[QE-users] bfgs failed after 11 scf cycles and 10 bfgs steps, convergence not achieved

[Kazume NISHIDATE]

Dear Mike Pionteck

> bfgs failed after 11 scf cycles and 10 bfgs steps, convergence not
> would like to use these settings. Does anyone know how I can fix
  this problem?


I'm not sure but the following procedure usually works for me.	

- Edit the output file of your failed bfgs run.

- Go to the last line of the file.
- Back search the 'ATOMIC_POSITIONS' section.


You will find the 'last' atomic positions at the 'bfgs' step.

- Copy all the atomic positions and paste it into the input file.


Now you have a new input file. Delete (or copy) the original output
file and run the next JOB with the new input file.


One reason why it works may be that the last atomic positions in the
'faled bfgs step' is not so bad assumption as the starting point for
the next bfgs steps (and also because the atomic positions at the
previous time step will not be used to construct the new bfgs
dynamics). Repeat the above procedure until you can get the converged
one.


You can also try the other option for the 'ion_dynamics' such as
'dump'. Alternately, quenching Molecular Dynamics, decreasing its
temperature to 0 K, may also work.


Check the input options.



西館数芽
Kazume NISHIDATE Ph.D

Department of Systems Innovation Engineering,
Graduate School of Science and Engineering, Iwate University
4-3-5 Ueda, Morioka, Iwate, 020-8551 JAPAN
Phone:+81-19-621-6391
kazume.nishidate at gmail.com, nisidate at iwate-u.ac.jp
https://sites.google.com/site/nisidatelab/