TARAS

Group Brief

TARAS &ndash We are a collaborative research group that focuses on scheduling real-time tasks on thermally constrained systems. Both uniprocessor and multicore systems are considered. The goals are to maximize system lifetime and work completed. Both classical optimization techniques and heuristic approaches are being used to solve the problem.

Participants

                   University of Notre Dame                    Xiaobo Sharon Hu
                                                                                   Tam Chantem
                   University of Michigan                         Robert Dick
                                                                                   Yun Xiang

Projects

This collaborative research effort attempts to develop algorithms for temperature-constrained embedded real-time systems on both uniprocessors and multicore systems. Due to shrinking device sizes and increasing transistor counts, most chips now have high power density, temperature, and hence shortened lifetime. Since it is expensive and impractical to design packages and cooling solutions to handle worst-case thermal profiles, we must rely on temperature-aware design techniques to reduce system temperature. We propose several algorithms to minimize or constrain processor temperature, both optimal and heuristic.

Temperature-Aware Scheduling and Assignment. We used mixed integer linear programming to optimally solve the temperature-aware real-time MPSoC assignment and scheduling problem and presented an efficient and flexible heuristic framework that allows for the trade off between accuracy and running time [1, 4].
Online Work Maximization for Temperature-Constrained Systems. In this work, we used dynamic voltage and frequency scaling (DVFS) to maximize work completed over a time interval. We showed that our algorithm is optimal and can handle non-negligible transition overheads [3].
Reliability-Aware Scheduling and Assignment. We presented a system-level reliability analysis tool that considers thermal traces to accurately predict system lifetime and failure probabilities. We are in the process of devising an efficient online task assignment and scheduling algorithms that directly maximizes the lifetime of multicore systems [2].

References

Thidapat Chantem, X. Sharon Hu, and Robert P. Dick, Temperature-Aware Scheduling and Assignment for Hard Real-Time Applications on MPSoCs, IEEE Transactions on Very Large Scale Integration Systems (TVLSI), to appear.
Yun Xiang, Thidapat Chantem, Robert P. Dick, X. Sharon Hu, and Li Shang, System-Level Reliability Modeling for MPSoCs, In Proceedings of the International Conference on Hardware-Software Codesign and System Synthesis (CODES+ISSS), pages 297--306, October 2010.
Thidapat Chantem, X. Sharon Hu and Robert P. Dick, Online Work Maximization under a Peak Temperature Constraint, In Proceedings of the International Symposium on Low Power Electronics and Design (ISLPED), pages 105-110, August 2009.
Thidapat Chantem, Robert P. Dick and X. Sharon Hu, Temperature-Aware Scheduling and Assignment for Hard Real-Time Applications on MPSoCs, In Proceedings of the Design, Automation, & Test in Europe (DATE), pages 288-293, March 2008.