Author(s)

J. Mushtaq, A.J. Davies

Abstract

Solving large computationally intensive problems requires significant processing power and the current trend in achieving this power is not by increasing the throughput of individual processors, but to sub-divide the problem into co-operating tasks so that many processors may be used to solve the problem. Stand-alone workstations, such as the PC and the SUN machines, delivering tens of millions of operations per second are common place, but to achieve high parallel efficiency, distributed systems must support rapid communication between a large number of proce

Keywords