could not get the same result in shear filed

Dear Lammps Users:

I have a problem that even I use the same data and input script , but I can not get the same result in the same machine. If there is no shear fileld, everthing is good. I can get the same result in the different machine. But only if I added shear filed to my system. I would get different results every time.Sometimes the results were just alike,but sometimes there were a big difference. I use the lammps’s version is 2014 Jun 28.The below is my input script.

system in shear flow

dimension 3
units lj
special_bonds lj 1.0 1.0 1.0
boundary p p p
processors 5 5 4
atom_style bond
read_restart riven
group nano type 3 4
group polymer type 1 2
neighbor 2 bin
neigh_modify every 1 delay 3
neigh_modify exclude molecule nano

bond_style harmonic
bond_coeff 1 50.0 0.7

pair coeff

pair_style dpd 1.0 1.0 34387
pair_coeff 1 1 25.0 4.5
pair_coeff 1 2 80.0 4.5
pair_coeff 1 3 5.0 4.5
pair_coeff 1 4 5.0 4.5
pair_coeff 2 2 25.0 4.5
pair_coeff 2 3 80.0 4.5
pair_coeff 2 4 80.0 4.5
pair_coeff 3 3 50.0 4.5
pair_coeff 3 4 50.0 4.5
pair_coeff 4 4 50.0 4.5
comm_modify vel yes

fix 1 nano rigid molecule
fix 2 polymer nve

#compute xz pressur_tensor and shear viscosity
compute mytemp all temp/profile 1 0 0 z 20
compute pre all pressure mytemp
variable pre5 equal -c_pre[5]
variable visc equal v_pre5/0.03

#creat shear and save data
fix 3 all deform 1 xz erate 0.03 remap v units box
fix 4 all ave/time 1 1000 1000 c_pre[1] c_pre[2] c_pre[3] v_pre5 file press.dat
fix 5 all ave/time 1 1000 1000 v_visc file visc.dat
fix 6 all ave/spatial 1 1 1000 z 0 0.05 vx units reduced file velprofile-20.dat

thermodynamic property

dump 1 all custom 10000 *.lammpstrj id type x y z
dump_modify 1 format "%06d %d %10f %10f %10f "
timestep 0.01
thermo 1000
thermo_style custom step ke etotal c_mytemp cpu
run 500000
write_restart 540000.riven

I don’t understand your Qs. What does
same machine and different machine mean? What
does adding shear mean?

Steve