Comments on P3M, FMM, and the Ewald method for large periodic Coulombic systems