VASP Forum

Dear All,
I'm trying to calculate the static dielectric constant in a bulk perovskite.

This is comparatively large system

1.00000000000000
23.5804004668999987 0.0000000000000000 0.0000000000000000
0.0000000000000000 9.8221998214999999 0.0000000000000000
0.0000000000000000 0.0000000000000000 17.1207008362000010
Eu S O N H C
8 20 124 16 232 64
Direct
0.3035240006161741 0.7044972780750686 0.6088067019479594
0.6964805902531316 0.2954869898532566 0.1088101149644477

1. First I have relaxed the system
2. Now i'm using this INCAR

SYSTEM = Eu-DABCO full static dielectric finite differences

ISTART = 0
ICHARG = 2
ISPIN = 1

ENCUT = 520
PREC = Accurate
EDIFF = 1E-8
NELM = 250
NELMIN = 6
ALGO = Normal

LREAL = .FALSE.
ADDGRID = .TRUE.
LASPH = .TRUE.

ISMEAR = 0
SIGMA = 0.01

IVDW = 12

LCALCEPS = .TRUE.

IBRION = 6
NFREE = 2
POTIM = 0.015
NSW = 1

ISYM = 2
SYMPREC = 1E-5

LWAVE = .FALSE.
LCHARG = .FALSE.

K-Spacing Value to Generate K-Mesh: 0.020
0
Gamma
1 1 1
0.0 0.0 0.0

The calculation runs due to the limit of the walltime.

My questions are
1. Is this setup correct?
2. What strategy can I use to restart the calculation when stops

Thank you

If you are interested in computing the static dielectic function only then I would remove the tags related to the finite differences ionic displacements:

IBRION = 6
NFREE = 2
POTIM = 0.015
NSW = 1

These doing finite displacements of the atoms which (wihout looking at the OUTCAR) is probably why the calculation is taking so long and you are running out of walltime. If you share the OUTCAR and OSZICAR I can tell you with more certainty.
We currently support CHECKPOINT_FD which allows you to continue a IBRION=6 calculation. However if you are only interested in the dielectric function, then this functionality will not help.