Dear all,
I am trying to run a longer simulation with SPCE water model provided with in the lammps benchmark files. I get the error: Out of range atoms - cannot compute PPPM, after I set the timestep for 1fs and run it for 2 million steps. The reason for that is because I want to run the simulation for 2 ns, because I need a longer time frame for my Structure Dynamic Factor calculation. Below is the input file. Can somebody please help me get rid of the error.
SPC/E water box benchmark
units real
atom_style full
read_data data.spce
replicate 2 4 1
pair_style lj/cut/coul/long 9.8 9.8
kspace_style pppm 1.0e-4
pair_coeff 1 1 0.15535 3.166
pair_coeff * 2 0.0000 0.0000
bond_style harmonic
angle_style harmonic
dihedral_style none
improper_style none
bond_coeff 1 1000.00 1.000
angle_coeff 1 100.0 109.47
kspace_style pppm 1.0e-5
special_bonds lj/coul 0.0 0.0 0.5
velocity all create 293.6 432567 dist uniform
fix 2 all npt temp 293.6 293.6 100.0 iso 1.0 1.0 100.0
neighbor 2.0 bin
neigh_modify delay 0 every 1 check yes
thermo 100
thermo_style custom step temp press vol
thermo_modify norm no flush yes
dump 1 all dcd 200 spce_bench_uwrapped_every_100fs_2ns_293.dcd
dump_modify 1 unwrap yes
timestep 1.0
Dear all,
I am trying to run a longer simulation with SPCE water model provided with
in the lammps benchmark files. I get the error: Out of range atoms - cannot
compute PPPM, after I set the timestep for 1fs and run it for 2 million
steps. The reason for that is because I want to run the simulation for 2 ns,
because I need a longer time frame for my Structure Dynamic Factor
calculation. Below is the input file. Can somebody please help me get rid of
the error.
there are two obvious and fundamental mistakes:
- you cannot run a flexible water potential at a 1fs time step. you
need to lower the time step to about 0.25 to accurately integrate the
vibrations of the hydrogen atoms. if you go larger, your integrator
will become increasingly inaccurate and ultimately diverge at some
large fluctuation.
- SPC/E is a rigid water potential and thus requires using fix shake,
which in turn will allow you to use a larger time step, as it
eliminates the high frequency motions.
on top of that, if you definitely want to increase your pppm
convergence for a variable cell calculations or else you will not have
reasonable energy conservation (or will have to hide it behind to much
thermostatting, which in turn will give you bogus trajectory data).
axel.
Thank you very much, and just to make sure, by increasing my pppm convergence you mean going from 1e-4 to 1e-8 for example, or the other way to, like 1e-1?