> I am using a Ge supercell with 64 atom and two defect sites. Can > someone give me an estimate > of the number of kpoints needed for this type of calculation? semiconductor + big cell => small number of k-points needed. A single one might be enough? Test ! Stefaan