智慧型個別工作耗電量分析工具

An Intelligent Task-Based Power Consumption Measurement Tool

曾信雄(Hsin-Hsiung Tzeng)

指導教授 ： 薛智文

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2008年)

https://doi.org/10.6342/NTU.2008.01093

摘要

現今的嵌入式裝置被廣泛的運用於日常生活中，由於半導體產業的高成長率，處理器的速度也變的越來越快；但是處理器速度越快，相對的能耗量也就越大。因此能耗量測成為了嵌入式系統裡面一個重要的議題。模擬量測可以提供使用者非常細微的分析結果，但是在分析過程當中所要花的時間太長；傳統的實體量測可以提供一個快速且真確的量測結果給使用者，但是往往無法觀察出系統在執行時的情形；在這篇論文中，我們在一個別工作量測工具上延伸了觀測硬體之功能，並且提出了一個方法對周邊輸入輸出裝置作觀察。相信這套能耗量測分析工具的實驗方法能對此工具的開發者與使用此工具之使用者有所幫助。

關鍵字

嵌入式系統 ； 實體量測 ； 模擬量測

並列摘要

Energy consumption measurement is very important for developers of modern embedded systems. Basically, there are two types of measurement methods: the simulation-based and the physical-based. For energy consumption analysis, the simulation-based measurement is slow and inaccurate. Though the typical physical-based measurement is fast and accurate, while it can not get detail system information. Therefore, we propose a task-based measurement that bases on the physical-based measurement to provide system information. However, previous works can not use system information to derive more detailed hardware dependent information well, such as the power consumption of CPU, memory, I/O devices, etc. In this thesis, we propose a novel idea to extract the power consumption of I/O devices, and verify our work through experiments on DMA. Experiments show that we can use our method to have an acceptable result. We believe this thesis is helpful for not only users who would like to measure the energy consumption of applications, but also provide a fast approach to developers who would like to get more hardware dependent information.

並列關鍵字

embedded system ； energy consumption ； simulation-based ； physical-based ； task-based

參考文獻

[2] J. Flinn and M. Satyanarayanan. Powerscope: A tool for profiling the energy usage of mobile applications. In Proceedings of IEEE Workshop Mobile Comput. Syst. Applicat., pages 2–10, Feb 1999.

[3] Chuling Hu, Danial A. Jimenez, and Ulrich Kremer. Toward an evaluation infrastructure for power and energy optimizations. In proceedings of the 19th IEEE Internationl Parallel and Distributed Processing Symposium, 2005.

[7] A. Sinha and A. P. Chandrakasan. Jouletrack - a web based tool for software energy profiling. In Design Automation Conference, pages 220-225, 2001.

[10] Bokyung Wang and Suresh Singh. Computational energy cost of tcp. In INFOCOM 2004. Twenty-third AnnualJoint Conference of the IEEE Computer and Communications Societies, volume 2, pages 785–795, March 2004.

[13] W. Ye, N. Vijaykrishnan, M. Kandemir, and M. J. Irwin. The design and use of simplepower: a cycle-accurate energy estimation tool. In Proceedings of the 37th conference on Design automation, pages 340–345, June 2000.