Search

Task scheduling in parallel computing is represented by a directed acyclic graph, and it is called multiprocessor task scheduling. The primary objective of task scheduling algorithm is to minimize the total execution time i.e. scheduling length. There are various types of task scheduling algorithms: Bound Number of Processor Scheduling, and Arbitrary Processor Network Scheduling Algorithms. Bound Number of Processor Scheduling algorithms are used for limited number of processors; since it does not use task duplication, link contention, and routing strategy for communication. But Arbitrary Processor Network scheduling algorithms are considered for link contention routing strategy in communication. It uses arbitrary network topology. In this paper, we have studied different Arbitrary Processor Network scheduling algorithms: Mapping Heuristic, Dynamic Level Scheduling, Bottom Up, and Bubble Scheduling and Allocation. The primary objective of this survey paper is to explain various Arbitrary Processor Network scheduling algorithms and their functionality, priority attributes, and time complexity. The performance metrics can also be used for comparison of different Arbitrary Processor Network scheduling algorithms. [ABSTRACT FROM AUTHOR]

A multiprocessor system is a computer system with more than one processor. Task scheduling on a multiprocessor environment is an important area of research as it has a number of applications in scientific and commercial problems. The objective of task scheduling is to minimize the total completion time of a given application program that is represented by Directed Acyclic Graphs (DAGs). Task scheduling may be classified into static task scheduling and dynamic task scheduling. In this paper, we have studied various types of task scheduling and their properties. Also, we have studied performance metrics for task scheduling. [ABSTRACT FROM AUTHOR]