[gmx-users] Errors faced in reproducing simulation in Gromacs 2018.1 that works fine in v. 4.6.x

[Neelima S.]

 Hello,
I am trying to simulate a graphene nano ribbon (GNR) (9.99nm x 5.57nm) in
water (spc216) using OPLS-AA forcefield. Successful simulations have been
carried out in v. 4.6.3 and 4.6.5 using the following settings:
Box size: 8x12x12
EM:
emtol = 100.0 ; emstep = 0.01 ; nsteps = 100000 ;
NVT:
nsteps = 10000 ;
dt     = 0.002 ;
NPT:
nsteps = 3000 ; dt = 0.002 ;
Final Run:
dt = 0.002 nsteps = 5000000

However, in V2018.1, under the same settings, the final MD simulation has
been showing either "X particles communicated to PME node Y are more than a
cell length out of the domain decomposition cell of their charge group" or
"A charge group moved too far between two domain decomposition steps"
error, suggesting that the system is not well equilibrated.

The energy minimization has been found to be okay by checking the minimized
potential for both V. 2018.1 and V. 4.6.5. Also the npt equilibrated
structure has no defect/overlapping atoms.
I have also tried:
1. Increasing npt and nvt equilibration time to 50ps
2. Increasing box size to 8.2x12.2x12.2
3. Decreasing npt and nvt equilibration step sizes to 0.001 ps.
However, the final mdrun still shows the same errors using dt=0.002
ps. Reducing
the final mdrun dt to 0.001ps allows the simulation to carry on but the
atoms in freezegroups do not stay stationary, like they do in v.4.6.x.
Here, I tried adding hydrogen bond constraints but then the simulation
produces the aforementioned errors even at 0.001 ps.
The mdp file used for final run:
integrator = md tinit = 0 dt = 0.002 nsteps = 5000000 init_step = 0
comm-mode = Linear nstlist = 10 nstcalclr = 100 ns_type = grid
cutoff-scheme = group nstenergy = 500 nstdihreout = 100 nstxout-compressed
= 500 pbc = xyz rlist = 1.2 coulombtype = PME fourierspacing = 0.5
rcoulomb-switch = 0 rcoulomb = 1.2 epsilon-r = 1 epsilon_rf = 0 vdw-type =
Cut-off rvdw-switch = 0 rvdw = 1.2 DispCorr = EnerPres Tcoupl = nose-hoover
tc-grps = System tau_t = 0.1 ref_t = 300 Pcoupl = berendsen Pcoupltype =
isotropic tau_p = 1 compressibility = 4.5e-5 ref_p = 1.0 gen_vel = yes
gen_temp = 300 gen_seed = -1 pme_order = 4 ewald_rtol = 1e-05
ewald_geometry = 3d epsilon_surface = 0 optimize_fft = no disre = No dihre
= No free-energy = no freezegrps = C_FIXED freezedim = Y Y Y
; this part was included to try to overcome the freezing problem at
0.001ps, but system shows errors ;constraints = h-bonds ;morse = yes

Can you please suggest what else could be done to successfully reproduce
the simulation in V.2018.1? Regards,
Neelima S.