The Design of Energy-Aware Task Scheduling Algorithms for Mixed Task Sets in Multicore Real Time Systems

Modern IC design techniques and wireless network techniques drive explosive developments in the embedded computing, personal computing, and pervasive computing etc. Many “smart” products, e.g., smart phones, game boxes, vehicle information systems, are also proposed in an amazing time-to-market speed. To satisfy the quality of services for these products, chip makers have adopted the multicore architecture in their microprocessor designs. The power consumption and thermal runaway issues, therefore, become more and more important. Many conventional researches have proposed the task scheduling algorithms with the dynamic voltage scaling mechanism to solve the energy consumption problem. Most of them, however, focus on scheduling periodic tasks. In most real time systems, in fact, periodic tasks and aperiodic tasks will exist simultaneously. Only a few studies have addressed this kind of mixed task set, but their discussions were limited in the single-core processor. Even in the typical discussion about the mutiprocessors or muticore processors, to the best of our knowledge, it still lacks well-designed scheduling algorithms. In this project, we will focus on designing an energy-saving scheduling algorithm for mixed task sets in multicore systems. Particularly, we will address two kinds of dynamic voltage scheduling architeecture: the first one assumed that the voltage of each core can be scaled independently, and the second one assumed that at any time all cores should run in a common voltage. We will derive all of these scheduling problems to the integer-linear programming models. Also, we will design the dynamic scheduling algorithms for runtime execution environment. By realizing this project, we expect that all members joining this project can learn more knowledge and experiences on low-power multicore processor designs.

Project ID:PB10108-2806
External Project ID:NSC101-2221-E182-067