potfit wiki

open source force-matching

User Tools

Site Tools



Following is a complete list of all potfit parameters, together with a short description of their meaning, and indications when they are needed. Floating point numbers should be given in decimal notation. Vectors are given as a list of components, separated by spaces or tabs (all on one line).

An asterisk (*) marks required parameters.

parameter name parameter type default value
short explanation

main parameters

ntypes* integer (none)
number of atom types (types vary in [0,..,ntypes-1])

file names

config* string (none)
file with atomic configurations in force format, see Configurations
startpot* string (none)
file name containing the starting potential, see Potential files
tempfile* string (none)
file name for temporary (intermediate) potentials, see tempfile
endpot string $startpot_end
file name for final potential, see Potential files
imdpot string (none)
file name prefix for imd potentials, see IMD
plotfile string (none)
file name for plotting potential, see plotfile
flagfile string (none)
file, whose presence terminates the current fit, see Flagfile

general options

write_pair boolean 0
write radial pair distribution, see Pair distribution
write_lammps boolean 0
write potential in LAMMPS format, see LAMMPS
cell_scale float 1.0
rescale all cells, see cell_scale

Minimization options

opt boolean 0
whether or not to perform an optimization
eng_weight* float (none)
weight of energy in minimization, see Algorithms
stress_weight float (none)
(* if compiled with stress) weight of stress in minimiziation, see Algorithms
anneal_temp float 1.0
starting temperature for simulated annealing, set to 0 to skip simulated annealing, see Simulated Annealing
evo_threshold float 1.0
stop the evolutionary algorithm if the difference in the population is smaller than this, set to zero to skip evo, see Differential Evolution
d_eps float 0.0
error margin for powell minimization - if the improvement is smaller, the optimization is stopped, see lsq
seed integer 4
seed for random number generator
apot_punish float 0.0
scaling factor for apot punishments, this does not affect out-of-bounds punishment, see Analytic potentials/Punishments

options only for tabulated potentials

plotpointfile* string (none)
file name for plotting the sampling points, see plotpointfile
distfile string (none)
file name for distribution file (* with compile option 'dist'), see bindistfile
maxchfile string (none)
file with maximal permissible changes for each parameter, see Maxchfile

options only for analytic potentials

plotmin float 0.0
starting point for plotting potentials, see plotfile
enable_cp boolean 0
whether or not to include chemical potentials, see pair_potentials

options for output

imdpotsteps integer 500
number of sampling points in each function in imd potential files, see IMD
output_prefix string (none)
prefix for various output files, see Output
extend float 2.0
how many steps should the imd potentials be extrapolated, see IMD
lammpspotsteps integer 500
number of sampling points in each funcation in LAMMPS potential files, see LAMMPS

parameter for option coulomb

dp_cut float 10
cutoff-radius for electrostatic interactions, doesn't depend on other pair-function-radii.

parameters for option dipole

dp_tol float 1.e-7
dipole iteration precision.
dp_mix float 0.2
mixing parameter for dipole convergence during iteration.

parameters for OpenKIM potentials

The options for OpenKIM potentials are described in more detail on the OpenKIM page.

kim_model_name string (none)
KIM model indentifier
kim_model_params string (none)
Special parameter option, see OpenKIM page for details
kim_model_output_directory string (none)
Output directory for the optimized KIM model
kim_model_output_name string (none)
Name for the optimized KIM model
parameters.txt ยท Last modified: 2019/08/11 12:11 by daniel