[gmx-users] Conserved Energy in csvr NVT runs

[Mark Abraham]

On Apr 16, 2014 2:04 PM, "janos.daru" <janos.daru at gmail.com> wrote:
>
> Dear All,
>
> I run simple md job (GMX version:4.6.5 double precision) for
> alanine-dipeptid with careful options.
> nstlist                  = 1
> ns_type               = simple
> pbc                     = no
> rlist                     = 0
> epsilon_r                = 1
> coulombtype          = cut-off
> rcoulomb                 = 0
> vdwtype                  = cut-off
> rvdw                     = 0
> dispcorr                 = no
> constraints              = h-bonds
> constraint-algorithm     = Lincs
> lincs-order              = 4
> lincs-iter               = 2
>
> For NVE runs I get very nice conserved energy values: ~10E-5
> K/numberOfDegreesOfFreedom/ps
>
> But when I change to v-rescale with the following settings, I get clear
> drift(~1 K/numberOfDegreesOfFreedom/ps).
>
> tcoupl                   = v-rescale
> tc-grps                  = Protein
> tau_t                    = 0.05
> ref_t                    = 300
> nsttcouple               = 1
>
> As total energy and temperature makes a proper oscillation I have the
> filling, that "Conserved Energy" is not calculated as I expected, or I
miss
> something important form the input.

IIRC there is a conserved quantity written, but as you can see in Bussi's
original paper, it has a t dependence. This has been discussed on this list
before, if you search. tau-t probably affects the slope.

> An other observation is that in NVE runs I need to use comm_mode=none to
get
> nice energy conservation, while in NVT I get slightly better results with
> comm_mode=angular for longer runs (aprox 2.5 ns).

You haven't said, but I assume you have a dipeptide in vacuum, which I
think should need angular if the action of the thermostat can add kinetic
energy to the COM dof.

> Could some experienced user comment on these observations and give me
> advice, how to proceed to get proper NVT ensamble without artifacts of the
> drift?

You probably did. What is your time step?

> I didn't find in the manual options for global and local thermostats. Can
I
> use local (each degrees of freedom) thermostat in gromacs?

Not really. Best you can do is a temperature coupling group per atom.

Mark

> Thank you in advance,
>
> Janos
>
>
>
>
>
> --
> View this message in context:
http://gromacs.5086.x6.nabble.com/Conserved-Energy-in-csvr-NVT-runs-tp5015881.html
> Sent from the GROMACS Users Forum mailing list archive at Nabble.com.
> --
> Gromacs Users mailing list
>
> * Please search the archive at
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
send a mail to gmx-users-request at gromacs.org.