[Wien] Convergence of magnetic moment

[Laurence Marks]

You have to be careful about semantics (language) here. In the user guide
-cc and -ec are referred to as "Charge Convergence" and "Energy
Convergence", but they could better be written as "Charge Consistency" and
"Energy Consistency". The scf iteration will continue until neither
changes, i.e. they are both consistent from one iteration to the next.

If you want to converge the charge and energy (or magnetic moment), then
you need to look at the technical parameters in particular the number of
k-points as well as the oversampling in case.in0. You may also need to
increase RKMAX if you want a converged absolute energy. (Of course accuracy
is another thing as has been mentioned on this list before, and depends
upon the functional.)

Realistically, you have zero magnetic moment in your calculations. Be
humane to electrons: don't worry about the last couple of digits!

(N.B., Peter just emailed a similar response.)

On Wed, May 23, 2018 at 8:40 AM, Rui Costa <ruicosta.r15 at gmail.com> wrote:

> Dear wien2k users,
>
> Can I converge simulations to the last decimal place in the magnetic
> moments? Or do the numerical errors stop me from achieving such accuracy?
>
> For example, I have made a calculation with -cc 0.000 001 but i get:
>
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00013
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00014
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00015
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00017
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00018
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00019
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00021
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00024
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00026
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00028
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00030
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00032
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00034
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00035
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00039
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00059
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00094
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00108
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00106
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00040
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00010
> :MMTOT:  SPIN MAGNETIC MOMENT IN CELL     =   -0.00014
>
> From the evolution I would say that it is converged to the second last
> but I really be sure that the magnetization is converged to the last
> decimal place. So my next step would be to run the simulation with -cc
> 0.000 000 1 until I achieve a smooth convergence.
>
> Additionally, is it possible to display more decimal places in the .scf
> file?
>
> Best regards,
> Rui Costa.
>
>
>


-- 
Professor Laurence Marks
"Research is to see what everybody else has seen, and to think what nobody
else has thought", Albert Szent-Gyorgi
www.numis.northwestern.edu ; Corrosion in 4D: MURI4D.numis.northwestern.edu
Partner of the CFW 100% program for gender equity, www.cfw.org/100-percent
Co-Editor, Acta Cryst A
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20180523/ce755568/attachment.html>