[DFTB-Plus-User] opinions about atomic cluster input file

[Zulkarnain Zulkipli]

Dear DFTB+ user,


I`m new in using DFTB+.

regarding to atomic cluster calculation by using geometry optimization, 


i try to calculate ground state energy of Boron atoms by using DFTB+, 

i used the parameter in input file which is randomly refer from difference reasonable input file sources,
the problem is, when i calculate small ( < 6 ) number of the atomic cluster, the result obtained in a short time (  2-3 minutes ),
but, when i calculate for bigger ( > 10 ) number of atomic cluster, the time taken to obtain result is much more longer ( 1-3 hours ),

i`m waiting for any opinions and suggestions to improve my input file parameter that can produces faster and accurate result. 


* my input file as follow:
--------------------------------------------------------------------------------------------------------------------
Geometry = GenFormat {
3  C
B
1   1    1.33879136    1.12680398    0.29618134
2   1    2.70291917    0.99759966    0.51653906
3   1    1.52016049    1.48686303    0.59399900
}
Driver = ConjugateGradient {
MovedAtoms = Range { 1 3 }
MaxForceComponent = 1.0E-4 
  MaxSteps = 6000 
  OutputPrefix ="geom.out"
}

Hamiltonian = DFTB {
  SCC = Yes 
  SCCTolerance = 1.0E-5  
  MaxSCCIterations = 1000 

  Mixer = Broyden { 
    MixingParameter = 0.2     
  }

  MaxAngularMomentum = { 
    B = "p" 
  }

  Charge = 0.0

  SpinPolarisation = {}

  Eigensolver = Standard {}

  Filling = Fermi {
    Temperature [Kelvin] = 0.0
  }

  SlaterKosterFiles = {
    B-B = "B-B.skf"
  }
  ReadInitialCharges = No
}

Options = {
  WriteAutotestTag = Yes
}

------------------------------------------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20120809/00fcdd35/attachment.htm>