My Community

I followed the example on wiki to do GW bandstructure of Si:
http://cms.mpi.univie.ac.at/wiki/index. ... NNIER90%29
Step by step exactly. I recompiled wannier90-1.2 and vasp-5.3 with correct flags without error.

I ran through the first 3 steps without any error but at this point, the three files wannier90.amn, wannier90.mmn and wannier90.eig are not generated. I have no idea why since there is no hint in the output file.

As a proof that the vasp compiled correctly, I checked the bandgap of Si given in Step 3. It was 1.1497 eV in good agreement with experimental result. INCAR I used for step 3 was:

ALGO = GW0
LSPECTRAL = .TRUE.
NOMEGA = 50
LRPA = .FALSE.
NBANDS = 64
LWANNIER90=.TRUE.

The value of NBANDS is the same as Step 2 (All input files are copied from the website). I saw one post with the same problem on the forum, but without any practical solution given. Any help would be appreciated.

I have a similar problem. In my case I run the wannier90 interface with the LWANNIER90 = .TRUE. tag and all I get is a wannier90.win file with just one line, indicating the number of bands as follows:

num_wann = 96 ! set to NBANDS by VASP

The wannier90.wout file contains the details of the system, like positions and k-points, but it doesn't enter the actual calculation loop.

Some help would be appreciated.