Wrong 'num_bands' determined from vasp in wannier input file

Question on input files/tags, interpreting output, etc.

Please check whether the answer to your question is given in the VASP online manual or has been discussed in this forum previously!

Moderators: Moderator, Global Moderator

Post Reply
Message
Author
zhishuo_huang
Newbie
Newbie
Posts: 5
Joined: Fri May 10, 2024 11:50 am

Wrong 'num_bands' determined from vasp in wannier input file

#1 Post by zhishuo_huang » Wed Jun 05, 2024 8:59 am

Dear developers and users,

I am trying to do wannier calculations with vasp 6.4.3 with wannier90 3.1.0, compiled with intel compiler 2022.
I did successfully the SCF calculations.
While in the wannier calculations, the NBANDS was set to 276, and I tried to exclude 264 irrelevant bands via "exclude_bands : 1-152, 165-276", which means that I want to get 12 wannier orbitals from 12 isolated bands.
However, after calculation, I found that the "num_bands=24" in the wannier.win file.
I also checked the necessary files generated by VASP, as shown below
There are 24 states included in wannier90.eig:
###############################################################
1 1 2.934803017705
2 1 2.934803017705
3 1 2.934803017705
4 1 3.460789566715
5 1 3.481895510668
6 1 3.481895510682
7 1 3.804205358207
8 1 3.804205358208
9 1 3.804205358216
10 1 3.815709548970
11 1 3.815709548971
12 1 3.815709548981
13 1 28.513006307570
14 1 28.563092703652
15 1 28.790261663631
16 1 29.102905859108
17 1 29.291569530572
18 1 29.329144655855
19 1 29.370129570034
20 1 29.509888350947
21 1 29.775100102854
22 1 30.204202553026
23 1 30.495424856830
24 1 31.605076223788
1 2 3.002098636571
2 2 3.065485811690
3 2 3.065485811690
4 2 3.393219140124
5 2 3.523364511400
6 2 3.531272172565
7 2 3.696287806069
8 2 3.696287806085
###############################################################

The head two lines of wannier90.amn is
###############################################################
File generated by VASP: VASP_600LT_Gm
24 216 12
###############################################################

The head three lines of wannier90.mmn is
###############################################################
File generated by VASP: VASP_600LT_Gm
24 216 6
1 2 0 0 0
###############################################################

The wannier90.win file is shown :
###############################################################
exclude_bands : 1-152, 165-276

Begin Projections
Os:dxy;dxz;dyz:r=2
End Projections

num_iter = 1000
dis_win_min=2.0
dis_win_max=4.0
write_hr=.true.
write_rmn=.true.
write_tb=.true.
wannier_plot=.true.

!restart = plot
bands_plot = true

begin kpoint_path
G 0.00 0.00 0.00 X 0.50 0.00 0.00
...
...
X 0.50 0.00 0.00 R 0.50 0.50 0.50
end kpoint_path
guiding_centres = true

# This part was generated automatically by VASP
num_bands = 24
num_wann = 12
begin unit_cell_cart
8.3998100 0.0000000 0.0000000
0.0000000 8.3998100 0.0000000
0.0000000 0.0000000 8.3998100
end unit_cell_cart
begin atoms_cart
Ba 2.0999525 2.0999525 2.0999525
...
...
O 4.1999050 2.2934661 0.0000000
O 4.1999050 4.1999050 1.9064389
O 4.1999050 4.1999050 6.4933711
end atoms_cart
mp_grid = 6 6 6
begin kpoints
0.083333333333 0.083333333333 0.083333333333
...
...
0.416666666667 -0.416666666667 0.416666666667
-0.416666666667 0.416666666667 -0.416666666667
end kpoints
###############################################################

All these information shows that the wannier procedure is done from 24 band states to 12 wannier orbitals.
Below attached the INCAR for this calculation:
###############################################################
SYSTEM = VASP_600LT_Gm

ALGO = All
EDIFF = 1E-7
EDIFFG = -1E-4
ENCUT = 600
LREAL = .FALSE.
LCHARG = .FALSE.
PREC = Accurate

LWAVE = .TRUE.
LCHARG =.TRUE.
NELM = 200
ISMEAR = 0
SIGMA = 0.01
NCORE=1

NBANDS=276
LWANNIER90 = .TRUE.
LWRITE_UNK = .TRUE.
LWRITE_MMN_AMN=.TRUE.
LWRITE_WANPROJ = .TRUE.

NUM_WANN=12

!LOCPROJ = 13 14 15 16 : dxy dyz dxz : Ps 2
!NTARGET_STATES
WANNIER90_WIN="
num_bands = 276
exclude_bands : 1-152, 165-276

Begin Projections
Os:dxy;dxz;dyz:r=2
End Projections

num_iter = 1000
dis_win_min=2.0
dis_win_max=4.0

mp_grid = 6 6 6
write_hr=.true.
write_rmn=.true.
write_tb=.true.
wannier_plot=.true.

!restart = plot
bands_plot = true
begin kpoint_path
G 0.00 0.00 0.00 X 0.50 0.00 0.00
X 0.50 0.00 0.00 M 0.50 0.50 0.00
M 0.50 0.50 0.00 G 0.00 0.00 0.00
G 0.00 0.00 0.00 R 0.50 0.50 0.50
R 0.50 0.50 0.50 M 0.50 0.50 0.00
M 0.50 0.50 0.00 X 0.50 0.00 0.00
X 0.50 0.00 0.00 R 0.50 0.50 0.50
end kpoint_path
guiding_centres = true

"
###############################################################

In the INCAR, it is indicated that the wannierization should be carried out between 12 band states and 12 wannier orbitals, contradicting with the results.
I will appreciate it if anyone can provide any clue for this problem.

Best regards
Zhishuo Huang

svijay
Global Moderator
Global Moderator
Posts: 73
Joined: Fri Aug 04, 2023 11:07 am

Re: Wrong 'num_bands' determined from vasp in wannier input file

#2 Post by svijay » Thu Jun 06, 2024 2:18 pm

Dear Zhishuo Huang,

Could you please share all the input files that you used to run VASP (https://www.vasp.at/wiki/index.php/Mini ... 20possible.)? That would make it easier for us to reproduce the issue.

Sudarshan

zhishuo_huang
Newbie
Newbie
Posts: 5
Joined: Fri May 10, 2024 11:50 am

Re: Wrong 'num_bands' determined from vasp in wannier input file

#3 Post by zhishuo_huang » Fri Jun 07, 2024 5:56 am

Dear Sudarshan,

Thank you for your commit.
Please find the attachment for the reproduction.
The calculation includes three steps, SCF procedure, NSCF calculation with more bands, and wannier calculation.
The procedure is controlled by the pbs file.

Best regards
Zhishuo Huang
You do not have the required permissions to view the files attached to this post.

svijay
Global Moderator
Global Moderator
Posts: 73
Joined: Fri Aug 04, 2023 11:07 am

Re: Wrong 'num_bands' determined from vasp in wannier input file

#4 Post by svijay » Wed Jun 12, 2024 9:21 am

Dear Zhishuo Huang,

Sorry about the delayed reply and thank you for the files. Could you also please send the OUTCAR and OSZICAR files for all of these calculations?

Sudarshan

zhishuo_huang
Newbie
Newbie
Posts: 5
Joined: Fri May 10, 2024 11:50 am

Re: Wrong 'num_bands' determined from vasp in wannier input file

#5 Post by zhishuo_huang » Fri Jun 14, 2024 9:20 am

Dear Sudarshan,

Please find the attachment for the calculation files with OUTCAR and OSZICAR in each step.

Best regards
Zhishuo Huang
You do not have the required permissions to view the files attached to this post.

svijay
Global Moderator
Global Moderator
Posts: 73
Joined: Fri Aug 04, 2023 11:07 am

Re: Wrong 'num_bands' determined from vasp in wannier input file

#6 Post by svijay » Fri Jun 14, 2024 3:30 pm

Unfortunately I was not able to reproduce your error, also with the latest release of VASP - notably I changed ALGO=Exact to ALGO=All in the NSCF step to make the calculation run quickly for the tests. Something to try would be to do all of this consistently in one-shot (wiki/index.php/Constructing_Wannier_orbitals) and see if the error persists.

Post Reply