Pressure tensor with nonzero xz and yz terms

We are working on a simulation that is meant to shear a liquid and calculate a pressure tensor using the compute_pressure command. From our understanding, a shear tensor should have zero terms for the xz and yz components when shearing in a single direction. However, all of our tensor terms have nonzero values. Does this mean that we are shearing the box incorrectly? Are we misinterpreting the pressure tensor itself (perhaps what LAMMPS outputs is not the pressure tensor in a traditional sense?) Thanks for your time.

Erica McCready
Northwestern University

Instantaneous pressure values will never be exactly 0.0. Do you mean
they are time-averaging to some non-neglible value?