[Pw_forum] problems with integrated charge

[Paolo Giannozzi]

On Oct 25, 2006, at 18:00 , Liping YU wrote:

>  So I think this is due to parallel algorithm in the code not the  
> bad data.

so please try if such problem is solved in the cvs version. It should  
be,
since the algorithm used in parallel subspace diagonalization has been
rewritten

> To let program continue running without this error message, I  
> changed the
> line in PW/electron.90 " ABS( charge - nelec ) / charge > *1.0D-3*  
> " to
> " ABS( charge - nelec ) / charge > *1.d0* ". When scf calculation is
> close to convergence, such WARNING message will disappear.
> I don't know if this is a good way to do.

it may work: if the error on the charge is due to some insufficent
convergence in the subspace diagonalization, it may disappear
(but it might also NOT disappear) with the progress of self-consistency.
In any event, an error of 1 electron is something to be worried about

P.