[Wien] large energy oscillation

[Mr C.A. Bridges]

Dear Stefaan and Steve,
 Thank you very much for your replies. The :warn label is given below.

:WARN : Iteration  1 occurs more than once in this scf file!!!
        ETOT is not calculated correctly !!!!!
        You should save your calculations using save_lapw and continue

It seems that midway through the SCF cycle the calculation just restarts
at iteration 1! The same warning is given 13 times for each initial
iteration. I wonder if it could have restarted due to NFS conjestion?

Craig




On Wed, 13 Aug 2003, Stefaan Cottenier wrote:

>
>>  I have just completed a calculation with Wien2K which converged to a
>> reasonable result. I am happy with the result, but I noticed that the
>> energy jumped up to extremely large values partway through the
>> calculation, as shown below. My question is, should I worry about the
>> result if this kind of jump in energy occurs, or is this a perfectly
>> acceptable, normal occurence?
>
>1) grep the :WARN label, which will tell you something more (probably bad
>integration)
>
>2) look at some properties (:MMI, :HFF, :EFG, :RTO, ...) just before and just
>after the jump. Are they more or less similar (as far as you can tell, because
>these are non-selfconsistent values)? If they are, then you can consider these
>high energies as an 'accident de parcours'. If they are not, then you might
>have jumped into a qualitatively different solution (local minimum of the Born-
>Oppenheimer surface) (your slightly larger energy after the jump compared to
>before, suggest this). In the latter case, start your calculation again from
>the original situation, with 'something' changed (for instance GAUSS instead
>of TETRA), hope you pass the problematic region smoothly this time, and change
>back to the original parameters (TETRA again) when you are near convergence,
>and converge further (it will jump here for a while, but that's normal).
>
>Stefaan
>
>_______________________________________________
>Wien mailing list
>Wien at zeus.theochem.tuwien.ac.at
>http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>