[Pw_forum] k points weights

[Paolo Giannozzi]

On Jul 27, 2012, at 10:39 , Davide Tiana wrote:

> As remembered in many QE tutorials, weights are meaningless in NSCF
> calculations such as band ones.

yes and no, actually.

'nscf' calculation. The output data is meant to be used in subsequent
calculations performing integrals over the BZ (e.g. dos); symmetry is
applied and occupancies are calculated (using weights among other
things) as for the 'scf' case

'band' calculation. Tthe output data is meant to be used only for  
visualization;
symmetry is not applied, occupancies need not to be calculated, weights
are in principle not used. Since the latest release, however, weights  
are
used in the 'band' case as well to re-compute occupancies, because there
is a specific case in which data to be exported to another code (not  
part of
QE) have to be calculated using 'band'  and not 'nscf', for reasons  
related
to the way symmetries are handled.

So, in conclusion: while weights in 'band' calculations are in most  
cases
meaningless, it is a good idea not to fill them with absurd numbers  
(e.g.
all zero or negative or very large) because the code could complain when
calculating occupancies

P.
---
Paolo Giannozzi, Dept of Chemistry&Physics&Environment,
Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
Phone +39-0432-558216, fax +39-0432-558222