Abstract
In case of algorithms with non-local data access patterns virtual shared memory implementations commonly result in a huge lack of performance because of virtual shared memory latency. In order to avoid this problem, a system providing access to distributed data structures has been developed, implemented as a C++ object library. By means of this library the programmer may use special access objects as agents to access the distributed data structures. In case of accessing arrays these access objects may be buffers or queues. If the algorithms are suited for using these access objects, the communication bandwidth instead of the communication latency becomes the major performance factor.
Preview
Unable to display preview. Download preview PDF.
References
K. Li (1986) Shared Virtual Memory on Loosely Coupled Multiprocessors, Ph.D. Thesis, Yale University
H. Hellwagner, Virtually Shared Memory Architectures for Scaleable Universal Parallel Computers, Applied Computer Science and Software (Proc. Int'l Symp.), Springer-Verlag (1991)
H. Hellwagner, A Survey of Virtually Shared Memory Schemes, SFB Report No. 342/ 33/90A, Techn. Univ. Munich (1990)
B. Nitzberg, V. Lo, Distributed Shared Memory: A Survey of Issues and Algorithms, COMPUTER 24, 8(1991)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1994 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Lüpke, S. (1994). Accelerated access to shared distributed arrays on distributed memory systems by access objects. In: Buchberger, B., Volkert, J. (eds) Parallel Processing: CONPAR 94 — VAPP VI. VAPP CONPAR 1994 1994. Lecture Notes in Computer Science, vol 854. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-58430-7_40
Download citation
DOI: https://doi.org/10.1007/3-540-58430-7_40
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-58430-8
Online ISBN: 978-3-540-48789-0
eBook Packages: Springer Book Archive