Consider a rA CA matrix A and a rB CB matrix B, with CA and Cij be the entry in the ith row and jth column of C, then 'B. Let C-A B, CA ! As can be seen from the above equation, the computation of Cy depends only on the matrices A and B. Hence, the entries of the matrix C can be computed independently of each other. In other words, they can be computed concurrently. This will speed up the computation significantly in a multiprocessor system, especially if the sizes of the matrices involved are very large