Recognizing the need for a comprehensive overview of new results, this book looks at the problems of creating fast numerical and nonnumerical algorithms. It also contains analyses of their properties and possibilities of their efficient implementation on specialized parallel computers. Instead of treating these problems separately, the book deals with them in their complexity, taking into account their mutual relationships. Attention is paid to the following topics: - methods of creating fast parallel algorithms - parallel algorithms for associative computers - systolic algorithms and their implementation on specialized processors - algorithms for pipeline and matrix processors and for multiprocessors - solution of a system of linear equations on specialized computers - lower time bounds for SIMD-type algorithms.