スパース行列の演算法について

概要

論文の詳細を見る
The process of matrix computation is very important in the system of linear problem. The lager problem, the higher ratio of zero element are included in the matrix. The matrix of which the proportion of non-zero element is very small is called sparse matrix. Thus, total computation efficency is influenced by treating sparse matrix in system of large size linear problem. Various kind of computation method for sparse matrix were proposed by many researchers. (e.g. Knuth et al.) In this paper, we present the computation characteristics of the sparse matrix by using the Row-wise format which may be considered to be most apropriate for the operation of row-wise computation of matrix. The results of our investigation are obtained by comparing the computation time between ordinary matrix operation and Row-wise format operation.
岡山理科大学の論文