Storage And Computing Of Matrix On Cryptanalysis

The most important cryptographic algorithm of Modern cryptography is AES, and the most threat attack method of AES is algebraic attack. As a result, we can see the importance of algebraic attack, and the form of the algebra attack is usually with sparse matrix and dense column vector product(SMVP). Therefore, matrix storage and computing in the cryptanalysis is more and more important.This article start from the importance of the storage and computing of matrix on cryptanalysis, and then were review various matrix storage and computing methods.On the basis of Huffman Coding,we research the block compressed row storage for SMVP, and based on Huffman Coding,we propose a improved block compressed row storage which is mainly suitable for the high order sparse matrix of the binary finite field. Later, we continue to broaden our thoughts, and we promote the new block compressed row storage for the general finite field.Finally, we discuss the SMVP algorithm for the implementation of project, and introduce the parallel computing method of SMVP based on the Compute Unified Device Architecture Then we combine the basic theory of SMVP closely together with the implementation of project.