Online learning of timeout policies for dynamic power management

Khan, Umair Ali; Rinner, Bernhard

doi:10.1145/2529992

Cited by 28 publications

(21 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…And DPM can be classified into timeout, predicative, stochastic, and machine-learning approaches. Timeout strategy 2,8 switches a device to a low-power state after it has been idle for a certain time period. Timeout strategy is easy to implement, but the drawback of this strategy is that it wastes power while waiting for the timeout to expire.…”

Section: Dpmmentioning

confidence: 99%

“…However, solving the stochastic optimization problem to find the optimal DPM strategy with linear programming is a complex process. Machine-learning strategies 3,8 apply machine learning to learn the request arrival patterns for DPM; they perform well under various workload conditions. But these strategies require offline data collection and training of a classifier that is a time-consuming and complex work.…”

Section: Dpmmentioning

confidence: 99%

“…The power-aware sensor model is similar to the system power model in the ACPI, which has been studied in Benini et al 1 The previous works on constructing DPM strategy are based on analytical approaches, such as linear programming 1,7 and reinforcement learning-based methods. 3,8 The analytical approaches require a mathematical modeling of the system, which is an intricate process. Thus, their complexity results in poor usability that impedes them from practical usage.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Synthesizing power management strategies for wireless sensor networks with UPPAAL-STRATEGO

Dai

Mei

Guo

2017

International Journal of Distributed Sensor Networks

View full text Add to dashboard Cite

Effective power management has become a key concern in the design of wireless sensor networks. Dynamic power management refers to strategies which selectively switch between several power states of a device during the runtime in order to achieve a tradeoff between power consumption and performance. In this article, we present a novel methodology that exploits current model-checking technology for automatic synthesis for dynamic power management. The generic system model for dynamic power management is modeled as a network of timed games. And the synthesis objectives are expressed as synthesis queries. Subsequently, automatic synthesis of power management strategies is performed by UPPAAL-STRATEGO with respect to the synthesis queries. Once a strategy has been constructed, its performance can be analyzed through statistical model-checking using the same tool. The modeling and synthesizing procedures are illustrated with a running example. Finally, the applicability of the methodology is assessed by synthesizing and evaluating a range of power management strategies for a concrete sensor node. Our methodology can be employed to help designers in constructing dynamic power management strategies for wireless sensor networks in practical applications.

show abstract

Section: Dpmmentioning

confidence: 99%

Section: Dpmmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Synthesizing power management strategies for wireless sensor networks with UPPAAL-STRATEGO

Dai

Mei

Guo

2017

International Journal of Distributed Sensor Networks

View full text Add to dashboard Cite

show abstract

“…Based on this a thermal aware scheduling approach is proposed to reduce the temperature of the system at run-time. Apart from these works, there are other studies to reduce the power consumption of a multicore system by scaling the hardware frequency dynamically [Dhiman and Rosing 2009;Javaid et al 2011;Ye and Xu 2014;Khan and Rinner 2014]. However, as shown in [Faruque et al 2010], these approaches cannot guarantee to minimize a system's thermal overhead effectively for all applications.…”

Section: Related Workmentioning

confidence: 99%

Adaptive and Hierarchical Runtime Manager for Energy-Aware Thermal Management of Embedded Systems

Das

Al-Hashimi

Merrett

2016

ACM Trans. Embed. Comput. Syst.

View full text Add to dashboard Cite

Modern embedded systems execute applications, which interacts with the operating system and hardware differently depending on type of workload. These cross-layer interactions result in wide variations of chipwide thermal profile. In this paper, a reinforcement learning-based run-time manager is proposed that guarantees application-specific performance requirements and controls the POSIX thread allocation and voltage/frequency scaling for energy-efficient thermal management. This controls three thermal aspectspeak temperature, average temperature and thermal cycling. Contrary to existing learning-based run-time approaches that optimize energy and temperature individually, the proposed run-time manager is the first approach to combine the two objectives, simultaneously addressing all three thermal aspects. However, determining thread allocation and core frequencies to optimize energy and temperature is an NP-hard problem. This leads to an exponential growth in the learning table (significant memory overhead) and a corresponding increase in the exploration time to learn the most appropriate thread allocation and core frequency for a particular application workload. To confine the learning space and to minimize the learning cost, the proposed run-time manager is implemented in a two-stage hierarchy: a heuristic-based thread allocation at a longer time interval to improve thermal cycling, followed by a learning-based hardware frequency selection at a much finer interval to improve average temperature, peak temperature and energy consumption. This enables finer control on temperature in an energy-efficient manner, while simultaneously addressing scalability, which is a crucial aspect for multi-/many-core embedded systems. The proposed hierarchical run-time manager is implemented for Linux running on nVidia's Tegra SoC, featuring four ARM Cortex-A15 cores. Experiments conducted with a range of embedded and cpu intensive applications demonstrate that the proposed run-time manager not only reduces energy consumption by an average 15% with respect to Linux, but also improves all the thermal aspects -average temperature by 14 • C, peak temperature by 16 • C and thermal cycling by 54%.

show abstract

“…Algorithm 1 provides the pseudo-code of the hierarchical RTM, where affinity selection is performed using a greedy heuristic (lines 1-26) and frequency selection is performed using the Q-learning algorithm (lines [28][29][30][31][32][33][34][35], which is the simplified version of the learning algorithm used in our earlier work [3].…”

Section: B Hierarchical Rtm Algorithmmentioning

confidence: 99%

Workload Change Point Detection for Runtime Thermal Management of Embedded Systems

Das

Merrett

Tribastone

et al. 2016

IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst.

View full text Add to dashboard Cite

Abstract-Applications executed on multicore embedded systems interact with system software (such as the OS) and hardware, leading to widely varying thermal profiles which accelerate some aging mechanisms, reducing the lifetime reliability. Effectively managing the temperature therefore requires (1) autonomous detection of changes in application workload and (2) appropriate selection of control levers to manage thermal profiles of these workloads. In this paper we propose a technique for workload change detection using density ratio-based statistical divergence between overlapping sliding windows of CPU performance statistics. This is integrated in a run-time approach for thermal management, which uses reinforcement learning to select workload-specific thermal control levers by sampling on-board thermal sensors. Identified control levers override the OS's native thread allocation decision and scale hardware voltage-frequency to improve average temperature, peak temperature and thermal cycling. The proposed approach is validated through its implementation as a hierarchical run-time manager for Linux, with heuristic-based thread affinity selected from the upper hierarchy to reduce thermal cycling and learningbased voltage-frequency selected from the lower hierarchy to reduce average and peak temperatures. Experiments conducted with mobile, embedded and high performance applications on ARM-based embedded systems demonstrate that the proposed approach increases workload change detection accuracy by an average 3.4x, reducing the average temperature by 4-25• C, peak temperature by 6-24• C and thermal cycling by 7-35% over stateof-the-art approaches.

show abstract

Online learning of timeout policies for dynamic power management

Cited by 28 publications

References 41 publications

Synthesizing power management strategies for wireless sensor networks with UPPAAL-STRATEGO

Synthesizing power management strategies for wireless sensor networks with UPPAAL-STRATEGO

Adaptive and Hierarchical Runtime Manager for Energy-Aware Thermal Management of Embedded Systems

Workload Change Point Detection for Runtime Thermal Management of Embedded Systems

Contact Info

Product

Resources

About