Thanks for the suggestion, but I fear it is not my case. I obtain the same meaningless results with version 5.0.2 installed on another pc of my university. I suspect it is something in Gromacs procedure for computing pressures. However, thanks Francesco