[gmx-users] Exploding temp/pressure.

[Mark Abraham]

Hi,

P-R coupling is only stable if the system is already near equilibrium, so I
would always vary other stuff (like step size) while using NVT or Berendsen
P coupling (or both in successive stages).

Although your observations are consistent with an initial structure that
has reached a reasonable local minimum, a possibility is that the
configuration is still frustrated somehow and that it is easier for a
larger time step to get over a saddle point towards a deeper minimum. But
then that time step could be too large for the resulting forces (which
would certainly be made even worse by P-R). Expending effort into
constructing a more reasonable original starting structure might pay off.
Simulated annealing at the small step size would be another possibility to
consider.

Mark

On Sun, Jan 1, 2017 at 5:02 PM Nash, Anthony <a.nash at ucl.ac.uk> wrote:

> Hi all,
>
> I¹m trying to equilibrate a Martini CG simulation from an initial
> atomistic structure. Eq and Fc values were derived using an atomistic
> system. I¹ve started the dt at 0.0005 for 600000 steps, moving through
> 0.001, 0.0015 and 0.002 for the same number of steps, using the .mdp
> details below (with the exception of the time step and the output
> delimiters - the remaining settings are from the Martini website). During
> these steps the temperature dropped from approx 60000K down to 310K and
> the potential energy dropped from positive figures down to around
> -1.4*10^6. A final 600000 steps with dt = 0.002 showed a converged temp,
> pressure, volume, potential energy (and total energy), and very little
> change in the unit dimensions.
>
> I now want to move the time step up to 0.025. I¹ve tried various dt values
> (0.005, 0.01, 0.015Š 0.025) using the below setup, but each time dt causes
> almost immediately structural explosion. The temperature and the pressure
> rockets up. I changed the pressure coupling to berendsen and the same
> happens. I have been through dozens of tau_p and tau_t combinations, each
> one yields a trajectory of varying length before exploding (the protein
> Œexplodes¹ at the exact time the the pressure and temperature increases).
> I have also changed the thermostat to berendsen and Nose-Hoover, using a
> range of 5*(dt*nsttcouple) and 20(dt*nsttcouple), respectively.
>
>
> I would appreciate any thoughts or suggestions on how to solve my
> temperature/pressure problem when I try increasing the dt to a CG value.
>
> Thanks
> Anthony
>
>
> title                    = Martini
> integrator               = md
> dt                       = 0.025
> nsteps                   = 300000
> nstcomm                  = 100
>
>
> nstxout                  = 10000
> nstvout                  = 10000
> nstfout                  = 0
> nstlog                   = 1
> nstenergy                = 1
> nstxout-compressed       = 0
> compressed-x-precision   = 0
> ;compressed-x-grps        =
> energygrps               = collagen solvent
> cutoff-scheme            = Verlet
> nstlist                  = 20
> ns_type                  = grid
> pbc                      = xyz
> verlet-buffer-tolerance  = 0.005
> coulombtype              = PME ;reaction-field
> rcoulomb                 = 1.1
> fourierspacing       = 0.16 ;0.2  ;0.12
> epsilon_r                = 15   ; 2.5 (with polarizable water)
> epsilon_rf               = 0
> vdw_type                 = cutoff
> vdw-modifier             = Potential-shift-verlet
> rvdw                     = 1.1
>
> tcoupl                   = v-rescale ;berendsen ;v-rescale
> tc-grps                  = collagen solvent
> tau_t                    = 0.4 0.4 ;2.0 2.0
> ref_t                    = 310 310
>
> Pcoupl                   = parrinello-rahman
> Pcoupltype               = semiisotropic
> tau_p                    = 12.0  ;parrinello-rahman is more stable with
> larger tau-p, DdJ, 20130422
> compressibility          = 4.5e-5 0
> ref_p                    = 1.0 1.0
> refcoord_scaling         = com
>
> gen_vel                  = no
> gen_temp                 = 310
> gen_seed                 = 473529
> continuation = yes
> constraints              = none
> constraint_algorithm     = lincs
> lincs-warnangle = 45
> lincs-order=18
> lincs-iter=4
>
>
>
> --
> 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.
>