A few new related features have been added to LAMMPS in
the last couple days, that may be of interest to some users.
Some recent Qs to the mail list have touched on issues that
are addressed by the new features.
a) new fix property/atom command that lets molecule IDs be
added to atom styles that don’t have them - also allows
new custom integer or floating-point vectors to be added
to store per-atom values
b) these values can be initialized by the read_data command
using its fix option to pass a chunk of the data flie to a fix
c) they can also be initialized by the set command, which now
allows atom-style variables to be used for nearly all of its
inputs - this means you can define a formula which is used
to set the values for many per-atom properties
d) a new atomfile-style variable was added to the variable command,
it allows one or more sets of per-atom values to be read from
a file and assigned to the variable - any LAMMPS command
that allows use of an atom-style variable (e.g. fix setforce)
can also use an atomfile-style variable, including the set command
as mentioned in ©
e) the compute property/atom command can access the custom
integer and floating-point attributes of (a), which means they
can become input to other commands (compute reduce, fix ave/spatial, etc)
as well as be output via dump custom
f) if you write a new class (pair, fix, compute) to add to LAMMPS,
it can access the custom integer/floating-point vectors created
by (a) - this is effectively a simple way to add custom per-atom
attributes to a model w/out having to write a new atom style
Steve