透過您的圖書館登入
IP:3.147.103.8
  • 學位論文

虛擬平台上針對智慧型手機程式之能源消耗估計與節能

Power Estimation and Energy Saving of Smartphone Applications on Virtualized Platforms

指導教授 : 洪士灝

摘要


在現今快速成長的移動通訊系統市場中,因為其有限的電池容量,如何節省其使用的能源消耗已成為硬體設計與軟體開發者的一項主要挑戰。為提昇產品競爭力,能夠盡早去估計其系統的能源消耗狀況並加入節能技術更是最主要的關鍵。在系統開發初期,虛擬平台不僅能幫助硬體設計者去驗證其平台的功能正確性,還能提供軟體開發者一個在硬體平台仍尚未就緒前的軟體測試環境。 在這篇論文中,我們提出一套虛擬效能評估框架,它加強了功能性模擬器來幫助軟硬體開發者去分析其目標系統上的效能表現與電源消耗。在此框架下,我們加入了兩種不同的能耗模型:處理器為主的能耗模型主要旨在分析於處理器上使用動態電壓與頻率調整所造成的影響,而週邊裝置的能耗模型則專注於全體系統的能源估計。在實驗中我們展現了這個虛擬效能評估框架可以很好的反應真實智慧型手機上的效能表現與電源消耗行為,如此一來便能有效地幫助軟硬體開發者去減少他們在這些系統上的能源消耗。

並列摘要


Energy-saving has become a major challenge for both hardware designers and software developers of mobile systems due to limited battery capacities. Analysis of the power consumption in such a system and employment of energy-saving techniques as early as possible is key to competitive products design. In the early design stage, virtualized platforms may help hardware designers in verifying the functional correctness of a design, as well as provide an environment for software developers to test their applications before the hardware is ready. In this thesis, we propose a framework, called Virtual Performace Evaluation (VPE) framework, which augments functional emulators to assist hardware and software developers in analyzing the performance and power of a target system. For power modeling, we provide two models: the processor-only model aims at the analysis of dynamic voltage and frequency scaling effects on the processor cores and the peripherals model targets at the power consumption of a full system. Our experimental results show that our VPE framework can reflect the performance and power behavior of smartphones and help hardware/software designers in reducing the energy consumption on these systems.

參考文獻


[17] S. Li, J. H. Ahn, R. D. Strong, J. B. Brockman, D. M. Tullsen, and N. P. Jouppi, “McPAT:
Scott, “Energy-efficient processor design using multiple clock domains with dynamic
voltage and frequency scaling.”
project - from VLIW DSP processor to multicore computing platform.”
2008, pp. 188–191.

延伸閱讀