[Pw_forum] hybrid calculation crashed

Tiana Davide davide.tiana at epfl.ch
Sat Nov 7 11:18:21 CET 2015


Dear
I run a single point calculation on a pbesol optimised structure using hse
After obtaining the 1st scf convergence, pw crashed during the exchange calculation after 16 cycles giving this error:

dexx is negative!   Check that exxdiv_treatment is appropriate for the system

I am running at gamma point using nc pseudopotential. This is the setting I've been using (there are 2 Tb atoms this is why such a high cut-off)

 &system
   input_dft='hse'
   ibrav=0
   nat=84,
   ntyp=5,
   ecutwfc=170.0,
   ecutrho=640
   ecutfock=600
   vdw_corr='grimme-d2'
   occupations='smearing',
   smearing='marzari-vanderbilt'
   degauss=0.05
   nspin=2
   starting_magnetization(1)=0.25,
   starting_magnetization(2)=-0.25,
 /

Any idea of why does this happen?

Many thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20151107/4564c697/attachment.html>


More information about the users mailing list