Principle and Algorithms of Real-Time Parallel Processing For Distributed Processor

The paper presents the novel principle on constructing a new class of highly parallel fast stable numerical algorithms of linear algebra. In accordance with it the parallel modification of classic Gram-Shmidt algorithm is developed. The new algorithm has super-linear convergence rate and highest numerical stability supported by easy controlled parallel cross-feedbacks. It provides for high computing density in distributed processors and proposed for using in area of multidimensional data stream real-time processing. Index Terms - Parallel processing, linear algebra, numerical algorithms, distributed processor. I. INTRODUCTION The problem of fast and stable solution of a large linear algebraic systems plays an important role in areas of real-time processing of multidimensional data streams. Well-known classic numerical linear algebra methods have hard sequential structure and low numerical stability and hence are restricted for the proper realization in parallel distributed processor..