Dear Chenhai,
indeed that's strange. I suggest you to reduce the number of bands to make some validation tests from scratch.
Consider reducing the number of bands in order to have a smaller BSE matrix and then try the different solvers (diagonal, haydock, inv.). The three should provide the same spectrum, and you can have an idea on how the spectrum looks like. Finally refine the calculation adding bands up to convergence.
I suggest you to remove the old database and regenerate them from scratch, also consider to work with more simplified input files keeping only what you really need.
Best,
Daniele
Incorreect Im(alpha)
Moderators: Davide Sangalli, andrea.ferretti, myrta gruning, andrea marini, Daniele Varsano
- Daniele Varsano
- Posts: 4207
- Joined: Tue Mar 17, 2009 2:23 pm
- Contact:
Re: Incorreect Im(alpha)
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/