Problems running VASP: crashes, internal errors, "wrong" results.
Moderators: Global Moderator, Moderator
-
guyohad
- Newbie
- Posts: 14
- Joined: Mon Feb 15, 2021 9:42 am
#1
Post
by guyohad » Sun May 07, 2023 1:21 pm
Dear VASP developers,
I am using VASP 6.4.1 in interface with wannier90 3.1.0, and I have encountered some issues.
1. The UNK file is written
only if exclude_bands is not specified
and projections are used. The UNK file is not written unless the two conditions are met.
2. Even when the two conditions are met, I get the following error regarding the projections:
Code: Select all
-----------------------------------------------------------------------------
| |
| W W AA RRRRR N N II N N GGGG !!! |
| W W A A R R NN N II NN N G G !!! |
| W W A A R R N N N II N N N G !!! |
| W WW W AAAAAA RRRRR N N N II N N N G GGG ! |
| WW WW A A R R N NN II N NN G G |
| W W A A R R N N II N N GGGG !!! |
| |
| The wannier interface is returning 12 projections while NUM_WANN is |
| 8. I will continue by using NUM_WANN but you should probably change |
| NUM_WANN or the wannier90 input file. |
| |
-----------------------------------------------------------------------------
even though my NUM_WANN is 12. Wannier90 then outputs the error:
Code: Select all
wannier90.amn has not the right number of projections
I attach the input and main output files.
I appreciate any help.
Best,
Guy
You do not have the required permissions to view the files attached to this post.
-
manuel_engel1
- Global Moderator
- Posts: 125
- Joined: Mon May 08, 2023 4:08 pm
#2
Post
by manuel_engel1 » Tue May 09, 2023 8:58 am
Dear Guy,
Thank you for posting on the forum. I have examined and run the system you provided.
Unfortunately, I am unable to reproduce the first issue. After specifying
exclude_bands and the projections block, VASP still generates the
UNK file. Could you provide additional information or input files regarding this first issue?
The second issue is due to a limitation in VASP's wannier90 interface (see
num_wann). I suggest to specify
in the
INCAR file instead of the
wannier90.win file. This always overwrites the setting in wannier90. Not specifying
num_wann in the
INCAR file allows VASP to generate a default value, which in your case equals 8.
The recommended way of providing the wannier90 input parameters is via the
wannier90_win tag.
I hope this information is helpful.
Best,
Manuel
Manuel
VASP developer
-
guyohad
- Newbie
- Posts: 14
- Joined: Mon Feb 15, 2021 9:42 am
#3
Post
by guyohad » Tue May 09, 2023 10:48 am
Dear Manuel,
Thank you for your reply. I have ran some more tests and have encountered 3 issues:
1. If no projections are specified, the UNK file is not generated. See problem01.tar.
2. As you said, the UNK file is indeed generated when specifying NUM_WANN in INCAR, and using projections. However, I get an error when running wannier90:
Code: Select all
wannier90.amn has not the right number of projections
See problem02.tar.
3. If I specify projections for the UNK file to be generated, and then remove them in the wannier90 run to avoid problem number 2, I get a system error when I run wannier90:
Code: Select all
end-of-file during read, unit 11, file UNK00001.1
See problem03.tar.
None of the above problems is solved upon using the tag WANNIER90_WIN.
Best wishes,
Guy
You do not have the required permissions to view the files attached to this post.
-
manuel_engel1
- Global Moderator
- Posts: 125
- Joined: Mon May 08, 2023 4:08 pm
#4
Post
by manuel_engel1 » Tue May 09, 2023 3:26 pm
Dear Guy,
Thank you for providing more details. I will try my best to address the issues you have listed.
- problem01 is indeed an issue that can be traced back to a bug in VASP. Not specifying the projections causes VASP to crash (or silently fail), which happens before the UNK files are written. Unfortunately, there is currently no way to only write UNK files.
- problem02 is related to the projections block in wannier90. By specifying
Code: Select all
begin projections
Si:sp3
end projections
wannier90 expects 8 local functions (4 for each Si atom). This is incompatible with num_wann = 4,
resulting in the error you observe. You can adjust the number of projection functions, for example by specifying
Code: Select all
begin projections
Si:s,px
end projections
Unfortunately, this will still not run without error due to problem03.
- problem03 is due to the use of the gamma-only version of VASP. It is likely there is a bug in VASP that causes the UNK files to be incorrect in this case. This causes the error message you reported when wannier90 tries to read the file. This issue definitely requires further investigation on our end. As a workaround, I suggest to use the standard (complex) version of both codes.
Here is a modified
INCAR that generates the
wannier90.win input file and runs without errors using the standard version of VASP:
Code: Select all
# comment
System = Si
ISTART = 1
ENCUT = 300
ALGO = N; EDIFF = 1.e-6
ISMEAR = 0; SIGMA = 0.01
NBANDS = 12
NUM_WANN = 4
LWANNIER90 = .TRUE.
LWRITE_UNK = .TRUE.
WANNIER90_WIN = "
exclude_bands = 5-12
wannier_plot = true
wannier_plot_list = 1
wannier_plot_supercell = 1
begin projections
Si:s,px
end projections
"
Best,
Manuel
Manuel
VASP developer
-
guyohad
- Newbie
- Posts: 14
- Joined: Mon Feb 15, 2021 9:42 am
#5
Post
by guyohad » Tue May 09, 2023 5:20 pm
Dear Manuel,
I highly appreciate your help and your quick replies, and hope that this report will contribute to future bug fix in VASP.
All the best,
Guy