Heuristic Task Scheduling in Heterogeneous Multiprocessor Systems for Power Optimization

International Journal of Computer Science and Technology | Vol. 12, No. 3, March 2010 | pp. 17–24

DOI: 10.46882/2010/IJCST/000024

Article Type: Original Research Paper

Title: Heuristic Task Scheduling in Heterogeneous Multiprocessor Systems for Power Optimization

Names of Authors: Dr. Seyed M. Mir¹, Elena K. Petrova²

Authors’ Affiliations: ¹Department of Computer Engineering, Sharif University of Technology, Tehran, Iran; ²School of Computer Science and Engineering, Sofia University, Sofia, Bulgaria

Abstract: Energy consumption in modern heterogeneous multiprocessor systems-on-chip (MPSoCs) has escalated significantly, requiring power-aware task allocation scheduling. This paper describes an energy-efficient allocation strategy titled Dynamic Power-Heuristic Scheduler (DP-HS) designed for real-time applications running on asymmetric processing architectures. DP-HS measures core frequencies alongside specific core power dissipation indices to allocate incoming thread demands without causing local thermal choke points. The software framework was verified experimentally across simulated ARM big.LITTLE architectures running various multi-threaded computing benchmarks. The data shows that DP-HS limits total core power requirements by up to 19.4% relative to standard Linux kernel completely fair schedulers (CFS). Additionally, miss-rate cycles for critical high-priority task deadlines dropped by 11.2% under peak execution pressures. We demonstrate that mapping application tasks utilizing hardware-specific power parameters is essential for high-performance mobile devices balancing performance bounds against thermal constraints.

Keywords: Task Scheduling, Heterogeneous Multiprocessors, Power Optimization, Dynamic Voltage Scaling, Real-Time Systems, MPSoC

Manuscript Timeline: Received: August 14, 2009; Revised: October 22, 2009; Accepted: November 15, 2009; Published: March 5, 2010

Citation: Mir, S. M., & Petrova, E. K. (2010). Heuristic Task Scheduling in Heterogeneous Multiprocessor Systems for Power Optimization. International Journal of Computer Science and Technology, 12(3), 17–24.

Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read