Unifying Programming Support for Parallel Computers

Highly parallel computation requires programming support tools distinct from those that have been developed for sequential computation. In this paper, we discuss characteristics of parallel programming support that we have found to be useful. In addition, we describe a compatible collection of existing tools — taken from the Poker Parallel Programming Environment [1], the Prep-P Mapping Preprocessor [2] and the Simple Simon Programming Environment [3] — that could form the basis of an integrated, comprehensive parallel environment.