[Pw_forum] takes time but converges

[Paolo Giannozzi]

Regassa Dhasaa wrote:

> WARNING ...not compatable with FFT or something and shows some matrix. 
> Does this affect the converged result which we found ~2hr later.

no

> The other question is about the 'relax' calculation for FCC (CaMnO3) it 
> starts well by showing some six atomic position relaxations at the 
> output but finnally it says "charge is wrong" and stops. I am starting 
> working with the mixing beta = 0.5. What else can I try?

gaussian broadening instead of fixed occupations

Paolo
-- 
Paolo Giannozzi, Democritos and University of Udine, Italy