Multi-word Atomic Read/Write Registers on Multiprocessor Systems
Journal article, 2004

Modern multiprocessor systems offer advanced synchronization primitives, built in hardware, to support the development of efficient parallel algorithms. In this paper we develop a simple and efficient algorithm for atomic registers (variables) of arbitrary length. The simplicity and better complexity of the algorithm is achieved via the utilization of two such common synchronization primitives. In this paper we also evaluate the performance of our algorithm and the performance of a practical previously know algorithm that is based only on read and write primitives. The evaluation is performed on 3 well-known, parallel architectures. This evaluation clearly shows that both algorithms are practical and that as the size of the register increases our algorithm performs better, accordingly to its complexity behavior.

Author

Andreas Larsson

Chalmers, Department of Computing Science, Distributed Computing and Systems

Anders Gidenstam

Chalmers, Department of Computing Science, Distributed Computing and Systems

Phuong Ha

Chalmers, Department of Computing Science, Distributed Computing and Systems

Marina Papatriantafilou

Chalmers, Department of Computing Science, Distributed Computing and Systems

Philippas Tsigas

Chalmers, Department of Computing Science, Distributed Computing and Systems

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

03029743 (ISSN) 16113349 (eISSN)

Vol. 3221 736-748

Subject Categories

Computer and Information Science

DOI

10.1007/978-3-540-30140-0_65

ISBN

3-540-23025-4

More information

Created

10/6/2017