[CP2K:1663] PM6 - MTD - Matrix diagonalization failed

Teodoro Laino teodor... at gmail.com
Fri Dec 19 11:17:11 UTC 2008


On the top of the above message just a couple of remarks on your input file:

>   &DFT
>      &POISSON
>        POISSON_SOLVER WAVELET
>        PERIODIC NONE
>      &END POISSON
>   
SE does not work with GRIDS. so WAVELET as decoupler is useless.. It is 
a bit more tricky to control..
but so far, by default SE is essentially only non-periodic.
>     &QS
>       METHOD PM6
>       &SE
>         ANALYTICAL_GRADIENTS F
>         STO_NG 6
>       &END SE
>     &END QS
>     &SCF
>       SCF_GUESS MOPAC
>       EPS_SCF 1.0E-04
>   
10^-4 is extremely larger threshold for convergence.. you may want to 
use something smaller (1.0E-6- E-7)...
Whatever you do with a WF like the one that you get out of EPS_SCF 
1.0E-4 is  essentially trash.
>       &PRINT
>         &RESTART OFF
>         &END RESTART
>       &END PRINT
>     &END SCF
>   &END DFT
>   



More information about the CP2K-user mailing list