[DFTB-Plus-User] LBFGS Minimizer

[Carlo Nervi]

Hello,
I'm not a dftb+ developer, but I can report my experience.
I get the same error whenever the geometry is almost optimized, so that
LBFGS has difficulties in proceeding.
Reducing the MaxForceComponent value may help, or sometimes it is only
necessary to repeat the same identical calculation with no changes.
Do not forget to do 'cp geo_end.gen DFTB.gen' before to restart (i.e. use
the last structure for the new input).
Hope this helps,
Carlo


Il giorno mar 15 set 2020 alle ore 11:26 toufik esssakhri <tousak at hotmail.fr>
ha scritto:

> Dear DFTB users,
> I am using LBFGS Minimizer for geometry relaxation of a finite system, the calculation was not able to reach convergence and I got the following error
>
> -> Error: x need to be different in each iteration. (LBFGS Minimizer).
> But the  charges.bin, detailed.out, eigenvec.bin … files are written!
> I tried also ConjugateGradient but the geometry didnt converge.
>
> DFTB+ 20.1 version and Ob2-1-1 skf files
>
> Any idea about what is going wrong?
>
> Geometry = xyzFormat {
>   <<< c24h36.xyz
> }
> Driver = LBFGS {
>   Memory = 15
>   LatticeOpt = Yes
>   MovedAtoms = 1:-1
>   MaxForceComponent = 1.0e-4
>   MaxSteps = 1000
>   OutputPrefix = "geom.out"
>   ConvergentForcesOnly = Yes
>   Constraints = {}
> }
>
> Hamiltonian = DFTB {
>   Scc = Yes
>   SccTolerance = 1e-3
>   MaxSCCIterations = 1000
>   Dispersion = LennardJones {
>     Parameters = UFFParameters {}
>   }
>   Mixer = Broyden {
>     MixingParameter = 0.3
>     InverseJacobiWeight = 1e-2
>     MinimalWeight = 1
>     MaximalWeight = 10000
>     WeightFactor = 1e-2
>   }
>   SlaterKosterFiles = Type2FileNames {
>   Prefix = "/home/dftb20.1/ob2-1-1/base/"
>   Separator = "-"
>   Suffix = ".skf"
>   }
>   MaxAngularMomentum {
>     C = "p"
>     H = "s"
>   }
>   Filling = Fermi {
>     Temperature [Kelvin] = 100
>   }
> }
> Am I missing something?
>
> Thank you in advance.
>
> Best regards,
> Taoufik
>
>
> _______________________________________________
> DFTB-Plus-User mailing list
> DFTB-Plus-User at mailman.zfn.uni-bremen.de
> https://mailman.zfn.uni-bremen.de/cgi-bin/mailman/listinfo/dftb-plus-user
>


-- 

------------------------------------------------------------
Prof. Carlo Nervi carlo.nervi at unito.it  Tel:+39 0116707507/8
Fax: +39 0116707855      -      Dipartimento di Chimica, via
P. Giuria 7, 10125 Torino, Italy.    http://lem.ch.unito.it/

ICCC 2020 5-10 July 2020, Rimini, Italy: http://www.iccc2020.com
International Conference on Coordination Chemistry (ICCC 2020)

 <http://www.iccc2020.com/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20200915/a8744368/attachment.htm>