[gmx-users] energy decreasing in NVE simulation

[Erik Marklund]

For more exotic NVE-systems I had to do some or several of the following 
things to get stable Etot:
* have an even shorter timestep than one would expect from the applied 
constraints and such.
* use double precision.
* apply the constraints with lower tolerance/more iterations etc.

Then there's a few things I notice in your setup:
* I see that you do not use constrints at all. I would give it a shot.
* Why do you have two separate comm-grps?

Erik Marklund

Yun-an Yan skrev:
>
> Dear all, 
>
> When I try to simulation with NVE ensemble, the 
> total energy keeps decreasing. Geometry optimization 
> and solvent equilibration are done before the NVE simulation. 
>
> I follow the requirements provided in 
> http://www.gromacs.org/Documentation/Terminology/NVE
>
> Would any one help me to figure out what was wrong with
> my settings?
>
> Millions of thanks and best wishes,
> Yun-an
>
>
> BTW: Please check the following for the .mdp file
>
> ;Run control: A leap-frog algorithm for integrating Newton's equations.
> integrator               = md
> ;time step in femtoseconds
> dt                       = 0.001
> ;number of steps
> nsteps                   = 10000000
> ;No constraints
> constraints              = none
> constraint_algorithm     = shake
> shake_tol                = 1e-08
>
> ; group(s) for center of mass motion removal
> comm-grps                = AA CCl4
> ; frequency for center of mass motion removal
> nstcomm                  =  1000
> ;frequency to write coordinates to output trajectory file
> nstxout                  = 1000
> ;frequency to write velocities to output trajectory file
> nstvout                  = 1000
> ;frequency to write energies to log file
> nstlog                   = 1000
> ;frequency to write energies to energy file
> nstenergy                = 1000
> ;group(s) to write to energy file
> energygrps               = System
>
> ;Frequency to update the neighbor list (and the long-range forces,
> ;when using twin-range cut-off's).
> nstlist          = 10
> ns_type          = grid
> rlist            = 1.5
>
> ; Use periodic boundary conditions in all directions.
> pbc = xyz
>
> ;cut-off distance for the short-range neighbor list
> ;treatment of electrostatic interactions
> coulombtype      = PME-switch
> fourierspacing   = 0.12
> pme_order        = 4
> ewald_rtol       = 1e-5
> rcoulomb_switch  = 1.0
> rcoulomb         = 1.2
>
> ;treatment of van der waals interactions
> vdwtype          = Switch
> rvdw-switch      = 1.0
> rvdw             = 1.2
> dispcorr         = EnerPres
>
> ;Temperature coupling
> tcoupl          = no
>
> ;Pressure coupling
> pcoupl           = no
>
> ;Velocity generation
> gen_vel          = no
>
>


-- 
-----------------------------------------------
Erik Marklund, PhD student
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,    75124 Uppsala, Sweden
phone:    +46 18 471 4537        fax: +46 18 511 755
erikm at xray.bmc.uu.se    http://folding.bmc.uu.se/