الفهرس 
AcknowledgementOnly 14 pages are availabe for public view
 |  | from | 120 |  |  |
| | | from | 120 | | |
Abstract
In this work, models for time–shared computer systems are described. Perf or–
i
ranee of these models is evaluated using analytical and simulation techniques. Two examples of time–shared computer systems are studied. In chapter 1, the nature of computer performance evaluation is discussed in zeneral. The computer system is described by its resources. Input to the sys¬tem is the workload. Different types of workloads are cited. Performance mea¬sures are presented including user and system performance. Then tools for evaluating these measures are described.In chapter 2, network models are analyzed using queueing theory. A computer program to predict the performance of a computer system is give. Chapters 3 and 4, present two case studies. One is a time–shared multiprogram¬med paged virtual memory computer system. The other is the PDP 11/40 running under the Resource Sharing Time Sharing / Extended system ( RSTS/E ). Approx¬imate solution using analytical techniques, and simulation results are presen¬ted. Performance of these systems is analyzed by varying system parameters. Optimum quantum size to give the best user performance is found. Tuning the PDP 11/40 running both interactive and batch jobs is studied. The appendices contain a glossary of notation queueing network theorems, test of the random number generator used in simulation, and a-, annotated bibliography.