Data Distribution Algebras - A Formal Basis for Programming Using Skeletons

Today there is no programming model or language support which ooers a uniform programming methodology on a wide variety of parallel architectures. In this paper functional languages are proposed as such a methodology using an extension of the concept of skeletons | higher-order functions coupled with parallel implementation templates. An essential part of the proposed methodology is the use of data distribution algebras as a structured language concept to specify distribution issues explicitly in a skeleton-based approach. It is shown how the skeleton composition problem can be solved using data distribution algebras. Finally the methodology is applied to the solution of partial diierential equations in the context of uid dynamics.