Multi-Processor System
HOME : RESEARCH : Multi-Processor System
 Power/thermal-aware scheduling for real-time tasks in multi-core systems
1. Goal
  • Determine the speed assignments on each CPU core and task mapping to minimize the energy consumption
  • 2. Constraints
  • The peak temperature must not exceed the temperature threshold.
  • Tasks have various power consumption characteristics
    (i.e., effective frequency factor).
  • 3. Motivation example
  • “Balanced” scheduling is not thermally safe
  • “Temperature-aware” scheduling is more thermal/energy optimal than “balanced” scheduling
  • 4. Experimental environment
  • LITMUS-RT-based real-time scheduling testbed
  • 5. Published papers
  • Changha Lee, Young-Ho Lee, and Jihong Kim, "Thermal-Aware Slack Distribution for Real-Time Systems," International Symposium on Low-Power and High-Speed Chips (COOL Chips XIV), Yokohama, Japan, April 20-22, 2011
  • 이창하, 송욱, 김지홍, "멀티코어 기반의 실시간 시스템을 위한 발열량을 고려한 태스크 할당 기법," 2010 대한임베디드공학회 추계학술대회, pp. 31-34, 2010.11