Error #IND (serial vs parallel)

Dear sir

While running LAMMPS using the following script, the simulation is running fine when I’m using the serial executable file (lmp_win_no-mpi.exe). But when using the parallel executable (lmp_mpi.exe), I am getting #IND values for the thermodynamic quantities during the equilibration stage. What could be the reason.

variable fname index C3S_halfwater_C2Sx2_20.txt
variable simname index C3S_halfwater_C2Sx2_20

units real
boundary p p p
atom_style full
pair_style lj/cut/coul/long 12
bond_style harmonic
angle_style harmonic
kspace_style ewald 1.0e-4

log log.{simname}.txt read_data {fname}

#---------------------Pair Coefficients-------------------------#

pair_coeff 129 129 0.24 3.29 #Ca
pair_coeff 30 30 0 0 #Hw
pair_coeff 31 31 0.155 2.82 #Ow
pair_coeff 3238 3238 0.06 3.03 #Os
pair_coeff 4146 4146 0.06 3.03 #Os
pair_coeff 5076 5076 0.06 3.03 #Os
pair_coeff 3940 3940 0.06 3.39 #O
pair_coeff 4749 4749 0.06 3.39 #O
pair_coeff 7785 7785 0.013 0.98 #Si
pair_coeff 8688 8688 0.24 3.29 #Ca

#----------------------Bond Coefficients------------------------#

bond_coeff 133 250 1.68 #Si O
bond_coeff 34 1054.20 1.0 #O H
bond_coeff 35
49 250 1.68 #Si O

#------------------------Angle coefficients---------------------#

angle_coeff 17 160 109.5 #O Si O
angle_coeff 8 75.90 109.5 #H O H
angle_coeff 9
73 160 109.5 #O Si O

#read_data potential.txt

#neighbor 2.0 bin
#neigh_modify every 1 delay 0 check yes

#--------------------MINIMIZATION----------------------------#
#velocity all create 300 12345 mom yes rot no
group water type 30 31
group C3S subtract all water
fix 1 C3S spring/self 250 xyz
dump 1 all dcd 1 dump_min.dcd
min_style cg
thermo 1
minimize 1.0e-25 1.0e-25 100000 100000
undump 1

#--------------------EQUILIBRATION---------------------------#

reset_timestep 0
timestep 1
#velocity all create 300 12345 mom yes rot no
variable temp equal “temp”
variable press equal “press”

dump 2 all dcd 200 dump_eq_1.dcd

#fix def1 all print 10 “{temp} {press}” file TempPress.txt screen no
#fix def1 all print 10 “${press}” file Pressure1.txt screen no

fix 2 water nve
fix 3 water langevin 1 1 5 123456

#Set thermo output
thermo 1
thermo_style custom step lx ly lz press pxx pyy pzz temp

Run for at least 10000 femtosecond (assuming 1 fs timestep)

run 10000
unfix 2
unfix 3
undump 2

Please post a simple (as possible) input script and
data file that reproduces the problem. Can you test
that a simple Linux build (e.g. with g++) reproduces
the issue in serial vs parallel?

Steve