Abstract
Parallel programs are described as action systems. These are basically nondeterministic do-od programs that can be executed in both a sequential and a parallel fashion. A method for refining the atomicity of actions in a parallel program is described. This allows derivation of parallel programs by stepwise refinement, starting from an initial highl level and sequential program and ending in a parallel program for shared memory or message passing architectures. A calculus of refinements is used as a framework for the derivation method. The notion of correctness being preserved by the refinements is total correctness. The method is especially suited for derivation of parallel algorithms for MIMD-type multiprocessor systems.KeywordsAction SystemParallel AlgorithmParallel ProgramSequential ProgramComputation SequenceThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
