[CP2K-user] [CP2K:21118] for hard convergence of geo_opt

[xuan Garrett]

Dear cp2k,

I am no using CP2K-v2025.1 to perform the GEO_OPT for my system (FeN4). I 
am now using DIAG algorithm with 3*3*1 MONKHORST PACK grid, along with 
FERMI-DIRAC smearing method. But I find that the SCF is very hard to 
converge each ionic step of 130, even with DZVP-MOLOPT-SR-GTH basis set. 

Previous tests I have found that the DIAG is always harder to converge than 
OT, but OT cannot support k-points, so I have to switch to DIAG with 
k-points. And, after the GEO_OPT with DZVP-MOLOPT-SR-GTH, if I want to 
calculate with TZVP-MOLOPT-GTH for highly accurate energy, hard convergence 
again. 

Attached is my .inp file and the unfinished .out file. I want to know if 
some other good solutions. Many thanks~

I learn that if the MOLOPT basis set is not very suitable for slab of 
surface calculations? Because I find that the same system is very easy to 
converge in VASP. 

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/9ca211f5-c29f-4883-988d-da5e9b592e80n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250206/6aab6d69/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: FeN4_geoopt.inp
Type: chemical/x-gamess-input
Size: 7636 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250206/6aab6d69/attachment-0001.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: FeN4_geoopt.out
Type: application/octet-stream
Size: 1396226 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250206/6aab6d69/attachment-0001.obj>