff to:
pair_style lj/cut/opt 17.0
and the largest sigma is 6.0 Angstroms. Additionally, I am using these settings:
neighbor 3.0 bin
neigh_modify delay 5 check yes
I find this to be a little odd, even with a timestep of 1 fs and masses being kind of large (smallest is 35 g/mol).
Any comments or troubleshooting would be appreciated.
My complete force field parameter set:
Hello Users,
I'm running simulated annealing on a coarse grain polymer and I'm having
trouble with the simulation terminating with either bonds, angles or
dihedrals missing on processors. Based on the information given in the
error messages in the documentation:
"Typically this is because the pairwise cutoff is set too short or
the bond has blown apart and an atom is too far away."
there is not one reason for this, but many possibilities.
this is just one. if you are starting fresh from a non-equilibrated
configuration, then you may also have very high local
potential energy.
I've set the pairwise cutoff to:
pair\_style lj/cut/opt 17\.0
and the largest sigma is 6.0 Angstroms. Additionally, I am using these
settings:
neighbor 3\.0 bin
neigh\_modify delay 5 check yes
at the beginning of a calculations you should set delay to 0
and every to 1 this reduces the chance of losing atoms.
I find this to be a little odd, even with a timestep of 1 fs and masses
being kind of large (smallest is 35 g/mol).
Any comments or troubleshooting would be appreciated.
first thing to try is to do a "run 0" and then look at the
individual potential energies. if some are very high,
then you should start with some minimization steps
before attempting MD.
another option to avoid particles flying away
is to use nve/limit as integrator.
it is difficult to make more precise suggestions
without being able to run the input.
lost atoms/bonds/whatever/...
are just too generic a diagnostic.
axel.
You may try neighbor multi instead of 'bin' as you have different
cutoffs. Also, you should check the neigh_modify 'one' and 'page'
numbers, apparently you are using the default setting. Also delay and
every numbers should be checked. These settings are reasons lammps
crashes without spitting out errors, however, you may check these out
just to make sure you are doing it right before doing something else.
If it's a huge system (large number of monomers) neighbor and
neigh_modify is important.
Monojoy