[CP2K-user] [CP2K:20512] Re: revPBE0 CELL_OPT not converging

[Frederick Stein]

Dear Alberto,
- Did you check the convergence of EPS_SCHWARZ and EPS_SCHWARZ_FORCES (not 
set by you, check the manual)? With ADMM, it needs to be chosen tighter 
than without ADMM. You should also check the convergence of nuclear forces 
and the stress tensor.
- Why do you calculate the stress tensor numerically? This is usually not 
too accurate.
- The cell might be too small. Beware that HF energies with cutoff radii 
far below 5 A are usually not well converged. Compare this with your setup 
in CASTEP (if I understand it correctly).
You may also pre-optimize cell size and geometry independently of each 
other before optimizing them simultaneously.
Best,
Frederick

alberto santonocito schrieb am Montag, 29. Juli 2024 um 10:48:18 UTC+2:

> Dear CP2K Users,
>
> I attempted a CELL_OPT calculation using the revPBE0 functional, but 
> unfortunately, the calculation is not converging, and after a few steps, 
> the structure deviates significantly from the experimental one.
> I chose to use revPBE0 with CP2K because, in CASTEP, I found a good 
> agreement between the experimental and calculated structures using the 
> revPBE functional. However, my computational resources are insufficient to 
> test revPBE0 in CASTEP.
> With CP2K, this is possible due to the Auxiliary Density Matrix Methods 
> that reduce computational costs. Unfortunately, the calculation with CP2K 
> does not converge.
> I would appreciate any suggestions regarding this issue.
> I have attached the CP2K input and output  files, along with the CASTEP 
> input and output.
>
> Thanks in advance for your help!
>

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/d75234d5-2f77-4612-bd34-ff6ebb0f166an%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240729/04e87548/attachment.htm>