Page 1 of 1

LSCDM method and w90?

Posted: Tue Mar 01, 2022 8:28 am
by michaela._fechner
Dear all,

I noticed that also VASP has an implementation of LSCDM method to compute WannierFunctions, however,
although the code mentioned it I cannot get it to work. I am using the most recent VASP version (6.2.1), linked against W90 3.1. Then I try create a simple example for SrTiO3 perovskite the input looks as follows (see below),
although specified the code complains that no projections are defined and than w90 crashes consequently.

So is there a way to get this working?

best
M


INCAR:

SYSTEM = STO

ISMEAR = 0
SIGMA = 0.01
EDIFF = 1E-8
ENCUT = 600
NELM = 200
ISYM = 2

LWANNIER90_RUN=.TRUE.
LSCDM=.TRUE.

LWANNIER90_WIN="
# disentanglement
dis_num_iter = 500
dis_conv_tol = 1e-8

#iterations
num_iter = 500
num_print_cycles = 50
conv_tol = 1e-4

"


POSCAR:
STO
1.00000000000000
3.8979836085520994 0.0000000000000000 0.0000000000000000
0.0000000000000000 3.8989836085520994 0.0000000000000000
0.0000000000000000 0.0000000000000000 3.8999836085520994
Sr Ti O
1 1 3
Direct
0.25000000000000000 0.25000000000000000 0.25000000000000000
0.75000000000000000 0.75000000000000000 0.75000000000000000
0.75000000000000000 0.75000000000000000 0.25000000000000000
0.75000000000000000 0.25000000000000000 0.75000000000000000
0.25000000000000000 0.75000000000000000 0.75000000000000000

KPOINTS:
A-Points
0
Gankov
3 3 3
0 0 0

Re: LSCDM method and w90?

Posted: Tue Mar 01, 2022 3:32 pm
by michaela._fechner
Okay, issue solved the recent vasp version is 6.3.0, where LSCDM works

best
Michael

Re: LSCDM method and w90?

Posted: Wed Mar 02, 2022 6:50 am
by ferenc_karsai
Yes VASP 6.3.0 has the method implemented. Thank you very much for reporting back.