Abstract
Many multiprocessors work inefficiently, because the individual processing elements are idle most of the time and the exploitation of parallelism on high level is restricted by synchronization mechanisms, while low level parallelism is not worth a distribution. The PESAI multiprocessor, which is designed for the execution of production systems, avoids these disadvantages by the projection of a data flow based graph on a folded pipeline of parallel processors. The pipeline stages are interconnected by a bus system. Several experiments about details of implementation have been performed. This paper describes the evaluation of the simulation results which are used to enhance the performance and optimize the configuration of the PESAI architecture. The performance could be increased by about 20 percent. Synchronization has to be performed only in a few cases that occur rarely during execution. Furthermore, the amount of utilizable parallelism is increased. Thus the number of processors performing nonredundant computation can be extended.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Schreiner, F.; Zimmermann, G.: “PESAI A Parallel Architecture For Expert Systems”, 16th International Conference On Parallel Processing, 1987, p 166–169.
Forgy, C.: “OPS5 Users Manual”, Technical Report CMU-CS-81–135, Carnegie-Mellon University, 1981.
Brownston, L.; Farrel, R.; Kant, E.; Martin, N.: “Programming Expert Systems in OPS5: An Introduction to Rule Based Programming”, Addison Wesley 1985.
Forgy, C.: “RETE: A Fast Algorithm for the Many Pattern/Many Object PatternMatch Problem”, Artificial Intelligence 19 (September 82), p 17–37.
Gupta, A.; Forgy, C.; Newell, A.; Wedig, R.: “Parallel Algorithms and Architectures for Rule-Based Systems” 13th Intern. Symp. on Computer Architecture, Tokyo 1986, p 28–37.
Oshisanwo, A.; Dasiewicz, P.: “A Parallel Model and Architecture for Production Systems”, 16th International Conference On Parallel Processing, 1987, p 147–153.
Stolfo, S. “Initial Performance of the DAD02 Prototype”, Computer, January 1987.
Forgy, C.; Gupta, A.: “Measurements on Production Systems”, Technical Report CMU-CS-83–167 Carnegie-Mellon University, 1983.
Gupta, A.: “Parallelism in Production Systems”, Technical Report CMU-CS-86–122, Carnegie-Mellon University, 1986.
Bentley, J.: “Thanks, Heaps”, Communications of the ACM Vol 28 Number 3, March 1985, p 245–250.
Miyazaki, J.; Amano, H.; Aiso, H.: “Manji: An Architecture for Production Systems” 20th International Hawaii Conference on Systems Sciences 1987, p 236–245.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1988 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Schreiner, F., Zimmermann, G. (1988). Parallelism in the Pesa I Multiprocessor. In: Valk, R. (eds) GI — 18. Jahrestagung II. Informatik-Fachberichte, vol 188. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-74135-7_13
Download citation
DOI: https://doi.org/10.1007/978-3-642-74135-7_13
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-50360-6
Online ISBN: 978-3-642-74135-7
eBook Packages: Springer Book Archive