Sholota V. Structure Organization of Parallel Job-oriented Processor for Matrix Problems of Linear Algebra.

Object of research - architecture of high-efficiency computing systems for parallel processing of matrixes of the large dimension. A subject of research - structural organization of computing systems for parallel processing of matrixes of the large dimension. Methods of researches - system analysis, theory of parallel processing of matrixes, theory of algorithms, theory of designing of computing systems and theory of digital optical processing of the information. The purpose of researches - development of parallel methods of calculations, algorithms and structures high-speed parallel job-oriented processor for matrix tasks in linear algebra, wich based on the principles of natural parallelism of digital optical computing. The theoretical and practical results - are offered parallel interpretations of methods Gaussa and Gaussa-Jordana for the decision of systems of the linear algebraic equations of the high orders and matrix invertion of the large dimensions on the basis of calculation of external mu ltiplication of vectors, structural organization and hardware-oriented algorithms of parallel job-oriented processors with a floating point for the decision of systems of the linear algebraic equations and matrix invertion, digital parallel structures for matrix arithmetics in the form with a floating point: addition of matrixes, division of a vector into number and calculation of external multiplication of vectors. The novelty offered - proves to be true by a high level by approbation of works, including at the international level, presence 1 patent of Ukraine. A degree of introduction - in borders of area. Area of using - high-perfomens systems of image recognition and signals and images processing.