[QE-users] problem computing cholesky

[BARRETEAU Cyrille]

Dear QE users & developers

I am facing a a problem when trying to perform magneto-crrystalline anisotropy calculation using the Force-Theorem implementation of QE (as explained in the Force Theorem example).
I first perform the scf calculation (scalar relativistic pseudo) that runs smoothly. The problem arises in the nscf calculation (relativistic pseudo) starting from the previous scf scalar relativistic charge-density.dat.

The calculation very often fails due to problem in computing cholesky.
This occurs in a rather erratic way depending on tiny changes in the input parameters or even on the number of nodes on the which the calculation is performed. I have tried changing the diagonalization technique (cg or david), increasing the diago_thr_init, other relativistic pseudo etc...But I could not find a reliable solution that systematically solves the problem.

I have another question related to the restarting nscf from previous scf scalar relativistic.
In the new version of qe (6.2.1) I got a warning message: "some spin components not found".
Should I worry about this message?

Thanks
Cyrille



========================
Cyrille Barreteau
CEA Saclay, IRAMIS, SPEC Bat. 771
91191 Gif sur Yvette Cedex, FRANCE
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+33 1 69 08 38 56 /+33  6 47 53 66 52  (mobile)
email:     cyrille.barreteau at cea.fr
Web:     http://iramis.cea.fr/Pisp/cyrille.barreteau/
========================
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180705/e66d41ae/attachment.html>