RATLLE vs SHAKE

Frank Uhlig uhlig... at googlemail.com
Mon Jun 28 13:27:32 UTC 2010


Dear CP2k users,

I am a little bit confused by the output of constrained molecular
dynamics as carried out by CP2K/Quickstep.
As far as my knowledge goes, the constraining of collective variables
via the SHAKE or RATTLE algorithm correspond to different methods of
integrating Newton's equation of motion (Verlet and Velocity-Verlet
respectively). And as far as I noted, you are using a Velocity-Verlet
scheme by default (not seen anything else in the manual or code).
It would be nice if someone could explain further details about this,
otherwise I would be tempted to use RATTLE (as for the combination with
the Velocity-Verlet).

Thanks in advance and best regards,

Frank Uhlig






More information about the CP2K-user mailing list