Abstract

Static Scheduling is the mapping of a program to the resources of a parallel system in order to minimize the execution time. This paper presents static scheduling algorithms that schedule an edge-weighted directed acyclic graph (DAG) to a set of homogeneous processors. The aim is to evaluate and compare the performance of different algorithms and select the best algorithm amongst them. Various BNP algorithms are analyzed and classified into four groups - Highest Level First Estimated Time (HLFET), Dynamic Level Scheduling (DLS), Modified Critical Path (MCP) and Earliest Time First (ETF). Based upon their performance considering various factors, best algorithm is determined.