[CP2K-user] RESTART files and convergence

[Nicholas Winner]

I have an .wfn restart file from an ENERGY_FORCE run that I am using as the 
initial guess for a GEO_OPT calculation. The restart file was generated 
from a converged SCF, so presumably it should only require 1 SCF step and 
be converged and then move to the geometry optimization, but the restarted 
calculation still takes many SCF steps.

Am I wrong in thinking that it should be converged right away?

For reference, I am using default diagonalization, direct_p_mixing 
(alpha=.05), and smearing at 300K.

Thank you
Nick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210712/8ff2c6e1/attachment.htm>