The ABINIT Users Mailing List ( CLOSED )

["Frank Grosse" <frank.grosse@physik.hu-berlin.de>]

Hello Marc,
thank you very much. I recompiled version 5.6.3. with the replacement and it
works now.

I appreciate especially your fast response.

Frank


============================================
Dr. Frank Grosse
Humboldt-Universität zu Berlin
Institut für Physik
AG Festkörpertheorie
Newtonstraße 15
D-12489 Berlin
Tel. +49-30-20934993
email Frank.Grosse@physik.hu-berlin.de


-----Original Message-----
From: marc.torrent@cea.fr [mailto:marc.torrent@cea.fr] 
Sent: Montag, 1. Dezember 2008 17:14
To: forum@abinit.org
Subject: Re: [abinit-forum] PAW Tutorial pawprtdos

Hello Franck,

Thanks for having pointed out an error !
(introduced between 5.5.x and 5.6.x versions of Abinit).

If you want your ABINIT to work, you have to replace
~abinit/src/13paw/partial_dos_fractions_paw.F90 by the enclosed file (and
recompile)...

... or wait for the 5.6.4 version to be soon released.

Marc Torrent
CEA-Bruyeres-le-Chatel - France


Frank Grosse a écrit :
>  
> Hello,
> I encountered an error with the PAW tutorial when trying to evaluate 
> the PAW contributions to the DOS. As input I use the modified 
> tpaw1_4.in as described in the tutorial (attached) with the 
> 6c.lda.atompaw pseudopotential.
>
> With abinit version 5.6.3. I get the following error message:
>
> STOP Error in cprj_alloc: wrong sizes !
>
> and it does not produce the DOS output file. The last output in the 
> log is for
>
> ...
>  get dens in sphere for ikpt,isppol =  110 1 ...
>
> With the angular analysis for every band. Last band
>
>   Angular analysis
>  
>  Atom #   1 is   C, in-sphere charge =  0.241300
>  l=0, charge=0.006570, m=-l,l splitting: 0.007  l=1, charge=0.185938, 
> m=-l,l splitting: 0.000 0.090 0.095  l=2, charge=0.047265, m=-l,l 
> splitting: 0.015 0.014 0.017 0.000 0.001  l=3, charge=0.001077, m=-l,l 
> splitting: 0.000 0.000 0.000 0.000 0.000 0.000 0.000  l=4, 
> charge=0.000451, m=-l,l splitting: 0.000 0.000 0.000 0.000 0.000 0.000 
> 0.000 0.000 0.000
>
>  Sum of angular contributions for all atomic spheres  l=0, charge 
> =0.006570 proportion =0.027226  l=1, charge =0.185938 proportion 
> =0.770566  l=2, charge =0.047265 proportion =0.195875  l=3, charge 
> =0.001077 proportion =0.004465  l=4, charge =0.000451 proportion 
> =0.001868
>
>  Total over all atoms and l=0 to 4 :  0.241300
>  
>
>
> I tried the parallel version, and a serial version compiled with Intel 
> 10.1 as well as the precompiled GNU version from the abinit website. 
> No difference. Using version 5.4.4. with the identical input files, 
> everything is fine.
>
> I also tried 6c.lda.test-2proj.atompaw and 6c.lda.test-6proj.atompaw 
> with identical behavior.
>
>
> The initial scf steps work fine, with numerical differences between 
> version
> 5.4.4 and 5.6.3 though. 
>
> Could someone confirm that and maybe help resolve the issue?
> Thanks in advance.
>
> Regards,
>
> Frank
>
>
>
> ============================================
> Dr. Frank Grosse
> Humboldt-Universität zu Berlin
> Institut für Physik
> AG Festkörpertheorie
> Newtonstraße 15
> D-12489 Berlin
> Tel. +49-30-20934993
> email Frank.Grosse@physik.hu-berlin.de
>
>
>