基于重用距离分析的多虚拟机缓存划分系统-计算机系统结构专业论文.docxVIP

华 华 中 科 技 大 学 硕 士 学 位 论 文 II II Abstract With popularization of multi-core technology, parallel computing model is used to improve processing ability of the whole computer system. However, multi-core technology also brings serious shared cache contention problem, which leads to cache hits decrease and performance degradation of the co-scheduled tasks. Especially in the high performance computing virtualized environment, cache resource contention breaks down the isolation mechanism of virtualization, and causes computing performance degradation and the loss of safety and reliability. A cache partition system in virtualized environment based on reuse distance analysis provides memory access addresses sampling and cache miss rate sampling of high performance computing applications. Via the file segmentation and cyclic compress algorithm, it computes reuse distance histogram of each application. Based on reuse distance analysis, this system classifies all applications on cache behavior characteristics. Cache partition mechanism receives virtual resource requests of multiple applications from users, and matches them with the applications’ classification result. With the cooperation of physical nodes, the system completes applications’ resource allocation and cache partition. This system provides cache partition in virtual machine granularity based on reuse distance analysis. It minimizes the influence of cache interference, and enhances applications’ computing performance and quality of service. The experiment result shows that this system significantly improves the performance of cache sensitive applications. It reduces execution time by 10%~21% as well as cache miss rate by 10%~25%, increases throughput by 10%~23%. Key words: Virtualization, High Performance Computing, Cache Partition, Reuse Distance III III 目 录 摘 要 I HYPERLINK \l _bookmark0 Abstract II HYPERLINK \l _bookmark1 1 绪论 HYPERLINK \l _bookmark2 1.1 问题的提出 (1) HYPERLINK \l _bookmark3 1.2