Methods of constructing high-performance systems based on the data compression

A description of the main methods of compact storage and processing of sparse matrices and provides a comparative evaluation of the effectiveness of data compression schemes on the basis of list and methods fixed format and method. Different variants of list data compression schemes and evaluation of their performance is done. It is noted that the practical implementation of carrots is associated with the need to complete conversion of all arrays in the conduct in the description of additional elements, due to technical difficulties in the process of building software and a decrease in performance of its further functioning. It is shown that using the fixed-format feature is the inability to be included in the description of additional non-zero elements, which does not apply these methods to process information for solving systems of equations of the inevitable emergence of new non-zero elements in the solution process. It sets the overall structure of a two-stage forming compact description and describes the implementation of data compression based on the character and numerical analysis.