[Pw_forum] mixing_beta and mixing_ndim

[Stefano de Gironcoli]

I think there is no general strategy.
you can try increasing  mixing_ndim and/or reducing mixing_beta.
Another possibility is to change mixing_mode from "plane" (the default) 
to TF or local-TF.
This should help in the case the instability is due to charge sloshing 
(large metallic systems)...
A source of instability can be that from one iteration to the next the 
list of occupied level
changes... increasing number of bands, number of k-points and/or 
smearing width may help.
If your system use gradient corrected functionals and has a lot of 
vacuum there may be some
problem related to spurious negative charge that can be  possibly solved 
by increasing ecutrho....
It is difficult to give a general advise...
Stefano de Gironcoli,

> Dear PWSCF users,
>       I have a question about 'mixing_beta' and 'mixing_ndim'.  In one 
> 'scf' calculations, I found it converges very slowly.  The 'estimated 
> scf accuracy'  remains almost a constant in the two subsequent 
> iterations.  Do you think whether it will help if I change 
> 'mixing_beta' or 'mixing_ndim', making them larger or smaller?
>       Thank you.
>     
>
> Hanghui Chen
> Department of Physics, Yale University
> ------------------------------------------------------------------------
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>