[lammps-users] virial stress in meam

LAMMPS LAMMPS Mailing List Mirror
1

All,

I’m having some trouble tracking down the form of the virial stress used with the MEAM potential. My initial thought was Fi dot Ri (and not Fij dot rij), because of its many-body nature. But a closer look at meam_force.F in the lib package showed that Fij is computed inside the loop of j atoms (I think). I’ve been trying to follow the vflag_* breadcrumbs through the code, but it’s pretty difficult. Can anyone provide a quick explanation?

Thanks,
Michael Sellers

2

What is effectively done for MEAM (and other many-body
potentials) is as follows. For each small group of
interacting atoms (i.e. in the loop over interactions),
compute sum (Fi dot Ri), which insuring that the Ri of all atoms
in the small group is "local", meaning the Ri are close to
each other and do not cross a periodic boundary.

Steve