[Wien] is there any ralashion between ghostband and number of k-points?

[arash fathi]

Dear Wien users,
I am working on an uranim based compound using
wien2k_05.2 which is runing under redhat linux 9
(intel compiler 8.1+ mkl 7.0).
I would find the best number of k-pints to optimize my
calculation. Thus I let be changed the number of
k-points from 100 to 10000 with a reasonable but
variable step.
I have found that except for the number of k-pints
100, k-pints=100, all my calculations have been
converged. For the former case, k-point=100, I got
surprisingly a ghostband message after :IET=5 as
follows:
 'l2main' - QTL-B.GT.7., Ghostbands, check scf files
I checked then :NECxx and found that no more electrons
are missing:
:NEC01: NUCLEAR AND ELECTRONIC CHARGE  388.00000
387.99819   1.00000

My question is this that why I have such a ghostband
problem only for k-points=100 (I have not yet checked
for less than this number of k-points) and e.g. not
for k-points=600, 3000, 6000 and so on.

Best,
A. Fathi.



		
__________________________________ 
Do you Yahoo!? 
Yahoo! Small Business - Try our new resources site!
http://smallbusiness.yahoo.com/resources/