[CP2K-user] convergence issue with diagonalisation

[Travis]

Hi,

Your cell has 12 atoms, the CIF should only define 6 for the unit cell. So 
you doubled the number of atoms in only one direction but you gave a cell 
size for a 2x2x2 supercell. Your calculation is slow to converge because 
~3/4 of the volume of your cell is vacuum and this has to collapse before 
you can get a meaningful structure. 

-T

On Wednesday, May 6, 2020 at 7:38:11 AM UTC-4, Hepsibahpriyadarshini C 
wrote:
>
> Hi,
> I tried out optimising the geometry of Ag2O using OT I obtained 
> convergence since I require the usage of K-points for my study I had to 
> switch over to diagonalisation. When I started with 6 atoms I was able to 
> obtain convergence with 4X4X4 K-points. When I started to increase the 
> no.of.atoms that is 12 I am not able to obtain convergence. Inspite of 
> using BFGS as the optimizer, changing EPS_SCF values and playing with 
> mixing parameters I am not able to obtain convergence. Should ELPA be 
> compiled for this. Can band structure,PDOS calculations and phonon 
> calculations be performed with CP2K-5.1 version
> I have attached my input file for your kind reference. 
>
>
>
>
>
> Regards
> Hepsibah.C
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200506/a7087f9d/attachment.htm>