The more critical stage of both these algorithms is the sparse product of the matrix R by a column vector which can be efficiently done with ΞΆ operations.