Abstract

This paper describes the Gauss-Jordan method of calculating inverse matrices. An analysis of several papers on a similar topic is described, and the ways of performing the parallelization of Gauss-Jordan method for calculating inverse matrices. The achieved results of the related works are considered. An analysis of technologies suitable for parallel implementation of the calculation of inverse matrices of a rare type was performed, and the reasons why the CUDA platform is selected. Finally, the expected goals of the work itself are described.