Comparative Analysis Of BNP Scheduling Algortihms

Abstract

Parallel Computing is the procedure of creating an environment where requirement of the job is equated with the available resources in system. Multiprocessor task scheduling is most important and very crucial issue in design of homogeneous parallel systems. This paper covers the study of scheduling algorithms related to Bounded Number of Processors (BNP) class of multiprocessor scheduling algorithm represented by a Directed Acyclic graph (DAG). An attempt has been made to evaluate their performance on basis of Processor Utilization, SpeedUp and Scheduled Length Ratio (SLR). Analysis of the performance has proved Dynamic Level Scheduling (DLS) as a better algorithm in homogeneous environment.

Keywords: Directed Acyclic Graph (DAG), Dynamic Level Scheduling (DLS), Homogeneous Environment, Multiprocessor Scheduling algorithm, Parallel Computing, Multiprocessor task scheduling, Processor Utilization, Scheduled Length Ratio (SLR), SpeedUp.