[gmx-users] Re: Potential energy from a previous configuration mismatch

[Justin Lemkul]

On 11/29/12 10:40 AM, escajarro wrote:
> This is my mdp file for energy minimization:
>
> constraints              = none
>
> ; integration parameters
> integrator               = l-bfgs
> dt                          = 0.00
> nsteps                    = 15000
>
> ; center of mass removal
> nstcomm                  = 10
> comm_mode             = Linear
>
> ; neighbour searching
> nstlist                     = 1
> ns_type                  = grid
> pbc                        = xyz
> rlist                        = 1.4
>
> ; electrostatics
> coulombtype              = Reaction-Field-zero
> rcoulomb                   = 1.1
> epsilon_r                   = 1.0
> epsilon_rf                  = 0.0
>
> ; van der Waals interactions
> vdwtype                  = cut-off
> rvdw                       = 1.4
> DispCorr                  = EnerPres
>
> ; Ewald parameters
> fourierspacing          = 0.12
> fourier_nx               = 0
> fourier_ny               = 0
> fourier_nz               = 0
> pme_order              = 4
> ewald_rtol               = 1e-6
> optimize_fft             = yes
>
> ; energy minimization
> emtol                    = 0.1
> emstep                  = 0.01
>
>
>
> And this is the one for MD
>
> constraints              = none
>
> ; integration parameters
> integrator               = md
> tinit                       = 0
> dt                          = 0.001
> nsteps                    = 10000000
> ;init_step                = 0
> comm_mode            = linear
> nstcomm                 = 10
>
> ; output control
> nstxout                 = 10000
> nstvout                 = 10000
> nstfout                  = 0
> nstlog                   = 10000
> energygrps            = hexane-UA ethanol
>
> ; neigbour searching
> nstlist                   = 1
> ns_type                = grid
> pbc                      = xyz
> rlist                      = 1.4
>
> ; electrostatics
> coulombtype              = Reaction-Field-zero
> rcoulomb                   = 1.1
> epsilon_r                   = 1.0
> epsilon_rf                  = 0.0
>
> ; van der Waals interactions
> vdwtype                  = cut-off
> rvdw                       = 1.4
> DispCorr                  = EnerPres
>
> ; Ewald parameters
> fourierspacing           = 0.12
> fourier_nx                = 0
> fourier_ny                = 0
> fourier_nz                = 0
> pme_order               = 4
> ewald_rtol                = 1e-6
> optimize_fft              = yes
>
> ; Generate velocities
> gen_vel                   = yes
> gen_temp                = 298.0
> gen_seed                 = -1
>
>
> To use as the starting configuration the last configuration of the
> minimization, I used the confout.gro file of the minimization as the input
> conf.gro for MD. I also tried using the traj.trr file generated after
> minimization with the grompp option -t for MD. I used the same top and ndx
> files in both simulations.
>

The .mdp files look reasonable enough, though setting nstlist = 1 during MD will 
cost you a lot of performance.  A value in the 5-10 range should suffice.

Do other ensembles work, like NVT?  Does a smaller timestep affect stability?

-Justin

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

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

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