[DFTB-Plus-User] Calculation of Kpoints

[Mat Toliday]

  Dear Bálint,
Am I correct in thinking that the calculation can be split up further by the number of spin channels?Mat Tolladay
    On Friday, 13 November 2020, 10:48:30 CET, Bálint Aradi <aradi at uni-bremen.de> wrote:  
 
 Dear Carlo,

> As far as I understood the best value of *n* should be a multiple of
> Kpoints, if possible.

That's true, but only if you use process groups (Parallel/Groups). If
you for example have nK K-points and nProc processes (with nProc = n *
nK), then creating n groups would be the most efficient parallelisation.

> My question is: *is there any way to compute in advance the number of
> Kpoints without running dftb+?*

Since DFTB+ only reduces the K-points by the inversion symmetry, it is
often easy to guess nr. of k-points in advance. For the commonly used
MP-scheme with

SupercellFolding {
  n1 0 0
  0 n2 0
  0  0 n3
  s1 s2 s3
}

it would be

- (n1 * n2 * n3) / 2 if n1, n2, n3 are even and s1 = s2 = s3 = 0.5

- (n1 * n2 * n3 + 1) / 2 if n1, n2, n3 are odd and s1 = s2 = s3 = 0.0
  or s1 = s2 = s3 = 0.5

- and probably (n1 * n2 * n3) in most other cases.


Best regards,

Bálint


Ps. In case, you are interested to contribute: It would actually not too
complicated to assamble a small tool, which parses the DFTB+ input and
spits out the k-points.

-- 
Dr. Bálint Aradi
Bremen Center for Computational Materials Science, University of Bremen
http://www.bccms.uni-bremen.de/cms/people/b-aradi/

_______________________________________________
DFTB-Plus-User mailing list
DFTB-Plus-User at mailman.zfn.uni-bremen.de
https://mailman.zfn.uni-bremen.de/cgi-bin/mailman/listinfo/dftb-plus-user
  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20201113/4aa7f714/attachment.htm>