Adaptive dynamic power management for hard real-time pipelined Multiprocessor Systems

Chen, Gang Chen; Huang, Kai; Knoll, Alois

doi:10.1109/rtcsa.2014.6910525

Cited by 5 publications

(9 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Chen et al studied energy optimization by combining DVFS (Dynamic Voltage and Frequency Scaling) and DPM (Dynamic Power Management) technologies for real-time systems in [21]. They also explored how to apply DPM in an adaptive manner to optimize leakage power consumption for pipelined multi-core systems under deadline constraints in [1]. As aforementioned, an effective power management may not be suitable for thermal management.…”

Section: Related Workmentioning

confidence: 99%

“…For the aforementioned system transformation, the EPBOO principle in [26] and the theory in [1] can calculate the aggregate service curve provided for stream S i for scenarios in which every stage provides a service curve in rate-latency format. However, under APTM, the service curve of each stage is no longer rate-latency.…”

Section: B Real-time Constraintsmentioning

confidence: 99%

“…While shifting processors towards more parallelism, multi-core architectures raise a serious question: how to effectively extract and utilize the high degree of parallelism for applications. The pipelined architecture, which, in principle, can increase the throughput by performing multiple operations simultaneously, is believed to be one of the promising computing paradigms of multi-core systems [1]. A pipelined architecture consists of a set of processing components connected in series and thus achieves parallel computing by handling sub-tasks simultaneously on the components.…”

Section: Introductionmentioning

confidence: 99%

“…Thus, it could lead to pessimistic results due to the runtime variability of event arrivals and execution times. Gang et al proposed an adaptive online Dynamic Power Management approach called Balanced Workload Scheme (BWS) in [1] for hard real-time pipelined systems. However, BWS cannot be directly shifted to temperature optimization problem since it is designed for energy saving.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Adaptive Periodic Thermal Management for Pipelined Hard Real-Time Systems

et al. 2019

Self Cite

View full text Add to dashboard Cite

This paper proposes a novel dynamic thermal management approach named Adaptive Periodic Thermal Management (APTM) for multi-core pipelined processors with hard real-time constraints. During online execution, the approach optimizes the peak temperature of the processor by dynamically calculating and deploying new adaptive periodic thermal management schemes, which periodically switch the stages to sleep state until next adapting. At each such instant of adapting, new APTM schemes which optimize the peak temperature under real-time constraints are calculated according to the extended pay-burst-only-once principle. To reduce the overhead of online computation, our approach utilizes thermal properties obtained from offline experiments in searching the ATPM scheme minimizing the temperature. We evaluate our APTM with simulation on the Intel SCC chip as well as real experiments on an Intel I7 processor. Compared to existing approaches, our approach reduces the peak temperature by up to 8 K for the Intel I7 processor, and 7.5 K for the Intel SCC using four to sixteen processor cores.INDEX TERMS Pay-burst-only-once, real-time systems, temperature optimization, thermal management.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: B Real-time Constraintsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Adaptive Periodic Thermal Management for Pipelined Hard Real-Time Systems

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…Firstly, we investigate the effectiveness of MCFTP in managing the temperature via the configuration manipulation interface. Then, we implement two existing thermal managements in MCFTP, one is proposed in [7], namely O-PBOO, and the other policy is presented in [6], which is termed as BWS. Finally, we report the running overhead of MCFTP on two platforms that have different computing power abilities.…”

Section: Experimental Evaluationmentioning

confidence: 99%

McFTP: A Framework to Explore and Prototype Multi-core Thermal Managements on Real Processors

Cheng

Zhao

Huang

et al. 2017

2017 IEEE Trustcom/BigDataSE/Icess

Self Cite

View full text Add to dashboard Cite

Nowadays, multi-core processor architectures have been widely adopted in main domains e.g., embedded, general-purpose, realtime systems, etc. Diverse thermal managements have been proposed to manage the temperature under various constraints. This has made the selection of the right thermal management policy difficult. Designers need to validate any resource distribution decision in design phase on the target architecture, e.g., by using a re-configurable thermal framework running in the user-space. In this paper, we first analyze the requirements that such a framework should satisfy. Then, we propose McFTP: a thermal framework fulfilling all the requirements. For this purpose, an intermediate interface is defined to isolate thermal management policies from the low-level implementations. A set of commonly used temperature control mechanisms are implemented as a library which can be accessed via the interface. With these features, McFTP can not only implement a thermal management policy at high-level of abstraction, but also execute the user-defined task-set for real thermal evolution. We demonstrate the effectiveness and efficiency of McFTP by implementing it with two works in the literature on a Dell hardware platform.

show abstract