[gmx-users] RE: System instability with switched cut-off
Berk Hess
gmx3 at hotmail.com
Fri Aug 7 16:06:16 CEST 2009
Please believe me, switched Coulomb potential are very bad!
I understand you want to have the energies correct,
but the first thing you want is the forces to be correct.
The forces matter for your sampling.
If you use a switch function you will always have artificially
large forces in the switching region. Switching between 1.4 and 1.7 nm
switches over 0.3 nm. This produces much larger forces than the plain
Coulomb ones around say 0.5 nm.
For RF I would always use epsilon_rf=infinity (0 in the mdp file).
This makes sure the potential and force are zero at the cut-off.
Because RF also applies to exclusions, energies are the same as with PME
up to many digits in systems without ionic groups.
Another option would be running with PME and using mdrun -rerun with
a long Coulomb cut-off.
Berk
> From: matteus.lindgren at chem.umu.se
> To: gmx-users at gromacs.org
> Date: Fri, 7 Aug 2009 15:51:47 +0200
> Subject: [gmx-users] RE: System instability with switched cut-off
>
> Thanks for the answer. PME can't be used since the interaction coulomb LR
> value between two groups can't be extracted for PME. I know switched cutoffs
> produce artefacts around the cutoff value but in this case I am more
> interested in the correct interaction energies than the correct forces which
> is why a shifted function can't be used. My plan was that with the rather
> large rcoulomb and rlist values I chose, a switched function would be ok.
>
> Would the reaction field method suit my purposes? How do I choose
> epsilon_rf? I have pure water and 10M urea as solvents. Still I would like
> to figure out why switch creates these enormous fluctuations in my system.
> Can you see any errors in the parameters I've listed?
>
> Matteus
>
>
> -----Original Message-----
> From: gmx-users-bounces at gromacs.org [mailto:gmx-users-bounces at gromacs.org]
> On Behalf Of gmx-users-request at gromacs.org
> Sent: den 7 augusti 2009 14:51
> To: gmx-users at gromacs.org
> Subject: gmx-users Digest, Vol 64, Issue 39
>
> Send gmx-users mailing list submissions to
> gmx-users at gromacs.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> or, via email, send a message with subject or body 'help' to
> gmx-users-request at gromacs.org
>
> You can reach the person managing the list at
> gmx-users-owner at gromacs.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of gmx-users digest..."
>
>
> Today's Topics:
>
> 1. RE: System instability with switched cut-off (Berk Hess)
> 2. Re: Energy Minimisation and Equilibration Problems (Bruce D. Ray)
> 3. Re: question about mdrun -append (Baofu Qiao)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 7 Aug 2009 13:46:31 +0200
> From: Berk Hess <gmx3 at hotmail.com>
> Subject: RE: [gmx-users] System instability with switched cut-off
> To: Discussion list for GROMACS users <gmx-users at gromacs.org>
> Message-ID: <COL113-W42EEC13FE2DE71D0118E668E0B0 at phx.gbl>
> Content-Type: text/plain; charset="iso-8859-1"
>
>
> Hi,
>
> Never ever use switched cut-off's!
> Switched cut-off's can give extremely strong artifacts.
> In 4.1 grompp will warn you about this.
>
> Switch gives you correct energies at short range, but incorrect forces
> in the switching region.
> What you want is correct forces up to (nearly) the cut-off.
> Shift does this.
> But reaction-field does this smoother and it even has a physical
> interpretation.
> So use RF if you can not use PME.
>
> But why can't you use PME, you have pbc, right?
>
> Berk
>
>
> From: matteus.lindgren at chem.umu.se
> To: gmx-users at gromacs.org
> Date: Fri, 7 Aug 2009 13:13:24 +0200
> Subject: [gmx-users] System instability with switched cut-off
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> Dear all
>
>
>
> Im trying to extract the interaction
> energies between protein and solvent. Since PME cant be used for this, plain
> cutoffs causes problems and shifted cut-offs gives incorrect energies I want
> to
> use switched cutoff but with rather long cutoffs to get the correct energy.
>
>
>
> Unfortunately my system is unstable and
> within a few ps the total energy and kinetic energy fluctuates several
> hundred
> per cent. I4ve tried different options but can4t get it to stabilize except
> if
> I4m using any other coulomb type. Any ideas?
>
>
>
> title =
> Dyna
>
> cpp =
> /usr/bin/cpp
>
> constraints =
> hbonds
>
> integrator =
> md
>
> dt =
> 0.002
>
> nsteps =
> 20000000
>
> nstxout =
> 20000000
>
> nstvout =
> 20000000
>
> nstfout =
> 5000
>
> nstlog =
> 250
>
> nstenergy =
> 500
>
> nstxtcout =
> 500
>
> xtc_grps =
>
> energygrps =
> polar hydrophobic peptide SOL
>
> nstlist = 1
>
> ns_type =
> grid
>
> coulombtype =
> switch
>
> rcoulomb_switch =
> 1.4
>
> rcoulomb =
> 1.7
>
> rlist =
> 2.0
>
> vdwtype =
> switch
>
> rvdw_switch =
> 1.4
>
> rvdw =
> 1.7
>
> tcoupl =
> nose-hoover
>
> tc-grps = Protein
> Non-Protein
>
> tau_t =
> 0.5 0.5
>
> ref_t =
> 298 298
>
> Pcoupl =
> parrinello-rahman
>
> tau_p =
> 1.0
>
> compressibility =
> 4.5e-5
>
> ref_p =
> 1.0
>
>
>
> ;disre_fc =
> 30
>
>
>
> gen_vel = yes
>
> gen_temp = 298
>
> gen_seed = 173529
>
> unconstrained_start
> = no
>
>
>
> Thanks
>
> Matteus
>
>
>
> ---------------------------------------------------------
>
> Matteus
> Lindgren, graduate student
>
> Department of Chemistry, Umee University
>
> SE-901 87 Umee, Sweden
>
> Phone: +46 (0)90-7865368
>
>
>
>
>
>
> _________________________________________________________________
> See all the ways you can stay connected to friends and family
> http://www.microsoft.com/windows/windowslive/default.aspx
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> http://lists.gromacs.org/pipermail/gmx-users/attachments/20090807/6c0d63c8/a
> ttachment-0001.html
>
> ------------------------------
>
> Message: 2
> Date: Fri, 7 Aug 2009 05:06:02 -0700 (PDT)
> From: "Bruce D. Ray" <brucedray at yahoo.com>
> Subject: Re: [gmx-users] Energy Minimisation and Equilibration
> Problems
> To: gmx-users at gromacs.org
> Message-ID: <664136.39384.qm at web35807.mail.mud.yahoo.com>
> Content-Type: text/plain; charset="us-ascii"
>
> On Thursday, August 6, 2009 at 7:59:43 PM Nancy <nancy5villa at gmail.com>
> wrote:
>
>
> > I ran the minimisation, and mdrun gave the following last few lines of
> output:
> >
> > ==================================
> > Step=19990, Dmax= 7.8e-06 nm, Epot= -7.14922e+04 Fmax= 3.19585e+04, atom=
> 2395
> > Step=19992, Dmax= 4.7e-06 nm, Epot= -7.14924e+04 Fmax= 8.72456e+03, atom=
> 2395
> >
> Step=19993, Dmax= 5.6e-06 nm, Epot= -7.14927e+04 Fmax= 3.70693e+04, atom=
> 2395
> > Step=19994, Dmax= 6.8e-06 nm, Epot= -7.14929e+04 Fmax= 2.10513e+04, atom=
> 2395
> > Step=19995, Dmax= 8.1e-06 nm, Epot= -7.14929e+04 Fmax= 4.71234e+04, atom=
> 2395
> >
> Step=19996, Dmax= 9.7e-06 nm, Epot= -7.14931e+04 Fmax= 3.68827e+04, atom=
> 2395
> > Step=19998, Dmax= 5.8e-06 nm, Epot= -7.14933e+04 Fmax= 1.35461e+04, atom=
> 2395
> > Step=19999, Dmax= 7.0e-06 nm, Epot= -7.14933e+04 Fmax= 4.77130e+04, atom=
> 2395
> >
> Step=20000, Dmax= 8.4e-06 nm, Epot= -7.14936e+04 Fmax= 2.41860e+04, atom=
> 2395
> >
> > writing lowest energy coordinates.
> >
> > Steepest Descents did not converge to Fmax < 10 in 20001 steps.
> > Potential Energy = -7.1493609e+04
> >
> Maximum force = 2.4185994e+04 on atom 2395
> > Norm of force = 8.1511212e+02
> > ==================================
> >
> > As you can be seen, the forces still do not converge to Fmax < 10, even
> after 20,000 steps.
> >
> > Does anyone know what the problem might be?
>
> The problem is my error in the specification of non-bonded interaction
> combination rules,
> FudgeLJ, and FudgeQQ for gromacs force fields.
>
> Line 7 of your file ffethanediol.itp currently reads:
> 1 2 yes 0.5 0.833333
>
> This was taken from ambconv and is correct for amber and gaff force fields,
> but not
> for gromacs force fields, nor for OPLS-AA.
>
> Change line 7 of the file ffethanediol.itp to read:
> 1 1 no 1.0 1.0
>
> The correction is in topolbuild 1.3 which is not yet released, but this
> error was not caught
> for the topolbuild 1.2 series. Quite frankly, I had forgotten that
> modification was made as
> I've been editing the set of tables needed to add OPLS-AA to the mix of
> force fields available.
>
> As a general note. For gromacs force field topologies generated with
> anything in the
> topolbuild 1.2 series, this edit to the itp file that specifies the defaults
> is required.
>
>
> Sincerely,
>
>
> --
> Bruce D. Ray, Ph.D.
> Associate Scientist
> IUPUI
> Physics Dept.
> 402 N. Blackford St.
> Indianapolis, IN 46202-3273
>
>
>
>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> http://lists.gromacs.org/pipermail/gmx-users/attachments/20090807/e002422d/a
> ttachment-0001.html
>
> ------------------------------
>
> Message: 3
> Date: Fri, 07 Aug 2009 14:49:45 +0200
> From: Baofu Qiao <qiaobf at gmail.com>
> Subject: Re: [gmx-users] question about mdrun -append
> To: Discussion list for GROMACS users <gmx-users at gromacs.org>
> Message-ID: <4A7C22E9.2080501 at gmail.com>
> Content-Type: text/plain; charset="us-ascii"
>
> An HTML attachment was scrubbed...
> URL:
> http://lists.gromacs.org/pipermail/gmx-users/attachments/20090807/ee691be4/a
> ttachment.html
>
> ------------------------------
>
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
>
> End of gmx-users Digest, Vol 64, Issue 39
> *****************************************
>
>
> _______________________________________________
> gmx-users mailing list gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
_________________________________________________________________
See all the ways you can stay connected to friends and family
http://www.microsoft.com/windows/windowslive/default.aspx
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20090807/08638ca5/attachment.html>
More information about the gromacs.org_gmx-users
mailing list