Efficient algorithm for sparse matrix computations

S. C. Park, Louisiana State University
J. P. Draayer, Louisiana State University
S. Q. Zheng, Louisiana State University


A new representation of a sparse matrix is introduced that is very efficient for matrix multiplication when the non-zero elements are partially or fully adjacent to one another as in band or triangular matrices. Space complexity is better than that of the existing algorithms when the number of the groups of adjacent non-zero elements is less than two-thirds of the total number of non-zero elements. Time complexity is better or much better than that of existing algorithms depending on the number of groups of non-zero adjacent elements in the factor matrices.