[Pw_forum] projwfc.x error with spin orbit

[Virk Naunidh]

Dear Lorenzo,

Thanks for your help. We tried your suggestion.

We find that prior to crashing the capel variable is much higher than 1d-9 (1.58024691358025 for the case we tested).

I guess this means that the matrix is not unitary. Do you have any other suggestions?

Kind regards,
--
Naunidh Virk
EPFL, ITP, Lausanne, CH.

________________________________________
De : pw_forum-bounces at pwscf.org [pw_forum-bounces at pwscf.org] de la part de Lorenzo Paulatto [lorenzo.paulatto at impmc.upmc.fr]
Date d'envoi : mercredi 20 juin 2012 11:05
À : PWSCF Forum
Objet : Re: [Pw_forum] projwfc.x error with spin orbit

On Wed, Jun 20, 2012 at 10:49 AM, Virk Naunidh <naunidh.virk at epfl.ch<mailto:naunidh.virk at epfl.ch>> wrote:
Dear Users,

We are trying to project the wavefunction for a series of slab systems with spin orbit using projwfc.x, however for a certain number of those systems the calculation did not work and we were presented with the following error message:

 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     from d_matrix_so : error #         2
     D_S (j=1/2) for this symmetry operation is not unitary
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


Dear Virk,
the epsilon used to test for unitarity is very small. Maybe the matrix is indeed unitary but some numerical noise make it looks like it's not. The fact that your system is a slab with plenty of vacuum, and has low symmetry may also contribute. The fact that it only happens for some systems suggest that it is caused by numerical noise. It may also suggest that your calculation is not completely converged wrt the wfc cutoff.

Try to open PP/src/d_matrix_so.f90 and at line 42 change
  REAL(DP), PARAMETER :: eps = 1.0d-9
to something a bit larger. I.e. 1.d-8

It may also be a good idea to go down to line 282 and following to have the variable "capel" printed on output, which would give you an idea of how much non-unitary the matrix is.

If these suggestions do not solve the problem, please come back with more details, if the do, please confirm.

bests


--
Lorenzo Paulatto IdR @ IMPMC/CNRS & Université Paris 6
phone: +33 (0)1 44275 084 / skype: paulatz
www:   http://www-int.impmc.upmc.fr/~paulatto/
mail:  23-24/4é16 Boîte courrier 115, 4 place Jussieu 75252 Paris Cédex 05