VASP Forum

Hello Experts,
System is NiSi (orthorhombic 59) without SOC or Magnetization. I'm using VASP 6.4.2
I'm trying to successfully connect for VASP to Wannier90. But when I do the MLWs in wannier90, I get the error

Error reading wannier90.mmn: 1 1 145 0 0 0 0
Exiting.......
Neighbour not found

This error occurs due to some mistakes in the VASP to Wannier90 step. I have attached my INCAR, KPOINTS, POSCAR, POTCAR, CHGCAR, OUTCAR .win., .wout. And below are the first few lines of .mmn and .amn, which seems to be the corrupt outputs.

.mmn (Number 145 in the third line seems to be incorrect)

File generated by VASP: NiSi Orthorhombic
          32       10648           6
    1  145    0    0    0
   -0.527214350066   -0.849436273426
   -0.013830554603    0.014763329182

.amn
File Edit Options Buffers Tools Help

File generated by VASP: NiSi Orthorhombic
          32       10648          16
    1    1    1    0.000000000000   -0.000000000000
    2    1    1   -0.000000000000   -0.000000000000
    3    1    1    0.000000000000    0.000000000000
    4    1    1    0.000000000000    0.000000000000
    5    1    1    0.000000000000    0.000000000000
    6    1    1    0.000000000000    0.000000000000
    7    1    1   -0.526171377196   -0.352825134400
    8    1    1    0.000000000000   -0.000000000000
    9    1    1   -0.000000000000   -0.000000000000

Thank you for your help in advance!
Dinura

Hello Dinura,

Thank you for your question and for uploading the files. I'm afraid that I could not repeat your error (using Wannier90 3.1.0 and VASP 6.4.2). The calculation ran without error running in library mode (LWANNIER90_RUN = .TRUE. rather than LWANNIER90= .TRUE.).

Would you be able to share the file where you encountered this error? Perhaps that will make the issue clearer. The

1 1 145 0 0 0 0

certainly seems unusual, with the extra 1 before and the extra 0 after.

Best wishes,

Chris

edit: corrected second LWANNIER90_RUN to LWANNIER90 in the first paragraph

Hi Chris,

Thank you so much for your reply. Yes, I was able to run the VASP calculation, without an error. But I suspect that the files it produced (.mmn, .amn, and .eig) may be corrupted.

Unfortunately, I cannot attach the .mmn, .amn, and .eig files as they are too large. The input files for my Wannier90 calculation (MLWs) are those generated by this VASP step, .eig, .mmn, .amn, and .win.

I have attached the .win, slurm.sh, .out, and .wout files from my Wannier90 calculation. As you can see in the .wout file, the calculation abruptly stops during the step where it reads the .mmn file.

In addition, I have attached two more files, wannier90_version2.wout and the .werr file, which I sometimes obtain (I'm not sure why?). These files clearly indicate the error.

Thanks,
Dinura

Hi Dinura,

Thank you for the additional files. The format of your .mmn file looks similar to my own, so I would be surprised if that were the issue. I have repeated the calculation with a smaller mesh (2x2x2) and have been unable to repeat the error. However, the .mmn, .amn, and .eig files are only a few kb or mb, so you could upload those files (and your .win file). If you do, I can try to run Wannier90 with those, to see if I get the same error.

Best wishes,

Chris

Hi Chris,

Thanks, I played with it a little bit, and yes, it worked for the smaller meshes. But when I moved to denser meshes (above 22 × 22 × 22), error appeared. It might be a memory issue after all. I will look into it more closely.

Thank you so much for your help!

Dinura

Hi Dinura,

Glad that you've got it working.

Best,

Chris