Parallel Processing Systems

TL;DR: An experimental parallel computer system which is expected to achieve supercomputing performance is described, which allows the execution of single large jobs on multiple processors to perform large scale simulations of biological systems.

TL;DR: A regular architecture is proposed for the evaluation of polynomials of arbitrary degree that is easily adapted to a pipeline structure and the constraint of regularity may make faster algorithms unsuitable for certain implementations.

TL;DR: Interestingly, the serial arithmetical operations count for the present algorithm is 0(1612;N); this makes it faster than the quadrant interlocking factorization method of Chawla and Passi, the partitioning method of Wang and cyclic reduction, each of which has a serial count of 0(17N).

TL;DR: A mathematical model of neurons for multilayer neural networks is proposed and an algorithm of parallel summation of numbers stored in a array with the use of the principle of difference slices is considered.