【作者】 赵晓丽； 陶品； 杨士强；

【Author】 Xiao-Li Zhao+, Pin Tao, Shi-Qiang Yang Computer Science Department, Tsinghua University, Beijing, China

【机构】 清华大学计算机系；

【摘要】 移动设备上的资源非常有限,特别是能量,而现在移动设备上也出现很多消耗能量特别大的应用,如视频编码应用等。所以解决这一问题的技术陆续出现。其中计算迁移技术是一种可行的既能节约能量又能提高应用程序执行效率的方法。在本文中,我们提出了计算迁移技术中几种客户端任务和服务器端任务的分割算法,对这些算法的性能进行比较,并且分析了这些算法应用到H.264编码上的可能性。更多还原

【Abstract】 The resources in mobile devices are very limited, especially energy resource, and now some energy-killer applications run on the mobile devices like video encoder. So some technologies to deal with the usage of resources are proposed. Computation offloading is proved to be a very efficient method to save energy resource and improve the quality of applications. In this paper, we focus on the task partition algorithms in computation offloading scheme. Take partition algorithms partition the application to two parts: one part runs on mobile device, another one runs on the powerful computer around the mobile device. So the load in mobile device reduces and the usage of resources reduces too. We propose five task-partition algorithms-----EDM, CSC, CCSC, JCCSC, TS, do some experiments on them, compare them with the result and execute time. At last, we analyze the complexity of H.264 video encoder compared to other encoder and realize that using the computation offloading to H.264 encoder is very necessary. In future, we will complete the computation offloading system and use it to some complexity applications like H.264 encoder.更多还原