Exact solving of a linear equations set

Theoretical and experimental results of application of exact computation for solving of linear algebraic equations are presented in the paper. In particular it is demonstrated that computational bit complexity of solving of a linear algebraic equations set with non-degenerate matrix are not exceeding 0(1 7/2), and computational bit complexity of computing of normal pseudo solution are not exceeding 0(15log2I), where / is bit volume of input data. For computational speedup it is reasonably to use multiprocessing. It is illustrated that computational speedup for considered problems under of exact computation equal to number of processors.