Comparing production system architectures

Authors:
Mark Lease;Mac Lively
Affiliations:
Computer Science Department, Texas A&M University, College Station, Texas;Computer Science Department, Texas A&M University, College Station, Texas
Venue:
ACM SIGARCH Computer Architecture News - Special Issue: Architectural Support for Operating Systems
Year:
1988

Citing 5
Cited 0

Rule-based systems

Communications of the ACM
NON-VON's performance on certain database benchmarks

IEEE Transactions on Software Engineering
Parallel algorithms and architectures for rule-based systems

ISCA '86 Proceedings of the 13th annual international symposium on Computer architecture
Computer Architectures for Artificial Intelligence Processing

Computer
Initial Performance of the DADO2 Prototype

Computer

Quantified Score

Hi-index	0.00

Visualization

Abstract

Expert systems have become very popular in recent years. But the forward-chaining production systems (PS) driving the expert systems have severe limitations when run on general-purpose computers. The Match phase of the Match-Act-Select cycle production systems execute is computationally expensive and creates a limitation in the performance the PS's can deliver. This paper examines the architecture of three computers (DADO2, NON-VON, and Production System Machine) designed to speed up the Match process by pattern matching many facts and rules in parallel. The speeds of the designs are compared and the systems are analyzed to determine how suitable they are for use in the commercial environment.