The ABINIT Users Mailing List ( CLOSED )

[zfhou@mail.edu.cn]

Thanks very much for all of your helps!

I recognize that I encounter this warning because I did't fully understand 
the input variable of 'rfatpol' before. Now I get it.

Anthoer problem is that in the d/dk calculation the maximum residual is very 
larger than tolwfr (=1.0d-18), and it is not converged within 100 steps. How 
to make it be converged? 

Regarding that the band gap of HfO2 is about 5.8 eV, I set the diemac = 3.0. 
In the GS calculation, it is converged with 30 steps. According to the manual 
of abinit, I used the tolwfr= 1.0d-18 for RF calculation. 

In other electronic structure codes, there are some mixing parameters for 
charge density mixer (such as Kerker mixing and Broyden mixing), which is 
related to the convergence of SCF calculation. As we know, in abinit code the 
input variables related to SCF convergence are 'dielng', 'diemac' and 
'diemix'. What scheme is used in abint? Are there some references related to 
the mixing scheme of abinit?

Your helps will be greatly appreciated!
Regards!
zhufeng