Thermo-style pe & compute pe

Dear Lammps users ,

I have a question about the potential energy of the system that we can have as an output with thermo_style command and also compute pe command .

I am doing MD simulation on atoms confined between two Lennard-jones walls .
The potential energy that i get by “thermo_style” command is positive but with “compute pe” is negative !

Can you please let me know what probably can be the reason of this strange output ?!

Thanks in advance.

Is any other quantity (e.g. the wall interaction) being added

to the system pe?

If you take away the walls (even for timestep 0) do the 2 PEs