lagrangian multipliers

[Axel]

On Jan 16, 12:14 pm, Teodoro Laino <teodor... at gmail.com> wrote:
> Hi,
>
> in the example you posted you've 1 constraint and in the lagrangian
> multipliers file you will find one number for SHAKE and
> one for RATTLE... maybe I didn't understand your question but to me
> looks ok..

hi teo,

i guess the problem is, that when one wants to do umbrella sampling,
one wants the constraint force "for the individual step" and if i
understand
this correctly, this means the lagrange multiplier on the position
update,
i.e. the one from SHAKE and _not_ the one from RATTLE (the velocity
update).
at least, this is what i found when comparing CP2K to CPMD, where only
one "constraint force" is written and it is the LM from SHAKE.

i hope this is not too much "dumbed down" to be incorrect.

cheers,
    axel.
>
> Regarding the thermostat.. The region MASSIVE does not work with
> constraints.
> CP2K should print a clear message about that. Don't you get the error
> message?
>
> Teo
>
> On 16 Jan 2008, at 17:33, xiaoliang wrote:
>
> > Actually the distance between two centers of mass are fixed. In the
> > *.LagrangeMultLog file, Rattle and Shake lagrange multiplier are both
> > printed out.
> > like:
>
> > Shake  Lagrangian Multipliers:             2.798220972
> > Rattle Lagrangian Multipliers:            -2.890137362
>
> > Other question is that when I run constrained MD, it seems that the
> > REGION in THERMOSTAT section can not be set to MASSIVE.