[gmx-users] EM did not converge crashing drude system

[Justin Lemkul]

On 3/15/17 1:19 PM, Dayhoff, Guy wrote:
> I have a drude system that I have minimized, and equilibrated through nvt
> and npt ensembles using position restraints (all with scf). As I perform the
> production runs i’m facing random crashes that seem to be related to EM
> did not converge messages. This has happened anywhere from tens of
> thousands of steps in, to millions of steps in. Below is an example and
> the mdp options for the run.
>

Please contact me off-list about this.  These are not officially supported 
features and there are possibly bugs but I will need better descriptions of what 
you're doing and all your input files.

-Justin

>
> My Best,
>    Guy Dayhoff
>
> -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
>
> step 393698: EM did not converge in 20 iterations, RMS force 15.668
> step 393699: EM did not converge in 20 iterations, RMS force 88.569
> step 393700: EM did not converge in 20 iterations, RMS force 341.002
> step 393701: EM did not converge in 20 iterations, RMS force 680.888
> step 393702: EM did not converge in 20 iterations, RMS force 1147.229
> step 393703: EM did not converge in 20 iterations, RMS force 1077.486
>
>
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
> MDP options:
>
> ; RUN CONTROL
> integrator               = md
> nsteps                   = 50000000 ;50ns
> dt                       = 0.001
>
> ; OUTPUT CONTROL
> nstxout 		 = 0
> nstvout			 = 10000 ;every 10ps for now
> nstfout			 = 10000
> nstlog			 = 10000
> nstenergy		 = 10000
> nstcalcenergy 		 = 1
> nstxout-compressed  	 = 10000
> compressed-x-grps   	 = System
>
> ; NEIGHBOR SEARCHING
> cutoff-scheme            = verlet
> nstlist                  = 10
> ns-type                  = Grid
> pbc                      = xyz
> rlist                    = 1.2
>
> ; BONDS
> constraints		 = h-bonds
> continuation             = yes
>
> ; ELECTROSTATICS AND VdW
> coulombtype              = PME
> rcoulomb                 = 1.2
> vdwtype                  = cutoff
> vdw-modifier		 = potential-switch
> rvdw-switch		 = 1.0
> rvdw                     = 1.2
> DispCorr                 = EnerPres
>
> ; EWALD
> pme_order                = 4
> fourierspacing		 = 0.16
>
> ; TEMPERATURE COUPLING
> Tcoupl                   = Nose-Hoover
> tc-grps                  = Water non-Water
> tau_t                    = 10.0 10.0
> ref_t                    = 300 300
> nh-chain-length          = 1
>
> ; VELOCITY GENERATION
> gen-vel                  = no
>
> ; PRESSURE COUPLING
> Pcoupl                   = Parrinello-Rahman
> Pcoupltype               = semiisotropic
> tau_p                    = 50.0
> compressibility          = 4.5e-5 4.5e-5
> ref_p                    = 0.0 3.0
>
> ; DRUDE POLARIZATION
> drude                   = yes
> drude-mode              = SCF
> drude-hyper             = yes
> drude-khyp              = 16736000.0
> drude-r                 = 0.02
> drude-pow               = 4
>

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================