Thermal-Constrained Energy-Aware Partitioning for Heterogeneous Multi-core Multiprocessor Real-Time Systems

Abstract

Next-generation multi-core multiprocessor real-time systems consume less energy at the cost of increased powerdensity. This increase in power-density results in high heatdensity and may affect the reliability and performance of real-time systems. Thus, incorporating maximum temperature constraints in scheduling of real-time task sets is an important challenge. This paper investigates thermal-constrained energyaware partitioning of periodic real-time tasks in heterogeneous multi-core multiprocessor systems. We adopt a power model which considers the impact of temperature and voltage on a processor's static power consumption. Two types of thermal models are used to respectively capture negligible and non-negligible amount of heat transfer among cores. We develop a novel geneticalgorithm based approach to solve the heterogeneous multicore multiprocessor partitioning problem. Extensive simulations were performed to validate the effectiveness of the approach. Experimental results show that integrating a worst-fit based partitioning heuristic with the genetic algorithm can significantly reduce the total energy consumption of a heterogeneous multicore multiprocessor real-time system.

A novel thermal-constrained energy-aware partitioning algorithm for heterogeneous multiprocessor real-time systems

“…The increase in the temperature of a processor due to an increased heat-density has resulted in a recent research interest in temperatureaware multiprocessor scheduling to improve the reliability and performance of homogeneous real-time systems [9], [10], [11]. In addition, in our previous work [12], we studied the thermal-constrained energy-aware multiprocessor scheduling in heterogeneous real-time systems where we developed a genetic-algorithm based approach.…”

“…We adopt a general power model that was first developed by Li et al [4] and a heat-independent thermal model assuming negligible heat transfer among processors [10]. Unlike our previous work [12], this paper introduces a new heuristic, called BaB-A, using a branch-and-bound based partitioning approach. Extensive simulations have been carried out to show the advantages of this new approach (BaB-A) over the best existing approach (i.e., HyWGA algorithm that we developed previously [12]).…”

“…Unlike our previous work [12], this paper introduces a new heuristic, called BaB-A, using a branch-and-bound based partitioning approach. Extensive simulations have been carried out to show the advantages of this new approach (BaB-A) over the best existing approach (i.e., HyWGA algorithm that we developed previously [12]). In addition, we have used a real world test bed Green Server Farm at University of Illinois-Urbana Champaign [13] to validate some of the simulation results and the effectiveness of our new algorithm BaB-A.…”

A Multi-core Testbed on Desktop Computer for Research on Power/Thermal Aware Resource Management

“…In regards to temperature reduction methods there has been much theoretical research done over the years [8,19,42,61,[74][75][76].Yeo et al [75] developed a predictive dynamic thermal management technique for multicore systems. Their work proposes to maintain system temperatures below a desired level by moving and running the application from the possible overheated core to the future coolest core and by reducing the processor resources within multicore systems.…”

On the Design of Real-Time Systems on Multi-Core Platforms under Uncertainty

“…While it is a common practice to use certain time-consuming techniques such as Genetic Algorithm (GA) [105,138] and/or Simulated Annealing Algorithm (SAA) [176,178] to solve this problem, this is not viable for the topology virtualization approach. Since the physical to logical topology mapping is performed when deploying (installing or initiating) the application software on an individual device, the key to the success of this approach is to develop computationally efficient mapping methods that can effectively optimize the performance metrics for application software.…”