[QE-users] How is the wavefunction supposed to be normalized?

[John McFarland]

Dear Stefano,


I tried to write some code that is like the code you wrote, using both s_1psi() ans s_psi(), and I tested it with a wavefunction I believe is normalized.  The result is a matrix that resembles the identity matrix, except that the diagonal elements are not one but range from about 0.2 to 0.3.  I got a similar result when I tested the subroutine check_ortho() in ortho_wfc.f90.  Is there something I'm missing?


Here is an example of some code I tested:


      call calbec( npw, vkb, evc, becp )
      call s_psi(npwx, npw, nbnd, evc, tddft_spsi)

      do ibnd = 1, nbnd
        do jbnd = 1, nbnd

          ! calculate <psi_i| S |psi_j>
          norm = 0

          do ipw = 1, npw
            norm = norm + conjg( evc(ipw,ibnd) ) * tddft_spsi(ipw,jbnd)
          enddo

          write(stdout,*) " i " , ibnd , " j " , jbnd , "<psi_i| S |psi_j>" , norm

        enddo
      enddo



Best regards,

John


________________________________
From: users <users-bounces at lists.quantum-espresso.org> on behalf of Stefano de Gironcoli <degironc at sissa.it>
Sent: Tuesday, December 11, 2018 10:26:15 AM
To: users at lists.quantum-espresso.org
Subject: Re: [QE-users] How is the wavefunction supposed to be normalized?


yes, except the S matrix is never stored anywhere but is applied to the wfc by the s_1psi routine,

so more something like


 sum = 0.d0

 call s_1psi( npwx, npw, evc(1,b), spsi )

 DO  i = 1,npw

     sum = sum + conjg(evc(i,b)) * spsi(i)
 END DO




stefano


On 11/12/18 15:36, John McFarland wrote:

Dear all,


As I understand it, the wavefunction for each k-point is given by ecv(:,:), with the first index representing a single particle basis and the second index representing each band.  I'm guessing that when a normalized wavefunction is contracted over the S matrix, the result should equal one for each band.  What I think this would look like in code is:


 sum = 0

 DO  i = 1,npwx

    DO j = 1,npwx

       sum = sum + evc(i,b) * S(i,j) * evc(j,b)

    END DO

 END DO


and sum should equal 1 for any band b.  Is my understanding correct?


Best regards,

John McFarland



_______________________________________________
users mailing list
users at lists.quantum-espresso.org<mailto:users at lists.quantum-espresso.org>
https://lists.quantum-espresso.org/mailman/listinfo/users<https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.quantum-2Despresso.org_mailman_listinfo_users&d=DwMD-g&c=MNHwOqQ8N1u91SoMLfIblwuGXKgp50OPUXjl8uRAbak&r=aBqxGb3Z1QZuWuPQzYblq8v7-w7hDZgkvgBYwm8sq8Y&m=0_ldi_kPdbF9G-LbLtc8g1MiFx1QDdFx9WUZmApxI8M&s=dRaIjvkziJzsYp5az0bdiOPHS0PsijOxraxkUi-8mkk&e=>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20181224/8511810a/attachment.html>