Hi all, I am using jan 2009 version of LAMMPS for the simulations. I am pulling a periodic box of polymer in one direction with NPT ensemble for 100000 steps, I got a 6 components of stresses in log file. LAMMPS calculating the stresses by using virial formula with all components like pair,bond,angle,dihedral,improper so on,right. Now i want to get the stresses due to pair, stresses due to bond, stresses due to angle, so on separately. For this i have used commands like

compute peratomstress all stress/atom

compute globalstress all reduce sum c_peratomstress[1] c_peratomstress[2] c_peratomstress[3]

thermo_style custom c_globalstress[1] c_globalstress[2] c_globalstress[3]

But these are not matching with log file values pxx pyy pzz even after dividing with volume.

i want to know what is the problem ?

thanks,

grn