converge_kpoints - repeat VASP jobs until kpoint convergence
usage: converge_kpoints [-h] [-c [COMMAND]] [-i
[{linear,inc}]]
- [-m [MAX_STEPS]] [-t [TARGET]]
converge_kpoints perform a KPOINTS convergence. What this script
will do is to run a particular VASP run with increasing multiples of the
initial KPOINT grid until a target convergence in energy per atom is
reached. For example, let's say you have vasp input files that has a k-point
grid of 1x1x1. This script will perform sequence jobs with k-point grids of
1x1x1, 2x2x2, 3x3x3, 4x4x4, ... until convergence is achieved. The default
convergence criteria is 1meV/atom, but this can be set using the
--target option.
- -h, --help
- show this help message and exit
- -c [COMMAND],
--command [COMMAND]
- VASP command. Defaults to pvasp. If you are using mpirun, set this to
something like 'mpirun pvasp'.
- -i [{linear,inc}],
--increment_mode [{linear,inc}]
- Mode for increasing kpoints. In linear mode, multiples of the existing
kpoints are done. E.g., 2x4x2 -> 4x8x4 -> 6x12x6. In inc mode, all
KPOINTS are incremented by 1 at each stage, i.e., 2x4x2 -> 3x5x3
->4x6x4. Note that the latter mode does not preserve KPOINTS symmetry,
though it is probably less expensive.
- -m [MAX_STEPS],
--max_steps [MAX_STEPS]
- The maximum number of KPOINTS increment steps. This puts an upper bound on
the largest KPOINT converge grid attempted.
- -t [TARGET], --target
[TARGET]
- The target converge in energy per atom to achieve convergence. E.g., 1e-3
means the KPOINTS will be increased until a converged of 1meV is
reached.
Author: Shyue Ping Ong