Using rerun for consecutive trajectories

Hi All,

I have recently started using LAMMPS. I am doing NPT simulation on polymers in a water box. I have a fairly large system with close to 50000 atoms. There is no coarse-graining involved, I used all-atom force fields.
I did 4 consecutive runs of 5ns each as a result I have 4 trajectories files for 5ns,10ns,15ns and 20ns. Since I used data files generated at the end of each simulation as the input for the next run, all my trajectory files start at timestep 0.
I used rerun to calculate rdf in the following manner:

units real
atom_style full

boundary p p p
pair_style lj/cut/coul/long 15.0
pair_modify tail yes
bond_style harmonic
angle_style harmonic
dihedral_style opls
special_bonds lj/coul 0 0 0.5
kspace_style pppm 0.00001

variable nevery index 500
variable nrepeat index 500
variable nfreq index 500000
variable nbins index 500
variable trjtimesteplast index 2500000
variable molname index 10napss36f1wat_spc_160mMnacl

read_data data.${molname}_npt_5ns

If all you dump files start at time 0 and overlap, I don’t
see how to read them back into the same rerun
script and try to sort them out.


a) use the reset_timestep command when you
perform the original 4 runs and write the 4 dump files to write them out with
non-overlapping timesteps

b) read them in and rerun them with 4 different input scripts
so that its clear what values you are reading from
which dump files