A reformed task scheduling algorithm for heterogeneous distributed systems with energy consumption constraints

• Published in: Neural computing & applications Vol. 32; no. 10; pp. 5681 - 5693
• Main Authors: Hu, Yikun, Li, Jinghong, He, Ligang
• Format: Journal Article
• Language: English
• Published: London Springer London 01.05.2020; Springer Nature B.V
• Subjects: Advances in Parallel and Distributed Computing for Neural Computing; Algorithms; Artificial Intelligence; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer networks; Computer Science; Data centers; Data Mining and Knowledge Discovery; Energy conservation; Energy consumption; Energy management; Image Processing and Computer Vision; Probability and Statistics in Computer Science; Schedules; Task scheduling; Task scheduling; Energy constraint; Parallel computing; Heterogeneous computing; Distributed computing
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-019-04415-2

As the scale increases and performance improves, the energy consumption of high-performance computer systems is rapidly increasing. The energy-aware task scheduling for high-performance computer systems has become a hot spot for major supercomputing centers and data centers. In this paper, we study the task scheduling problem to minimize the schedule length of parallel applications while satisfying the energy constraints in heterogeneous distributed systems. The existing approaches mainly allocate unassigned tasks with minimal energy consumption which cannot achieve optimistic scheduling length in most cases. Based on this situation, we propose a reformed scheduling method with energy consumption constraint algorithm, which is based on an energy consumption level to pre-allocate energy consumption for unassigned tasks. The experimental results show that compared with the existing algorithms, our new algorithm can achieve better scheduling length under the energy consumption constraints.