[Pw_forum] Why does Ewald sum calculation need so much RAM

[Paolo Giannozzi]

On Tue, 2012-07-17 at 13:52 +0200, Thomas Gruber wrote:

> I am doing phonon calculations with PAW pseudopotentials for systems 
> with metallic character and wondering/have problems with the small RAM 
> from the cluster. Example 1:  A system with 36 atoms (fcc) a have done 2 
> scf calculation with an k-grid of 6x6x6 with occupation="tetrahedra" and 
> "smearing" (generatet k points=12). For the phonon dispersion the 
> symmetry get reduced and the

and the ... ?
-- 
Paolo Giannozzi, IOM-Democritos and University of Udine, Italy