List Scheduling Algorithm Based on Pre-Scheduling for Heterogeneous Computing

Abstract

Efficient application scheduling algorithms are important for heterogeneous computing systems to achieve high performance. List scheduling is a type of classical and widely used algorithm for static application scheduling, with low time complexity but limited result optimization. In this paper, we proposed a novel pre-scheduling-based list scheduling (PSLS) algorithm for heterogeneous computing systems. The PSLS algorithm has the same low time complexity with classical list scheduling algorithms HEFT and PEFT, but significantly reduces the makespans of applications. The PSLS algorithm consists of three phases: pre-scheduling phase, task prioritizing phase and processor selection phase. The simulation results based on randomly generated graphs and graphs of real-world applications show that the PSLS algorithm has obvious optimizations in terms of schedule length, efficiency and robustness compared with classical list scheduling algorithms.