[CP2K-user] Moving from semi-empirical, to DFT for Metadynamics run while keeping bias

[Bach Nguyen]

I have a pretty huge QM region in my QM/MM run and it is prohibitive to do 
the whole metadynamics calculation using DFT. 

What I would like to do is sample the region using semi-empirical first and 
use the bias potential as the initial point for my DFT QM/MM. The keyword I 
am using is RESTART_METADYNAMICS. Are there any pitfalls I ought to concern 
myself with this?

I also see the SPAWNED_HILLS_X functions, which I understand as to used to 
add extra hills on top of what's already there. 
So the TOTAL_BIAS == RESTART_METADYNAMICS + SPAWNED_HILLS_X

Is this interpretation correct? 

Many thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210428/a59c17f0/attachment.htm>