long range Coulomb interaction


Axel is right--the way that your textbook describes the long-range
Coulombic interaction and the way it is implemented in LAMMPS is the
same--just two different ways to look at the same method. Normally
the user sets the real-space cutoff and also specifies an accuracy
value. The code then predicts what the guassian charge width (named
g_ewald in the LAMMPS code) needs to be to achieve this accuracy in
real space, and also the reciprocal lattice vector cutoff needed to
achieve this accuracy in reciprocal space based (using the previously
determined g_ewald value). You can also specify the g_ewald value
using kspace_modify ( http://lammps.sandia.gov/doc/kspace_modify.html


Thank you very much for pointing out the “g_ewald” thing, I think I understand how it works now.