The ABINIT Users Mailing List ( CLOSED )

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

Another comment:

In PAW, the use of prtdos=3 is actually not recommended...
because only the "pseudo" wave functions are (for the moment) used to 
generate the projected dos, not the "true" wave functions; the "on-site" 
terms of the wave functions are not taken into account...
And the pseudization is not norm-conserving ; the density you project 
does not have the right norm !  (you forget the compensation charge).

Of course, you do not have this problem using norm-conserving 
pseudopotential.

This is a feature to be implemented ; in a future version the code will 
print a warning.

You can use prtdos=1 or 2, because, in these cases, only the eigenvalues 
are used to generate the DOS (and they are the right one in PAW).


Marc

Marc Torrent a écrit :
> Dear Nuno,
>
> I've just opened your file.
> And I have a few comments:
>
> If you want to compare norm-conserving and PAW runs, you have to 
> explicitely put the mixing scheme:
> - potential is mixed by defautl in norm-conserving with a conjugate 
> gradient algorithm (iscf=5)
> - density is mixed by default in PAW with an Anderson algorithm (iscf=14)
> Put iscf=14 in both cases.
> Doing this test could help to determine wether the pb is related to 
> mixing or PAW...
>
> I just wanted to
> > > 4-unit 62130756389765768 should be 9 !  But I can see the reason why 
> > > it is not (this is a parameter defined inside a module).
>
> I just wanted to say that the unit number should not be 
> tmp_unit=62130756389765768 but explicitely tmp_unit=9 and that it is 
> defined in a very basic module (called by every routine of the code).
>
>
> Marc Torrent
> CEA-Bruyeres-le-Chatel - France