write_data bug? No dihedral coeffs in new data-file

Hi all,

I’ve encountered a very strange problem and I haven’t found any similiar threads yet.
Basically I’m just equilibrating a system consisting of 100 pa6-chains each with 100 repetitive units.

However, when I check the output data-file at the end of the simulation, the dihedral-coeffs section is missing even though the header says that there are 13 different dihedral types and any subsequent simulation using this file as input will produce the error “no dihedral coeffs” (or something like this) …

I’ve attached a minimal-example to this mail. Please check it, maybe I’ve done an obvious mistake, but none of my colleagues was able to help me.

here’s my input-script (also attached to this mail as a file)

General

units real
atom_style full
dimension 3
#newton on
boundary p p p

Styles

pair_style lj/cut 12.5 #/coul/long 12.5 9.5 #pair_cutoff charge_cutoff
bond_style harmonic
angle_style harmonic
dihedral_style multi/harmonic
improper_style harmonic

#kspace_style pppm 1.0e-4

special_bonds lj/coul 0 0 0.5

System definition

read_data “polymer.data”

Settings

dielectric 1.0
neighbor 2.0 bin
neigh_modify delay 0 every 1 check yes page 500000 one 50000

thermo_style custom step temp etotal pe evdwl ecoul epair ebond &
eangle edihed eimp elong press vol density
thermo 100
timestep 1.0

Equilibration

fix 1 all npt temp 270.0 270.0 50 aniso 1 1 1000 drag 2
run 1000
unfix 1

write_data stage2.data

polymer.data (250 KB)

equi.in (762 Bytes)

Hi all,

I've encountered a very strange problem and I haven't found any similiar
threads yet.
Basically I'm just equilibrating a system consisting of 100 pa6-chains each
with 100 repetitive units.
However, when I check the output data-file at the end of the simulation, the
dihedral-coeffs section is missing even though the header says that there
are 13 different dihedral types and any subsequent simulation using this
file as input will produce the error "no dihedral coeffs" (or something like
this) ....

you probably want to use write_restart and read_restart for this? any
particular reason you prefer write_data?

I've attached a minimal-example to this mail. Please check it, maybe I've
done an obvious mistake, but none of my colleagues was able to help me.

the explanation is in the write_data documentation (and i has been
discussed on the lammps-users mailing list multiple times): not all
styles are supported (in fact, at the time of the writing of that
documentation, only a few pair styles would support writing
coefficients to the data file). obviously, your dihedral style doesn't
have support for the write_data command yet.

you can either put the dihedral coeff statements into the input file
(or an include files), which is preferred by many LAMMPS users, or add
the missing subroutines to the dihedral class for your dihedral style
or try using binary restarts.

axel.

Thank you very much axel.

regards,
frank.

Hi all,

I’ve encountered a very strange problem and I haven’t found any similiar
threads yet.
Basically I’m just equilibrating a system consisting of 100 pa6-chains each
with 100 repetitive units.
However, when I check the output data-file at the end of the simulation, the
dihedral-coeffs section is missing even though the header says that there
are 13 different dihedral types and any subsequent simulation using this
file as input will produce the error “no dihedral coeffs” (or something like
this) …

you probably want to use write_restart and read_restart for this? any
particular reason you prefer write_data?

I’ve attached a minimal-example to this mail. Please check it, maybe I’ve
done an obvious mistake, but none of my colleagues was able to help me.

the explanation is in the write_data documentation (and i has been
discussed on the lammps-users mailing list multiple times): not all
styles are supported (in fact, at the time of the writing of that
documentation, only a few pair styles would support writing
coefficients to the data file). obviously, your dihedral style doesn’t
have support for the write_data command yet.

you can either put the dihedral coeff statements into the input file
(or an include files), which is preferred by many LAMMPS users, or add
the missing subroutines to the dihedral class for your dihedral style
or try using binary restarts.

axel.