Problem CELL_OPT

[Matthias Krack]

Hi Carlo

did you try BFGS as optimizer with TYPE DIRECT_CELL_OPT instead of CG, 
since your system is not too large?

Best regards

Matthias

On Tuesday, 13 June 2017 09:36:02 UTC+2, c.pignedoli wrote:
>
> Dear all I have a problem with CELL_OPT
> I am using cp2k 5.0 svn 17619.
>
> If I do 
>
> TYPE DIRECT_CELL_OPT
>
> i have in the output that Max grad is always 0.001
> but the atoms are not moving and the job runs forever.
>
>
> If I do 
> TYPE GEO_OPT
>
> in the several .out files of the geometry optimization steps
> I see that Max gradient converges to my requested threshold 0.0001
>
> but in cp2k.out I have Max Gradinet 0.001 and the job never stops.
>
> Also, I would like to avoid having a .restart and a .out and a .wfn saved 
> for each step of the CELL_OPT
> but what I specified in the cp2k.inp seems not effective.
>
> Can you help me?
>
> Kind regards
>
> Carlo
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170613/9256f9c0/attachment.htm>