I am from Computer Science background and I am new to MD simulations.
I have ran the ‘crack’ example. I wanted to know whether LAMMPS allows multiple simultaneous simulations for crack propagation. For example, if there are multiple cracks, then is it possible to simulate each crack simultaneously in greater detail (may be using different simulation parameters). If yes, then do these multiple simulations interact among themselves? I did not find anything in the documentation related to this, it will be great if you can point me to some documentation on this.
I am from Computer Science background and I am new to MD simulations.
good luck. you have a steep road ahead of you.
I have ran the 'crack' example. I wanted to know whether LAMMPS allows
multiple simultaneous simulations for crack propagation. For example,
if there are multiple cracks, then is it possible to simulate each
crack simultaneously in greater detail (may be using different
simulation parameters). If yes, then do these multiple simulations
interact among themselves? I did not find anything in the
documentation related to this, it will be great if you can point me to
some documentation on this.
you have to clarify this a little bit.
do you want to run multiple instances of LAMMPS at
the same time and have each instance do a simulation
with different parameters?
yes, that is in principle supported via the -partitions
option. there are a number of parallel replica methods
available. check out:
if you want to do a different parallel replica method,
you would have to implement it. if those individual
simulations, are not coupled in any way, then you
can just run independent calculations.
or do you want to study multiple cracks in the same system?
then you are actually studying a different system, since
the individual cracks will be influencing the residual
pulling force on the atoms of the other ones.
Hello Prof. Axel,
Thank you very much for your reply.
I had a quick question related to "studying multiple cracks in the same
system" - does LAMMPS currently have this option?
that is not really a question about what LAMMPS takes as an input
parameter, but rather how the system you are modeling would behave.
LAMMPS will take a system that you set up and run an MD
or minimization on it. how it behaves follows (usually deterministic)
from the interactions between the atoms. so whether a system
will show multiple cracks depends on the input not on LAMMPS.
of course you can induce multiple defects that would lead to
one or more cracks.
at the very abstract level, it is difficult to discuss details, though.
you'd have to tell us what your goal is and why you want to study
multiple cracks at the same time.