Why is Diagonalization so much slower than OT?

[Christopher O'Brien]

To all,
Am am attempting to perform MD on a system with ~900 atoms. Using OT, I can 
run hundreds of time steps in a day. Using Diagonalization, I can only run 
4. I have appended my diagonalization section below. Are there any 
recommended settings for the Mixing parameters? Are there any 
recommendations for getting an initially converged WFN before starting MD?

The reason I don't want to use OT is that it prevents orbital occupation 
from changing. I am trying to use METADYNAMICS to study a reaction, so I 
don't want to constrain multiplicity. 

    &SCF

 

      EPS_SCF 1.0E-6 

      SCF_GUESS atomic 

      MAX_SCF 100 


      ADDED_MOS 100 

      &SMEAR .TRUE. 

        METHOD FERMI_DIRAC 

        ELECTRONIC_TEMPERATURE [K] 300. 

      &END SMEAR 

      &DIAGONALIZATION  ON 

        ALGORITHM DAVIDSON 

        &DAVIDSON 

          PRECONDITIONER FULL_SINGLE_INVERSE 

        &END DAVIDSON 

      &END DIAGONALIZATION 

      &MIXING  T 

        METHOD BROYDEN_MIXING 

        ALPHA 0.4 

        NBROYDEN 8 

      &END MIXING 


     &END SCF



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150918/590b9a28/attachment.htm>