[gmx-users] Fwd: NVT to NVE ensemble, energy drifting problem
Mark Abraham
mark.j.abraham at gmail.com
Fri Nov 6 11:41:31 CET 2015
Hi,
On Fri, Nov 6, 2015 at 10:32 AM gozde ergin <gozdeeergin at gmail.com> wrote:
> Hi Mark,
>
> As your suggestion I used cutoff sheme = Verlet with running in Gromacs 5.1
> Now instead of decreasing, total energy kept increasing! Do you think this
> behavior is normal?
>
Some drift is normal, but you need to think about it on a
per-degree-of-freedom-per-unit-time basis (which you'd know if you'd read
the manual sections and inspected the graphs I already suggested?). You
also need to inspect all your output (e.g. grompp also), respond
appropriately to the diagnostics that may have been offered, know what you
should expect from the short-range scheme documentation, and e.g. use
constraints suitably per .mdp documentation. Nor should you be expecting a
run that generates velocities to be immediately equilibrated and to show
equilibrium properties like energy conservation! One reasonable strategy is
to pre-equilibrate with NVT, and then do an exact restart in NVE.
Mark
>
> Energy Average Err.Est. RMSD Tot-Drift
>
> -------------------------------------------------------------------------------
> Kinetic En. 8537.49 180 386.122 1253.25
> (kJ/mol)
> Potential -20741.9 530 1092.62 3652.89
> (kJ/mol)
> Total Energy -12204.4 710 1449.39 4906.14
> (kJ/mol)
> Temperature 319.933 6.7 14.4695 46.9643 (K)
>
>
> Thanks in advance.
>
>
> On Thu, Nov 5, 2015 at 2:15 PM, Mark Abraham <mark.j.abraham at gmail.com>
> wrote:
>
> > Hi,
> >
> > So that means you are using the group cutoff scheme without a cutoff
> > buffer, which is known not to conserve energy (see manual sections on
> > neighbour searching). You can either choose a group-scheme model physics
> > where the buffer (set with rlist) is suitably larger than the cutoffs
> > (which is awkward to set up and will run slowly), or use the Verlet
> scheme
> > (which has a buffer by default and runs fast) which is now default in
> 5.1.
> > Even with that fixed, the use of constraints intrinsically leads to
> drift,
> > also. Whether the remaining drift matters for your simulation (vs the
> > decrease in sampling you would have to accept in order to start to deal
> > with it) is an open research question.
> >
> > Mark
> >
> >
> > On Thu, Nov 5, 2015 at 1:23 PM gozde ergin <gozdeeergin at gmail.com>
> wrote:
> >
> > > - system 512 water molecule covered by 25 organic on each surface,
> > > - I used both gromacs 4.5 and 5.0 with and without double precision
> > > - .mdp file is :
> > >
> > > integrator = md
> > > dt = 0.002
> > > nsteps = 10000000 ; 20 ns
> > > nstxout = 10000
> > > nstvout = 10000
> > > nstlog = 10000
> > > nstenergy = 10000
> > > continuation = no
> > > constraint_algorithm = lincs
> > > constraints = all-bonds
> > > lincs_iter = 1
> > > lincs_order = 4
> > > ns_type = grid
> > > nstlist = 5
> > > rlist = 1.2
> > > rcoulomb = 1.2
> > > rvdw = 1.2
> > > coulombtype = PME
> > > pme_order = 4
> > > fourierspacing = 0.16
> > > tcoupl = nose-hoover
> > > tc-grps = System
> > > tau_t = 1.0
> > > ref_t = 300
> > > Pcoupl = No
> > > pbc = xyz
> > > DispCorr = EnerPres
> > > gen_vel = yes
> > > gen_temp = 310
> > > gen_seed = -1
> > > nstcomm = 10
> > > comm-mode = Linear
> > > comm-grps = System
> > >
> > > hope it is readable now!
> > >
> > > best
> > >
> > > On Thu, Nov 5, 2015 at 11:44 AM, Mark Abraham <
> mark.j.abraham at gmail.com>
> > > wrote:
> > >
> > > > Hi,
> > > >
> > > > You're making it very hard to help you, by not providing a readable
> > .mdp
> > > > file, not specifying a GROMACS version, nor describing what is in
> your
> > > > system. If there was a magic "do this and energy is conserved" button
> > > then
> > > > it'd be already pushed for you ;-)
> > > >
> > > > Mark
> > > >
> > > > On Thu, Nov 5, 2015 at 11:34 AM gozde ergin <gozdeeergin at gmail.com>
> > > wrote:
> > > >
> > > > > Hi Ganesh
> > > > >
> > > > > I tried your suggestion and this time instead of decreasing,
> > potential
> > > > and
> > > > > kinetic energy kept increasing!
> > > > > So still could not get the converged-constant energy for NVE.
> > > > >
> > > > >
> > > > > On Wed, Nov 4, 2015 at 4:23 PM, Ganesh Shahane <
> > > ganesh7shahane at gmail.com
> > > > >
> > > > > wrote:
> > > > >
> > > > > > Hi Gozde,
> > > > > >
> > > > > > You could try modifying and introduce some of the parameters as
> > > follows
> > > > > > (keeping the rest same) -
> > > > > >
> > > > > > lincs_iter = 2
> > > > > > rvdw-switch = 1.1
> > > > > > vdwtype = Cut-off
> > > > > > vdw-modifier = Force-switch
> > > > > >
> > > > > > I recently faced some of the same problems as yours and the above
> > > > > > parameters worked for me.
> > > > > >
> > > > > > On Wed, Nov 4, 2015 at 2:47 PM, gozde ergin <
> gozdeeergin at gmail.com
> > >
> > > > > wrote:
> > > > > >
> > > > > > > Dear Chaban,
> > > > > > >
> > > > > > > I decreased the time step and also I run by using double
> > precision
> > > > > > however
> > > > > > > energy kept decreasing and did not converge.
> > > > > > >
> > > > > > > Does anybody has an opinion how to run NVE simulation in
> Gromacs
> > > > > without
> > > > > > > energy drifting?
> > > > > > >
> > > > > > > Bests
> > > > > > >
> > > > > > >
> > > > > > > On Tue, Nov 3, 2015 at 7:27 PM, Vitaly V. Chaban <
> > > vvchaban at gmail.com
> > > > >
> > > > > > > wrote:
> > > > > > >
> > > > > > > > Decrease the time-step.
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > On Tue, Nov 3, 2015 at 3:00 PM, gozde ergin <
> > > gozdeeergin at gmail.com
> > > > >
> > > > > > > wrote:
> > > > > > > >
> > > > > > > > > ---------- Forwarded message ----------
> > > > > > > > > From: gozde ergin <gozdeeergin at gmail.com>
> > > > > > > > > Date: Tue, Nov 3, 2015 at 5:39 PM
> > > > > > > > > Subject: NVT to NVE ensemble, energy drifting problem
> > > > > > > > > To: mailing list Gromacs <
> > > > > gromacs.org_gmx-users at maillist.sys.kth.se>
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > Dear Gromacs user,
> > > > > > > > >
> > > > > > > > > I did some simulation in NVT ensemble now I need to switch
> > and
> > > do
> > > > > > some
> > > > > > > > more
> > > > > > > > > simulations in NVE ensemble.
> > > > > > > > > What properties do I need to change in .mdp file to get a
> > good
> > > > > energy
> > > > > > > > > conversion?
> > > > > > > > > I used the same properties like NVT without temperature
> > > coupling
> > > > > but
> > > > > > my
> > > > > > > > > potential and kinetic energy drifted a lot.
> > > > > > > > > What should I do prevent this drifting?
> > > > > > > > >
> > > > > > > > > Here is my nvt.mdp:
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > *integrator = md dt
> =
> > > > > > > > > 0.002nsteps = 10000000 ; 20 nsnstxout
> > > > > > > > > = 10000nstvout = 10000nstlog
> > > =
> > > > > > > > > 10000nstenergy = 10000continuation =
> > no
> > > > > > > > > constraint_algorithm = lincs constraints =
> > all-bonds
> > > > > > > > > lincs_iter = 1 lincs_order = 4
> > > > > > ns_type
> > > > > > > > > = grid nstlist = 5 rlist
> > =
> > > > > > > > > 1.2 rcoulomb = 1.2 rvdw
> > =
> > > > > > > > > 1.2 coulombtype = PME pme_order
> > =
> > > > > > > > > 4 fourierspacing = 0.16 tcoupl
> > > > > > =
> > > > > > > > > nose-hoovertc-grps = Systemtau_t
> > > > > > =
> > > > > > > > > 1.0ref_t = 300Pcoupl =
> > > Nopbc
> > > > > > =
> > > > > > > > > xyz DispCorr = EnerPres gen_vel
> > > =
> > > > > > > > > yes gen_temp = 310 gen_seed
> =
> > > > > > > > > -1 nstcomm = 10comm-mode =
> > > > > > > > > Linearcomm-grps = System *
> > > > > > > > > For nve, I just make the temperature coupling off and used
> > the
> > > > same
> > > > > > > > > properties however as I mentioned my energy could not
> > converged
> > > > in
> > > > > 20
> > > > > > > ns.
> > > > > > > > >
> > > > > > > > > Thanks in advance
> > > > > > > > > --
> > > > > > > > > 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.
> > > > > > > > >
> > > > > > > > --
> > > > > > > > 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.
> > > > > > > >
> > > > > > > --
> > > > > > > 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.
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Best Regards,
> > > > > > Ganesh Shahane
> > > > > > --
> > > > > > 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.
> > > > > >
> > > > > --
> > > > > 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.
> > > > >
> > > > --
> > > > 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.
> > > >
> > > --
> > > 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.
> > >
> > --
> > 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.
> >
> --
> 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.
>
More information about the gromacs.org_gmx-users
mailing list