Modified Critical Path and Top-Level Attributes (MCPTL)-Based Task Scheduling Algorithm in Parallel Computing

Abstract

The parallel processing provides speedup to the computational tasks usually not solved efficiently by a sequential machine. It is basically, a transformation of an application program into task graph where tasks of the graph are allocated to the available processors. In this paper, we have proposed a new task scheduling algorithm which is a modified version of Critical Path and Top-Level Attributes-based Scheduling (CPTL) algorithm. This new algorithm is known as MCPTL algorithm. The proposed algorithm excludes the communication time between the tasks during allocations onto the processors. The MCPTL algorithm gives minimum scheduling length as compared the CPTL algorithm and also performance study is done with heuristic algorithms which is based on performance metrics, Scheduling Length Ratio (SLR), Speedup (SP), Efficiency (EFF), Load Balancing (LB), and Cost (C).