The order of Wannier orbital indices in the output files.

[JQ]

Dear CP2K developers,

Could you help to confirm that the order of Wannier centers output are not 
always consistent for large sized system? Thank you.

On Friday, 20 December 2013 16:24:24 UTC+2, JQ wrote:
>
> Dear CP2K users,
>
> I have a moderate sized system which I am interested to print out the 
> Wannier orbitals at a particular site. For example, the Wannier orbital 
> positioned between carbon and phosphorus atoms.
>
> First, I started with an energetic calculation and print out Wannier 
> centers. After getting the corresponding centers' indices, I can determine 
> the interested sites and I will restart the calculation with exactly the 
> same input files and parameters, except that asking CP2K to print out the 
> selected cube files for the selected sites. However, I notice the indices 
> of Wannier centers will not always be consistent as compared to the first 
> calculation and I can't always get the orbitals I am interested in.
>
> I guess one possibility could be due to the parallel calculation on a 
> computer cluster. The collection of wannier information may not be arranged 
> in the same order due to the parallel processing. Or could this be the case 
> related to the algorithm calculating the Wannier orbitals (which can't be 
> unique)?
>
> Please find the attached files for your information (system size = 91) and 
> the output for your reference. Input parameters like SCF and etc are for 
> testing purpose only. Thanks.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140109/28d28228/attachment.htm>