It is actually MgB2. The lattice is almost P6/mmm but not quite a!=c, and, I guess, this additional degreeof freedom requires an additional position relaxation to obtain zero force.
It is very possible. Try to relax using variable cell.
MeshCutoff was 100 Ry. If any of these parameters are increased further, then the iterations converge tothe same value of Etot.
At 100Ry I will never believe this.
