并行计算平台上数据索引技术研究.pdfVIP

摘 要 实时性。同时，我们将其应用于高性能绿色计算平台KD-60，使之在一定 的程度上实现了搜索的高效能。 3. 基于并行计算平台的时间序列索引。提出一种基于并行计算平台的时间序 列索引，并针对该索引结构提出相应的并行功耗分析模型。该方案在解决 时间序列索引问题的同时，打破了传统索引方式只注重索引效率而忽视索 引效能的单一思路，实现了系统索引过程的高效能低功耗。实验证明，基 于并行计算平台的时间序列索引很好的解决了时间序列索引的性能问题， 在索引的功耗方面也有良好表现。 关键词：并行计算平台；密集数据；一体化研究方法；索引；高效能；功耗 Abstract Abstract With the popularity of computer and network size continues to expand, an explosion of digital information growth. The generation, dissemination, collection and query become the worlds most basic life needs. How intensive data information for the user faster and more accurate index of information become an urgent problem. At the same time, human society needs high-performance computing applications also continue to promote the popularization of parallel computing platforms and improved. The main method of data retrieval is to build a variety of suitable index, and by the relative efficiency of the index to retrieve various types of data. However, with the retrieval of data and increase the size of user queries, information retrieval systems and processing capabilities to enhance the scale of need for a good application of parallel computing platform to solve. So how to make use of parallel computing platform to improve information retrieval system performance, improve system capacity, as important issues in this field. With the development of parallel computing platform, the index used in parallel computing platform is an inevitable trend. In this paper, parallel computing, Architecture - Algorithms - Programming - Application of the integration of research methods, focusing on how to solve the intensive parallel