temperature monitor and thermostat regions

[Axel]

On Nov 29, 1:54 am, Teodoro Laino <teodor... at gmail.com> wrote:
> > not so far, but i can imaging, there is a bunch of trouble
> > lurking when using constraints, particularly, if they cross
> > thermostat region boundaries.
>
> eheh ;-)
> I disabled this kind of possibility.. cp2k should stop in all cases
> like these ones ;-)
> If not please point me the problem.. (it would be anyway a bugged run).

well, for quite a few (simple) cases, you can account for them, but
allowing
a fractional number of DOFs and then removing an equal share of a
constraint
from each atom involved. i think this is how it is done, e.g. in CPMD.
but anyway,
but before discussion irrelevant problems, let us first wait and see
whether there
actually is a  case where this would be useful. i'm certain your TODO
list
has many items, that are more relevant.


> >> We could discuss about the possibility to have different regions to
> >> different temperatures, but I'm quite
> >> skeptical whether it will really work!
>
> > i guess there are some people that would be interested in this,
> > but i would worry about this when they actually do ask.
>
> It's trivial to give in input several temperatures but nowhere is
> guaranteed that the regions are really decoupled. i.e. you try
> to thermostat to a certain   temperature but
> the real equilibrium temperature will not be the one in input.
> That's something physical.. that's why I'm skeptical.

true. temperature has no meaning in those cases. a few
applications that i see for multiple thermostat regions at different
temperatures are:
- keeping an active center "cold" during (pre-)equilibration without
  completely freezing the coordinates.
- enhancing phase space sampling by heating up a small subsystem.
in both cases you would not care about no longer being in a well
defined ensemble. but as noted before, lets worry about this
when it is actually needed.

> > what would be the cherry on of the cream on _my_ pie would
> > be that the MASSIVE flag (i.e. one thermostat per DOF) would
> > not be only a global option but a per region option. i assume
> > this would make everything a lot more complicated and it is
>
> I don't really understand that.. MASSIVE you have 1 thermostat per DOF..
> This is not a global option.

hmmm.... it is! i can specify REGION MASSIVE or REGION MOLECULE,
or REGION ..., or REGION DEFINED.

how do i enable a MASSIVE for a REGION DEFINED? as you know,
you currently (well, last time i checked, which is a long time ago in
terms of the pace of cp2k development) cannot use MASSIVE on
anything with a constraint. e.g. in a QM/MM scenario, it could be
advantageous to have a  MASSIVE thermostat for the QM region
and _single_ thermostat, for the rest which usually has rigid water,
ie
lots of constraints.

> Defining regions you have a thermostat per region (like a global one but
> restricted to the selected number of atoms).
> If you define a MASSIVE within each region you get exactly the same as
> using MASSIVE (since the temperature is everywhere the same).
> did I misunderstood something?

probably. i don't want _alL_ region to be MASSIVE only some
(e.g. those that are compatible with the "no constraints" condition).

> > thanks again for your efforts,
>
> I will accept payments ONLY with the "delirium" ones ;-)

i'm working on that. in fact, the next batch is currently bubbling so
loud, it woke me up,. i'll save you a couple of "bearhugger" and
"mad axel" and your delirium tremons is guaranteed (should i aim
for pink elephants with parachutes?).

if you have a favorate style that we don't have in our portfolio yet,
please let me know. i'll start a flemish sour on the weekend
but should have time to set up another batch before i leave for
christmas. due to too many travels, production has slowed down
significantly.

cheers,
    axel.

>
> teo.
>
>
>
> >> p.s.: I will add soon also the possibility to dump on file the
> >> temperatures of the different thermostats (tomorrow should be in)..
>
> Here we are.. I totally forgot about this point.. I will add this
> print_key during the week end.