[Pw_forum] new bfgs: strange behavior doing vc-relax

[Stefano Baroni]

On Apr 20, 2011, at 10:21 PM, Stefano de Gironcoli wrote:

>>> With QE I have always seen the energy to decrease when the cutoffs are
>>> increased, but with VASP, I always see an oscillation in the total
>>> energy
>>> when the cutoff is incremented.
>> just guessing: it might be a consequence of real-space treatment of
>> augmentation
>> functions.
> 
> actually it can happen for US pseudo, if the cutoff for the density is 
> not large enough, that the energy increases with increasing cutoff.. 
> just because the augmentation charges are not integrated correctly and 
> the error can have any sign. If ecutrho is sufficient to integrate 
> correctly augmentation charges the energy should be variational w.r.t. 
> ecutwfc .

Just a short, half-philosphoc comment, in the hope to make the whole issue even more clear. DFT is variational with respect to the size of the 1-electron basis set (irrespective of the fact that ionic potentials are local or not), if the numerics of the calculation is otherwise well converged with respect to any possible parameter of the calculation. When it is not, no guarantee of variational bounds. This is obvious in a sense. Stealthy problems may arise when some of the auxiliary numerical parameters of the calculations are surreptitiously correlated with the size of the basis set (such as, e.g. the number of plane waves in the Fourier expansion of the [augmented] charge density or the number of mesh points used for the real-space integration of different functions).

Hope this halps
Stefano

---
Stefano Baroni - SISSA  &  DEMOCRITOS National Simulation Center - Trieste
http://stefano.baroni.me [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype)

La morale est une logique de l'action comme la logique est une morale de la pensée - Jean Piaget

Please, if possible, don't  send me MS Word or PowerPoint attachments
Why? See:  http://www.gnu.org/philosophy/no-word-attachments.html



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20110421/b1dd24cd/attachment.htm