Page 3 of 3

Re: Gap convergence with vacuum space between layers

Posted: Sat Apr 16, 2016 9:28 pm
by Fabiof
Dear Daniele,
I have done two calculations with the same variables but with different QPKrange indices:

First calculation:

Code: Select all

%QPkrange                    # [GW] QP generalized Kpoint/Band indices
  1|  7|  3|6|
%
Second calculation:

Code: Select all

%QPkrange                    # [GW] QP generalized Kpoint/Band indices
  1|  7|  1|8|
%
In both report files, [07.02] QP properties and I/O shows the same QP energies for the different calculations (no problems here)

But when using the ypp to interpolate bands i get different energies for the gaps.

Perhaps the problem that i had before, different energies for he same calculations, is related to this (different indices, different ypp interpolated band energies).

Is this expected to happen? The report files and interpolated bands are attached.

Re: Gap convergence with vacuum space between layers

Posted: Mon Apr 18, 2016 4:20 pm
by Davide Sangalli
Dear Fabio,
from the files you attached it seems you did everything correctly.

I suspect the difference is generated when loading the ndb.QP
If I'm not wrong there was a bug related to this problem and we released the fix for that in the non GPL version only up to now.
We hope to release the fix in the GPL version as well soon.

To understand better the problem could you perform a simple IP absorption calculation for the two runs (yambo -o b) including in input the variable
GfnQPdb= "E < ../5-PPA200/ndb.QP" # [EXTQP G] Database
and attach output and report files

Best,
Davide

Re: Gap convergence with vacuum space between layers

Posted: Wed Apr 27, 2016 4:56 pm
by Fabiof
Dear all,

Sorry for taking so long to attach the report files.
I have done the yambo -o b calculations for the two different PPA runs. The fiput and output are attached (1 for Band indices 3 to 6) and (2 for Band indices 1 to 8).


Thanks,

Fabio