Jointly sponsored by DIMACS and the Center for Research on Parallel Computing.
The ability to construct and understand the behavior of complex high-level parallel programs requires a rich abstract specification framework, and equally expressive language abstractions and tools.
The extent to which specification techniques and language abstractions can hide low-level details about specialized architectural features, and the degree to which they can enable the design of new parallel algorithms and help to implement existing ones are thus topics of important current research.
The goal of the workshop is to bring together some of the best researchers in specification techniques, parallel languages, algorithms, and systems to present and discuss recent developments in their areas of expertise, and interact on current issues of common interest. The emphasis of the workshop will be placed on specification methoc.ologies, new language and system abstractions, and techniques that bridge theory and practice.
Some of the topics that will be discussed at this workshop include: