[Pw_forum] gk_sort igk(ig) versus igk_k(ig, ik) possible bug in old versions of wfck2r.f90?

[aritz leonardo]

Dear all,

I have just successfully installed qe-6.0 and I realized this version, 
the k+G to G correspondence has changed. The old igk(ig) has been 
changed for igk_k.

My question is regarding the program "wfck2r.f90" of the PP section 
which transforms the wave-function from k-space to r.

**************** wfck2r.f90 from  qe-5.3.0 (and previous versions) reads:
   DO ik = 1,nks
      !
      !    prepare the indices of this k point
      !
      CALL gk_sort (*xk (1, ik)*, ngm, g, ecutwfc / tpiba2, npw, &
           igk, g2kin)

      CALL davcio (evc, 2*nwordwfc, iunwfc, ik, - 1)

      do ibnd=1,nbnd
      !
      ! I perform fourier transform
      !
         evc_r = cmplx(0.d0, 0.d0)
         do ig = 1, npw
            evc_r (nls (igk (ig) ),1 ) = evc (ig,ibnd)
         enddo
         CALL invfft ('Wave', evc_r(:,1), dffts)

********************************************************************

Shouldn't one pass  to gk_sort the hole k point instead of the first 
coordinate? *i.e. xk(:,ik) instead of xk(1,ik) *

(I am aware that in qe-6.0 this part has been replaced with igk_k.)

Thanks in advance!



-- 

============================================================
Aritz Leonardo Liceranzu
Department of Applied Physics II,
Faculty of Science and Technology,
University of the Basque Country (UPV/EHU)
Bº Sarriena s/n, 48940 Leioa, Spain

http://fisica.ehu.es/~aritz       Fax: +34-946013500
Mail: aritz.leonardo at ehu.es       Phone: +34-946015338
============================================================

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20170302/4e6a2394/attachment.html>