Iterative first variational solver

[andrei]

Having tried to use iterative solver (tseqit=true) I found that in many cases, including examples, I could not reach the convergence. For most examples that I tried, even total charge convergence could not be achieved. For some, like GaAs, this could be rectified by first doing several SCL iterations using direct solver, and then finishing with the iterative solver. For the other, like Fe, this did not help and iterative solver never converged. I tried varying nseqit parameter, but this did not help. Is there anything else which I  may be missing?

Thanks!
andrei

[exciting]

Hi Andrei,

Thanks for the message. The iterative solver is still in the experimental stage, and seems to work well for some systems and not for others: why this is I don't yet know. It doesn't work at all when the second-variational step is included.

I'll keep working on it, but if you have any ideas then let me know...

Cheers,
Kay.