On computational algorithm for inners

A modification is suggested to the computational algorithm for inners in [1], [2] in order to lessen the roundoff errors arising in computer calculations taking into consideration fixed structure of sparseness of left triangle of zeros without destroying the inners of original matrix.