In: Microprocessing and Microprogramming, Vol. 25, pages 133-138. 1989.
Abstract: Efficiently carrying out a parallel caculation with a conventional multiprocessor machine poses difficulties: a limited concurrence, a complex control system, difficulties of programmation. In Data Flow architecture, these problems are non-existent; but a representation model of parallel calculation is necessary: Among the existing models, an extesnion of the basic Petri nets, dubbed ``Data Flow Petri Nets'', has been selected.