As computing demands increase, emphasis is being placed on parallel architectures- To efficiently use parallel machines, software must be designed to take advantage of these machines. This research concentrates on an abstraction of algorithm design to permit the expression of parallel programs. The abstraction emphasizes thought about algorithms at a high level as opposed to algorithm implementation at a statement level. A model based on data flow allows algorithm expression using flow diagrams. The model specifies operating system requirements that support parallel programming at a module level. Paths are used to carry data between modules. Data enter modules through ports. Module activation is triggered by the satisfaction of data availability conditions. Continual module presence within the system, dynamic activation criteria, and a high level of programming distinguishes this model from other parallel programming systems.
Eggen, Roger E. and Metzner, John R., "A Semantic Basis for Parallel Algorithm Design" (1986). Computer Science Technical Reports. 74.
© 1986 University of Missouri--Rolla, All rights reserved.