[CP2K-user] Convergence of OT with transition metals

[Nicholas Winner]

Hello all, 

I have been working on some calculations for transition metal oxides and 
have found the convergence to be very poor. Apparently this is a well known 
issue, but most recommendations for how to deal with this are for 
diagonalization based methods, and I would like to take advantage of OT if 
possible.

Are there agreed upon settings to help converge systems with TMs? I can 
simply try every preconditioner, vary the band gap guess, try different 
minimizers, etc... but this brute force approach may take a long time.

Best,
Nick Winner
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200318/b7501ef6/attachment.htm>