Abstract. The author considers the ultrafast cellular method of matrix multiplication, which operates by cellular submatrices, interacts with well-known matrix multiplication cellular methods, and minimizes by 12.5% the computational complexity of cellular analogs of well-known matrix multiplication algorithms derived on their basis. The interaction of the ultrafast cellular method with the unified cellular method of matrix multiplication provides the highest (in comparison with well-known methods) percentage (equal to 45.2%) of minimizing of the multiplicative, additive, and overall complexities of the well-known matrix multiplication algorithms. The computational complexity of the ultrafast method is estimated using the models of getting cellular analogs of the traditional matrix multiplication algorithm.

Keywords: linear algebra, cellular methods, family of cellular methods of matrix multiplication, cellular analogs of matrix multiplication algorithms.

¹ V.M. Glushkov Institute of Cybernetics of National Academy of Sciences of Ukraine, Kyiv, Ukraine

 larisaelf@gmail.com