[CP2K:6995] Mixing Method Recommendation

[Chris Campbell]

Hello S Ling,

Thanks for the quick reply.

I have inserted this new scf, I had gotten rid of direct diagonalization 
already but my convergece still takes around 117 steps.

Regarding the &BS section I was wondering why two columns were required 
(please see 
https://github.com/misteliy/cp2k/blob/master/tests/QS/regtest-bs/cu2cl6_m3_clp6cud9.inp)?
I would expect just a number per keyword, i.e., NEL +1 and NEL -1 for Cu1 
and Cu2, respectively. Then L should be 0 in both cases and N should be 4 
(unpaired 4s electron in [Ar] 3d10 4s1 configuration).

Thanks for the help and previous replies to other posts which have been 
invaluable,
Chris


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150915/e21a8901/attachment.htm>