Programming with Shared Data Abstractions

ing with shared data abstractions Simon Dobson1 and Don Goodeve2 1 Well-Founded Systems Unit, CLRC Rutherford Appleton Laboratory, UK 2 School of Computer Studies, University of Leeds, UK We present a programming system based around a set of highly nt, highly distributed data structures. These shared abstract data types offer a uniform interface onto a set of possible implementations, each optimised for particular patterns of use and more relaxed coherence conditions in the data. They allow applications written using a shared data model to approach the performance of message passing implementations through a process of pattern analysis and coherence relaxation. We describe the programming system with reference to a sample application solving the travelling salesman problem. Starting from a naïve direct implementation we show significant speed-ups on a network of workstations.

• Type: Conference
• Year: 1997
• Where: IRREGULAR
• Authors: Simon A. Dobson, Don Goodeve
• Keywords: CLRC Rutherford Appleton, Data Abstractions, Distributed And Parallel Computing, IRREGULAR 1997, Message Passing Implementations