1971•Unpublished venueRequires access

Throughput analysis of computer systems - multiprogramming versus multiprocessing

Wei Xing Chang, Y. Paternot, J. A. Ray

Open publisher page 2 citations

Abstract

To increase system throughput, multiprogramming and multiprocessing systems are introduced. The first improves the throughput by means of interleaving the individual service requests, and the latter by exploiting the parallelism in the system design. The basic philosophy of multiprogramming is compared with that of multiprocessing to put the concepts in their proper perspective. Simple mathematical models, based on the theory of queues, were used to analyze the throughput capacity of these systems. The throughput of a given multiprogramming system configuration is compared with a number of multiprocessing systems. Within each multiprocessing system, the throughput is maximized by determining the optimum number of processors, and the optimum number of control and supervision programs which are required to control the system. The mathematical analysis is completely substantiated by the GPSS simulation results.

About this research paper

What this paper is about

To increase system throughput, multiprogramming and multiprocessing systems are introduced. The first improves the throughput by means of interleaving the individual service requests, and the latter by exploiting the parallelism in the system design. The basic philosophy of multiprogramming is compared with that of multiprocessing to put the concepts in their proper perspective. Simple mathematical models, based on the theory of queues, were used to analyze the throughput capacity of these systems. The throughput of a given multiprogramming system configuration is compared with a number of multiprocessing systems. Within each multiprocessing system, the throughput is maximized by determining the optimum number of processors, and the optimum number of control and supervision programs which are required to control the system. The mathematical analysis is completely substantiated by the GPSS simulation results.

Why it matters

OpenAlex reports 2 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

To increase system throughput, multiprogramming and multiprocessing systems are introduced. The first improves the throughput by means of interleaving the individual service requests, and the latter by exploiting the parallelism in the system design. The basic philosophy of multiprogramming is compared with that of multiprocessing to put the concepts in their proper perspective. Simple mathematical models, based on the theory of queues, were used to analyze the throughput capacity of these systems. The throughput of a given multiprogramming system configuration is compared with a number of multiprocessing systems. Within each multiprocessing system, the throughput is maximized by determining the optimum number of processors, and the optimum number of control and supervision programs which are required to control the system. The mathematical analysis is completely substantiated by the GPSS simulation results.

Key concepts: Computer multitasking, Multiprocessing, Computer science, Throughput, Parallel computing, Human multitasking, Operating system, Distributed computing

Related papers

Back to paper searchBrowse research topicsOriginal source
Throughput analysis of computer systems - multiprogramming versus multiprocessing — Research Paper | ScholarLens