[CP2K:4719] dbcsr error when applying periodic efield

[Florian Schiffmann]

Hi David,

when I looked in the code the problem was that ROKS has only a single set 
of MO derivatives but nspins=2. I can see that it makes sense for ROKS and 
from a quick look it should work the same way for the MO derivatives from 
the electric field. Nevertheless, I am not quite sure about it. Furthermore 
there is a second problem: The berry phase approximation for the electric 
field is only implemented for the DFT part. If you run QMMM, the MM part 
won't feel the same effect. It might be possible to fix that as well but to 
make sure both will work properly requires a bit more work (maybe a week of 
my time as the MM part of CP2K is not one of my major points of interest). 
Currently I am fairly busy so I can't tell you when I will have time to 
look into that.

Flo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20131027/98e09068/attachment.htm>