Document Type





Parallel computing




Computer Sciences


We introduce a model of parallel computation that retains the ideal properties of the PRAM by using it as a sub-model, while simultaneously being more reflective of realistic parallel architectures by accounting for and providing abstract control over communication and synchronization costs. The Hierarchical PRAM (H-PRAM) model controls conceptual complexity in the face of asynchrony in two ways. First, by providing the simplifying assumption of synchronization to the design of algorithms, but allowing the algorithms to work asynchronously with each other; and organizing this "control asynchrony" via an implicit hierarchy relation. Second, by allowing the restriction of "communication asynchrony" in order to obtain determinate algorithms (thus greatly simplifying proofs of correctness). It is shown that the model is reflective of a variety of existing and proposed parallel architectures, particularly ones that can support massive parallelism. Relationships to programming languages are discussed. Since the PRAM is a sub-model, we can use PRAM algorithms as sub-algorithms in algorithms for the H-PRAM; thus results that have been established with respect to the PRAM are potentially transferable to this new model. The H-PRAM can be used as a flexible tool to investigate general degrees of locality (“neighborhoods" of activity) in problems, considering communication and synchronization simultaneously. This gives the potential of obtaining algorithms that map more efficiently to architectures, and of increasing the number of processors that can efficiently be used on a problem (in comparison to a PRAM that charges for communication and synchronization). The model presents a framework in which to study the extent that general locality can be exploited in parallel computing. A companion paper demonstrates the usage of the H-PRAM via the design and analysis of various algorithms for computing the complete binary tree and the FFT/butterfly graph.

Additional Information

School of Computer and Information Science, Syracuse University, SU-CIS-91-06





To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.