Large sparse matrices multiplication often appear in scientific or engineering applications when solving partial differential equations. By this means, if we store and calculate every element of the matrix in computer it may waste much space. So we need find a way to calculate the sparse matrix more efficiently in order to make sparse matrix-vector multiplication at high speed. If we could find a way not only to store but also to calculate the large sparse-vector multiplication very efficiently it would benefit in engineering or scientific domain to save time.