Experiments with a Knowledge-Based System on a Multiprocessor
Abstract
This paper documents the results the authors obtained and the lessons they learned in the design, implementation, and execution of a simulated real- time application on a simulated parallel processor. Specifically, their parallel program ran 100 times faster on a 100-processor multiprocessor compared to a 1- processor multiprocessor. The machine architecture is a distributed-memory multiprocessor. The target machine consists of 10 to 1000 processors, but because of simulator limitations, we ran simulations of machines consisting of 1 to 100 processors. Each processor is a computer with its own local memory, executing an independent instruction stream. These is no global shared memory; all processes communicate by message passing. The target programming environment, called Lamina, encourages a programming style that stresses performance gains through problem decomposition, allowing many processors to be brought to bear on a problem. The key is to distribute the processing laod over replicated objects, and to increase throughput by building pipelines sequences of objects that handle stages of problem solving. The authors focused on a knowledge-base application that simulates real-time understanding fo radar tracks, called Airtrac. This paper describes a portion of the Airtrac application implemented in Lamina and a set of experiments that we performed. Following hypotheses were confirmed: (1) Performance of this concurrent program improves with additional processors, and thereby attains a significant level of speedup. (2) Correctness of the concurrent program can be maintained despite a high degree of problem decomposition and highly overloaded input data conditions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1987
- Accession Number
- ADA198708
Entities
People
- Masafumi Minami
- Russell Nakano
Organizations
- Stanford University