[Pw_forum] calculation = "nscf" with CP

[Paolo Giannozzi]

On Friday 23 December 2005 00:09, Young-Su Lee wrote:

> ver 2.0 ==========================================
>                 total energy =       -8.07495 a.u.
> [...]
> ver 3.0 ==========================================
>                 total energy =       -8.07495 Hartree a.u.

I don't really see a problem here. The error on the total energy is a
quadratic function of the error on the charge density. This is not true 
for the single pieces composing the total energy, so the latter will 
be more sensitive than the total energy to lack of self-consistency. 
The average potential, in particular, is very sensitive to errors because
gradient-corrected functionals have a nasty numerical behavior for 
\rho=>0

Paolo

-- 
Paolo Giannozzi             e-mail:  giannozz at nest.sns.it
Scuola Normale Superiore    Phone:   +39/050-509876, Fax:-563513 
Piazza dei Cavalieri 7      I-56126 Pisa, Italy