Error on bond for bond_harmonic potential while np is greater than 1

Hello everyone,

I am using LAMMPS to simulate wave propagation in an 1D
bead-spring chain with bond_harmonic potential, but after it reads my
data file, the following error message prints for 2 or greater number
of processors:

"Bond atoms 800 801 missing on proc 0 at step 0" (for two processors)

I would be grateful if someone could give me any advice.
My input file is listed as follows and its data file is attached.

Input script:

1D wave propagation

clear

units lj
dimension 2
boundary s p p
atom_style bond
bond_style harmonic

atom_modify map array sort 0 0.0
neighbor 2.0 nsq

read_data data.wave1d-noper

mass * 1.0

timestep 0.001

group source type 2
group other type 1

variable A equal 30.3
variable t0 equal 3.2
variable Zeta equal 2.0

variable xbef equal x[100]
variable xbef0 equal ${xbef}

variable xaft equal x[900]
variable xaft0 equal ${xaft}

variable xbedel equal x[100]-{xbef0} variable xafdel equal x[900]-{xaft0}

variable timer equal step*dt

thermo_style custom step

variable Lx equal -A*(step*dt-{t0})*exp(-{Zeta}*(step*dt-{t0})^2)
variable Ly equal 0.0
variable Lz equal 0.0

variable Vx equal A*(2*{Zeta}(stepdt-{t0})^2-1)*exp(-{Zeta}(stepdt-${t0})^2)
variable Vy equal 0.0
variable Vz equal 0.0

fix 1 other nve
fix 2 source move variable v_Lx v_Ly NULL v_Vx v_Vy NULL units box

fix PrintData all print 5 “{timer} {xbedel} ${xafdel}” file data.fourier

dump visualize all xyz 100 dump_wave1d_no-per.xyz

run 96000

With the Best Regards!

Zia Movaghar

data.wave1d-noper (92.2 KB)

Hello everyone,

    I am using LAMMPS to simulate wave propagation in an 1D
bead-spring chain with bond_harmonic potential, but after it reads my
data file, the following error message prints for 2 or greater number
of processors:

"Bond atoms 800 801 missing on proc 0 at step 0" (for two processors)

yup. the reason for that is that communication of atom information
in lammps is based on non-bonded atom cutoffs (which are usually
much longer than typical bond lengths). you only have harmonic
bonds, so this cutoff defaults to 0.0. you can adjust this using the

communicate cutoff

option. you have to make sure that this cutoff value is large enough
to handle maximally stretched bonds. you won't see this when running
on a single processor, since there are no atoms to be communicated.

let us know, if this doesn't solve your problem.

axel.