Hi All. Is the SCF Tolerance flag in DMOL3 calculated per atom or per the whole system? CASTEP gives an option about this, but DMOL3 does not, apparently. If this is the SCF Tolerance per the whole system; then doesn't an appropriate value for a medium-sized system (based on the number of atoms) become too small (too accurate of tolerance) for a larger system? We should treat this quantity with a cautionary note that it should be system-size-dependent right?