Hi,
I was wondering if someone could explain how exactly the angle lists in LAMMPS are maintained and updated, and in particular whether multiple processors can ever have the same angle (i.e. same set of three atoms, not just same angle type) listed on both their angle lists. I’m running into an issue where a fix I’m writing uses the angle list to modify the parameters of a specific set of atoms, and even though there’s only one angle in an otherwise diatomic fluid, when I check I’m finding that multiple processors claim to own those atoms on their angle list. I think this is probably a bug of some sort, but I’m having trouble getting rid of it, and even after reading up on some of the neighbor list information, I still feel fuzzy on the subject.
Thanks,
Thomas Allen
Hi,
I was wondering if someone could explain how exactly the angle lists in
LAMMPS are maintained and updated, and in particular whether multiple
processors can ever have the same angle (i.e. same set of three atoms, not
just same angle type) listed on both their angle lists. I'm running into an
yes, they can, when newton_bond is off. then all angle interactions
that straddle subdomains are computed in all of them and only the
parts that belong to the local atoms are use (with newton_bond on,
each angle is only computed twice and the forces on ghost atom
communicated back to their original local atoms.
issue where a fix I'm writing uses the angle list to modify the parameters
of a specific set of atoms, and even though there's only one angle in an
otherwise diatomic fluid, when I check I'm finding that multiple processors
claim to own those atoms on their angle list. I think this is probably a bug
of some sort, but I'm having trouble getting rid of it, and even after
reading up on some of the neighbor list information, I still feel fuzzy on
the subject.
as usual, there needs to be an input example to discuss this for real.
axel.