Page 1 of 1

Hybrid functional band structure: parameters not read

Posted: Wed May 29, 2024 12:46 pm
by johan_felisaz
Hi,

I would like to run hybrid functional calculations for some class of transition metal dichalcogenides materials. However, VASP seems to not take into account some of the parameters of the INCAR file, that is:
- the WAVECAR is not read, even though ISTART=1, and there is a WAVECAR file, coming from a previous normal DFT run, with the same k mesh and number of bands.
- there is a warning about the tag KPOINTS_OPT_NKBATCH, which is supposed to be a supported tag (wiki page: https://www.vasp.at/wiki/index.php/KPOINTS_OPT_NKBATCH )

Here is the INCAR file:

Code: Select all

lhfcalc = true
nbands = 56
ispin = 2
system = few-layers-hybrid-full/ptse2-2
magmom = 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
hfrcut = -1
nwrite = 3
lasph = true
kpar = 1
algo = Damped
encut = 250
time = 0.5
gga_compat = false
sigma = 2.0e-5
lsorbit = true
lmaxmix = 4
ismear = 0
istart = 1
icharg = 1
KPOINTS_OPT_NKBATCH = 72
Finally, I just wanted to know if this sort of calculation is supposed to be this long, or if some parameters I have chosen are less than optimal. For 56 bands, 9x9x1 grid + 72 ks for band structure, It did not finish in 24 hours on 16 GPU nodes...

Thanks in advance

Re: Hybrid functional band structure: parameters not read

Posted: Wed May 29, 2024 1:15 pm
by henrique_miranda
Could you share your KPOINTS, POSCAR and OUTCAR files?

I would say that calculations should run in less than one hour on a few CPU cores (8 or so).
This of course depends on the size of the cell you are simulating (I can only guess from the number of bands).

Re: Hybrid functional band structure: parameters not read

Posted: Wed May 29, 2024 1:44 pm
by johan_felisaz
Sorry I prepared the input files, but forgot to submit them.

They are now in the attachment.