The ABINIT Users Mailing List ( CLOSED )

[Marc Torrent <marc.torrent@cea.fr>]

Dear Bo liu,

The warning message seems to be explicit: you have to make your 
projectors smoother...
It can be achieve by changing the values of reference energies (here for 
s projectors). Too close reference energies produce high projectors !

Marc Torrent


iloveabinit@gmail.com a écrit :
> Dear All,
>
> I generate the Cl atom data by atomic.
> When I test the Cl2 calculation, the out file shows the following 
>
>  ==== Results concerning PAW augmentation regions ====
>
>  Total pseudopotential strength Dij (hartree):
>  Atom #  1
>    0.12274  -9.33465   0.00000   0.00000   0.00011   0.00000   0.00000   
> 0.00002
>   -9.33465 203.32921   0.00000   0.00000  -0.02806   0.00000   0.00000  
> -0.07413
>    0.00000   0.00000   0.35067   0.00000   0.00000   0.96443   0.00000   
> 0.00000
>    0.00000   0.00000   0.00000   0.35067   0.00000   0.00000   0.96443   
> 0.00000
>    0.00011  -0.02806   0.00000   0.00000   0.34914   0.00000   0.00000   
> 0.96013
>    0.00000   0.00000   0.96443   0.00000   0.00000   2.65904   0.00000   
> 0.00000
>    0.00000   0.00000   0.00000   0.96443   0.00000   0.00000   2.65904   
> 0.00000
>    0.00002  -0.07413   0.00000   0.00000   0.96013   0.00000   0.00000   
> 2.64694
>  print_ij: WARNING -
>   The matrix seems to have high value(s) !
>   (  1 components have a value greater than  100.0).
>   It can cause instabilities during SCF convergence.
>   Action: you should check your atomic dataset (psp file)
>           and look for "high" projector functions...
>
> How could we resolve this problem ?
>
> the input files for atompaw is 
> Cl  17 
> LDA-PW  loggrid 2001
> 3 3 0 0 0 0
> 3 1 5
> 0 0 0
> c
> c
> v
> c
> v
> 1
> 2.0
> y
> 0
> n
> y
> 0
> n
> Vanderbilt  
> 2 0
> 1
> 2.0
> 2.0
> 2.0
> 2.0
> 0
>
>
> Thank you in advance!
>
> Bo Liu