[Pw_forum] k points in epsilon.x

[Gabriele Sclauzero]

First:

    Andrea Ferretti is Sir, Layla Martin Samos is Madam, not the opposite...

dev sharma wrote:
> Dear Madam,
> 
> May be my question was not clear to all. originally my question was the 
> same . That as u suggested do a scf and then  a nscf with the nbnd and 
> regular mesh of k points. What i want to know , that in nscf calculation 
> i have to give
> kpoints {automatic}
> 4 4 4
> or like
> Kpoints
> 72
> 0 0 0
> 0 0.1 0
> /
> /
> 0.5 0 0
> 
> and i asked that if i have to give k points in second option, How to 
> give that ????  can i give these 72 k points generated by the scf 
> output  ????

Second:

Many people has already replied you, giving different possible solutions to your problem. 
I'm quite astonished that after a dozen of mails we have not been able to close this 
thread. Maybe there is some problem in communication/understanding which I cannot capture.
I summarize what are the possible solutions if you still want to use epsilon.x and don't 
want to try Yambo or SaX and co.
In the nscf run, either use

nosym=.TRUE.
noinv=.TRUE.

K_POINTS AUTOMATIC
4 4 4 0 0 0

Such that the code will NOT use symmetry to reduce the 4x4x4 k-point mesh.

Or use:

nosym=.TRUE.

K_POINTS CRYSTAL


and then append a list of k-points corresponding to a 4x4x4 regular grid of key points 
that you can easily generate either using the piece of code given by Layla, or adapting 
the code in pwtools/kpoints.f or PW/kpoint_grid.f90.

Obviously you must test for convergence with respect to the BZ sampling (i.e. number of 
k-points)

HTH


GS







> Thnaks
> Dev Sharma,
> Unoveristy of Delhi
> 
> On Mon, Oct 12, 2009 at 8:21 PM, Andrea Ferretti <ferretti at mit.edu 
> <mailto:ferretti at mit.edu>> wrote:
> 
> 
> 
>     Hi,
> 
>      > Sir, i want to calculate dielectric constant as a function of
>     frequency.  I
>      > have read the manual And it is written there that
>      > Epsilon.x doesn't support the reduction of the k-points grid into the
>      > unreducible Brillouin zone, so the previous PW runs must be
>     performed with a
>      > uniform k-points grid and all k-points weights must be equal to
>     each other,
>      > i.e. in the k-points card the k-points coordinates must be given
>     manually in
>      > \emph{crystal} or \emph{alat} or \emph{bohr}, but not with the
>      > \emph{automatic} option.
>      > Thanks
>      >
> 
>     I think this is correct..
>     a good procedure could be first to run a scf calculation using whatever
>     symmetrized kpt-mesh. Then you can run a nscf calculation with a
>     suitable
>     number of empty bands (according to your needs) and a regular mesh
>     of kpts
>     over the whole BZ.
> 
>     andrea
> 
>     DMSE, MIT
>     Massachusetts
> 
>     _______________________________________________
>     Pw_forum mailing list
>     Pw_forum at pwscf.org <mailto:Pw_forum at pwscf.org>
>     http://www.democritos.it/mailman/listinfo/pw_forum
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum

-- 


o ------------------------------------------------ o
| Gabriele Sclauzero, PhD Student                  |
| c/o:   SISSA & CNR-INFM Democritos,              |
|        via Beirut 2-4, 34014 Trieste (Italy)     |
| email: sclauzer at sissa.it                         |
| phone: +39 040 3787 511                          |
| skype: gurlonotturno                             |
o ------------------------------------------------ o