[CP2K-user] Large energy difference between md step and single point calculation

[Massimo Bocus]

Dear CP2K users/developers,

I am extracting some snapshots from an umbrella sampling simulation (with 
PLUMED as dependency of CP2K 5.1) and I would like to recalculate their 
energy at a higher level of theory. So to start I performed a single point 
calculation at the same level of theory of the US simulation (PBE-D3) and, 
surprisingly for me, I got very different results in the energy. In 
particular, the energy of the MD step is:
ENERGY| Total FORCE_EVAL ( QS ) energy (a.u.):            
-3520.987951048081868
While for the single point calculation:
ENERGY| Total FORCE_EVAL ( QS ) energy (a.u.):            
-3520.903314334958850
Which is an enormous difference. Attached the two input scripts, the main 
differences being the RUN_TIPE, the cell parameters (for the single point 
they are just taken from the md output at the selected step) and the lack 
of the MOTION section.
Does anybody have an idea on where this large energy difference could come 
from?

Thank you in advance, 
Massimo Bocus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200922/08918002/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: md.inp
Type: chemical/x-gamess-input
Size: 3647 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200922/08918002/attachment.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: single_point.inp
Type: chemical/x-gamess-input
Size: 2261 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20200922/08918002/attachment-0001.inp>