[CP2K-user] Issue with FULL_SINGLE_INVERSE

Nicholas Winner nwi... at berkeley.edu
Fri Aug 21 16:28:07 UTC 2020


Hello all,

I've had an issue come up that is strange that I would like to bring to the 
community's attention. On very rare occasions, the FULL_SINGLE_INVERSE 
preconditioner gets stuck, and cp2k will hang forever (example of the exact 
part where it freezes below). This has only happened a few times to me, so 
I cannot ascertain what the root cause is, but when it does happen, 
switching to FULL_ALL fixes it. Are the developers aware of this issue, and 
does anyone have an idea about how to fix it?

Also, I have tried verying the PRECOND_SOLVER when this problem occurs, and 
it does not help.

Preconditioner : FULL_SINGLE_INVERSE       : diagonalization, state 
selective
Precond_solver : DEFAULT
stepsize : 0.15000000 energy_gap : 0.01000000
eps_taylor : 0.10000E-15 max_taylor : 4
----------------------------------- OT 
---------------------------------------

Step Update method Time Convergence Total energy Change
------------------------------------------------------------------------------

Best,
Nick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200821/59515796/attachment.htm>


More information about the CP2K-user mailing list