СПЕЦІАЛІЗОВАНИЙ ПРОЦЕСОР ДЛЯ УЩІЛЬНЕННЯ ДАНИХ

Abstract

One of the effective approaches to data compression is the approach based on the use of optimizing properties of Fibonacci numbers. The essence of the approach is that in the process of compaction the block of digital data is considered as a large positive integer, given in the form of a linear Fibonacci form. The implementation of data compression methods based on the linear form of Fibonacci software requires a lot of time, which is associated with calculations over large numbers (up to 8000 binary digits). For some applications, such time is unacceptable, so there is a need to create a specialized processor that will speed up the process of data compression. The development of mathematical and structural models of a specialized processor and its components is carried out using a functional-structural approach to the design of digital devices. Based on the generalized model of the process of adaptive data compression based on the linear Fibonacci form, the main functions to be implemented by a specialized processor are identified. This processor is part of a computer system and is in some way connected to the computer's CPU. Because the files to be compressed and the compressed files are stored in computer memory, the CPU is expected to read and write the file, generate P and P* sequences, and implement a sequence-level optimization function. The specialized processor is responsible for calculations over large numbers. To implement a set of all functions, it is proposed to build more than one operating machine, and to decompose it into machines, each of which implements the corresponding function. Mathematical models and structures of such modules of the specialized processor are considered: modeling of a data source, coding, decoding, optimization at the level of blocks, formation of structure of sequence P*. Hardware implementation of calculations over large numbers and the ability to implement basic functional transformations of individual modules in the pipeline mode provides acceleration of the data compression process compared to software implementation.