Browse Prior Art Database

PERFORMANCE MODELING OF REAL-TIME MULTIPROCESSORS WITH TIME-SHARED BUSES

IP.com Disclosure Number: IPCOM000128489D
Original Publication Date: 1984-Nov-01
Included in the Prior Art Database: 2005-Sep-16

Publishing Venue

Software Patent Institute

Related People

Woodbury, Michael H.: AUTHOR [+4]

Abstract

The workload effects on computer performance are addressed for a highly reliable unibus multiprocessor used in real-time control. Because of the strict performance criteria required by a system of this type, it would be desirable to be able to determine the significant effects of workload distribution and scheduling on performance. As an approach to studying these effects, a modified stochastic Petri net (SPN) is used to describe the synchronous operation of this system. From this model the vital components affecting performance can be determined. However, because of the complexity in solving the modified SPN, a simpler model is constructed that presents the same critical aspects. This model is a closed queueing network. It consists of multiserver nodes and a non- preemptive priority queue. The use of this model for a specific application requires the partitioning of the workload into job classes. The steady state solution of the queueing model directly produces useful results, such as idle processing time, system bus contention, and bus queueing times, that are necessary in any performance evaluation. The use of this model in evaluating an existing system, the Fault Tolerant Multiprocessor (FTMP), is outlined with some experimental results.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 6% of the total text.

Page 1 of 16

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

PERFORMANCE MODELING OF REAL-TIME MULTIPROCESSORS WITH TIME-SHARED BUSES

Michael H. Woodbury and Kang G. Shin

THE UNIVERSITY OF MICHIGAN COMPUTING RESEARCH LABORATORY CRL-TR-46-84

November 1984

Room 1079, East Engineering Building
Ann Arbor, Michigan 48109
USA
Tel: (313) 763-8000

PERFORMANCE MODELING OF REAL- TIME MULTIPROCESSORS WITH TIME-SHARED BUSES 1 Michael H. Woodbury and Kang G. Shin

Computing Research Laboratory

Department of Electrical Engineering and Computer Science The University of Michigan
Ann Arbor, Michigan 48109

ABSTRACT

The workload effects on computer performance are addressed for a highly reliable unibus multiprocessor used in real-time control. Because of the strict performance criteria required by a system of this type, it would be desirable to be able to determine the significant effects of workload distribution and scheduling on performance.

As an approach to studying these effects, a modified stochastic Petri net (SPN) is used to describe the synchronous operation of this system. From this model the vital components affecting performance can be determined. However, because of the complexity in solving the modified SPN, a simpler model is constructed that presents the same critical aspects. This model is a closed queueing network. It consists of multiserver nodes and a non- preemptive priority queue. The use of this model for a specific application requires the partitioning of the workload into job classes. The steady state solution of the queueing model directly produces useful results, such as idle processing time, system bus contention, and bus queueing times, that are necessary in any performance evaluation. The use of this model in evaluating an existing system, the Fault Tolerant Multiprocessor (FTMP), is outlined with some experimental results.

[ Chapter ] 1. INTRODUCTION

1 The work described in this report was supported in part by NASA Grant No. 1-296 and NASA Training Grant No. NGT 23-005-801. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of NASA.

University of Michigan Computing Research Laboratory Page 1 Nov 01, 1984

Page 2 of 16

PERFORMANCE MODELING OF REAL-TIME MULTIPROCESSORS WITH TIME-SHARED BUSES

Representing the operation of a computer system by a structured model is a popular and natural approach to the study of a computer's performance. Many factors need to be incorporated into the model so that it accurately describes the system that is being modeled. The type of analysis desired dictates which factors of the computer's operation need to be incorporated into the modeling framework. A factor that is almost always included, especially in the study of computer performance, is the representation of the workload handled by the computer system being analyzed. The workload is an essential part of the performance evaluation o...