[CP2K-user] Possible bug in electrostatics in FIST.

[Matt Ellis]

Hi,

I'm getting some strange results when using electrostatics in a CP2K 
simulation. I have taken 128 pentacene molecules and created a small 
crystal. I've created 1 directory with all the necessary files to run FIST 
on this for 1 step with Ewald electrostatics. However, in my psf file I've 
set all charges to be 0 -so there should be no electrostatic contribution. 
I then copied this directory and in the newly formed directory set the 
EWALD_TYPE to NONE -the rest of the input file is exactly the same. Now, I 
get different forces and energies being calculated. Why is this -am I 
misunderstanding how the electrostatics/nonbonded forces are being 
calculated? This seems like a fairly major bug if I aren't.

I've attached the input directories I've been using. I've split the large 
input file into multiple files to make them easier to manage! I've tested 
with CP2K versions 4.1 and 6.1.


After looking into this a bit more thoroughly, I found that in the file 
'fist_nonbond_force.F' the Van der Waals forces are being calculated 
differently due to differences in the splines. I don't understand why this 
would be the case if all parameters are the same.
Best,

Matt

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210511/aa232d13/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: InputDirectories.zip
Type: application/x-zip
Size: 1110669 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210511/aa232d13/attachment.bin>