POL_SCF and Quadrupoles

Johannes Wagner joh... at gmail.com
Wed Sep 24 16:27:30 CEST 2014


Hi everyone,

I am trying to implement quadrupole damping and anion deformations for 
classical force field calculations and right now I am just testing what is 
already working.

While doing so I realised that as soon as I set the MAX_MULTIPOLE_EXPANSION 
to Quadrupole in &POISSON &EWALD 
and define Quadrupole polraizabilities (CPOL),  the conjugate_gradient 
method for the polarization cycle doesn't seem to work anymore.

Putting POL_SCF to self_consistent reproduces the forces of the reference, 
conjugate_gradient however is far of (see datasets attached).
However reducing to Dipole expansion and and deleting cpol terms reproduces 
the respective reference just fine, no matter the algorithm (as it should 
be, I guess?)
What could be the matter here?

I am using the latest svn checkout.

Attached the input files, the two sample outputs (forces, one using 
conjugate_gradient, one using self consistent), and the reference forces.


The only section I am changing at all is

   &EWALD 
        ...
     &MULTIPOLES T
         MAX_MULTIPOLE_EXPANSION QUADRUPOLE
         POL_SCF SELF_CONSISTENT
         EPS_POL 1.0e-6
         MAX_IPOL_ITER 100
       &END MULTIPOLES
     &END EWALD


Thanks very much for every hint and help! 
Johannes

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140924/4c051805/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: force_cp2k_con_grad.dat
Type: application/octet-stream
Size: 23400 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140924/4c051805/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: force_cp2k_self_cons.dat
Type: application/octet-stream
Size: 23400 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140924/4c051805/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: force_reference.dat
Type: application/octet-stream
Size: 26280 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140924/4c051805/attachment-0002.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: aim.inp
Type: chemical/x-gamess-input
Size: 3556 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140924/4c051805/attachment.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: coord.xy
Type: application/octet-stream
Size: 20647 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140924/4c051805/attachment-0003.obj>


More information about the CP2K-user mailing list