计算机科学 ›› 2012, Vol. 39 ›› Issue (5): 282-286.

• 体系结构 • 上一篇    下一篇

基于神经网络的独立程序在单机上运行功耗的预测

谭一鸣,曾国荪   

  1. (同济大学计算机科学与技术系 上海200092);(国家高性能计算机工程技术中心同济分中心 上海200092)
  • 出版日期:2018-11-16 发布日期:2018-11-16

Power Analysis for Executable Program on Single Computer Based on Artificial Neural Network

  • Online:2018-11-16 Published:2018-11-16

摘要: 程序运行能耗分析是目前绿色计算、高效能计算的研究热点。由于应用程序自身的复杂性、硬件平台的异构 性、环境因素对功耗影响的不确定性,导致很难直接给出程序运行功耗的预测公式,因此提出一种基于神经网络的程 序运行功耗预测方法。通过构造一个PP神经网络,以影响程序运行功耗的软件、硬件和环境因素为神经网络的输 入,以程序运行功耗和执行时间为神经网络的输出,并选取多个应用领域的典型实用程序来提取相应特征参数,且将 其作为训练样本来训练神经网络,从而得到程序功耗预测模型。重复实验表明,提出的功耗预测方法具有合理性和可 行性。

关键词: 绿色计算,功耗分析,神经网络,可执行程序

Abstract: Power management of application program is a hot research in the area of green computing and high produc- tivity computing. Because of the complexity of application programs,heterogeneity of processors and uncertainty of the running environment, it is difficult to propose an accurate method to predict energy consumption for application program directly. So we presented a power analysis paradigm for application program based on artificial neural network. First, we built a power analysis model based on back propagation neural network(I3PNN). The three factors of software, hard- ware and environment were taken as the inputs of BPNN, and energy consumption and finish time as the outputs of 13PNN. Next, we chose lots of classic application programs from different fields as training samples. After learning and training, an expected I3PNN was obtained which can be used to predict energy consumption for other new programs. Re- pcated experiments show that this power analysis paradigm is rational and feasible.

Key words: Green computing, Power analysis, Artificial neural network, Executable programs

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!