Abstract

Introduction In the last years two different paradigms for parallel computers have been established. In the first one, called parallel computer networks or networks for short, processors communicate by neighbour-to-neighbour message passing via a network, typically working asynchronously. The second one connects the processors with a shared memory so that information exchange can be realized via access to public variables. This machine type, called PRAM as a theoretical model, works synchronously and has the great advantage of easy programming. On the other hand, current technology does not allow a direct realization of (large) PRAMs in hardware. All real machines with shared memory use some kind of simulation. Both machine types are reflected in programming languages with communication facilities of either the message passing or shared memory type. There have been several approaches, which tried to support a pure PRAM programming style on netw

Citations