Cache performance tuning tools are conducive to develop program with good locality and fully use cache to decrease the influence caused by speed gap between processor and memory. This paper introduces the design and i...Cache performance tuning tools are conducive to develop program with good locality and fully use cache to decrease the influence caused by speed gap between processor and memory. This paper introduces the design and implementation of a cache performance tuning tool named CTuning, which employs a source level instrumentation method to gather program data access information, and uses a limited reuse distance model to analyze cache behavior. Experiments on 183.equake improve average performance more than 6% and show that CTuning is proficient not only in locating cache performance bottlenecks to guide manual code transformation, but also in analyzing cache behavior relationship among variables, thus to direct manual data reorganization.展开更多
程序最坏执行时间(Worst Case Execution Time,WCET)是嵌入式实时系统时间属性验证的基础,在采用静态分析技术的WCET估算中需要分阶段对不同的执行环境约束条件进行分析,并整合所有约束信息、结合程序控制流结构估算全局最坏路径,因此...程序最坏执行时间(Worst Case Execution Time,WCET)是嵌入式实时系统时间属性验证的基础,在采用静态分析技术的WCET估算中需要分阶段对不同的执行环境约束条件进行分析,并整合所有约束信息、结合程序控制流结构估算全局最坏路径,因此各阶段分析的中间结果对最终的WCET估算性能具有较大影响.在现代嵌入式系统中,硬件平台中的Cache机制成为对执行时间影响较大的硬件体系结构,对其进行精确的行为分析在WCET估算中具有重要的现实意义.采用抽象解释理论对Cache行为进行分析已有较为成熟的技术成果和相关工具,但由于静态分析技术具有较难理解和使用的特点,对于技术没有覆盖、工具没有支持的硬件架构,针对这类硬件架构进行相关研究和验证工具开发都具有较大难度和挑战.该文以抽象解释为理论基础,以复用Cache分析过程为目标,提出了基于抽象解释的模块化Cache行为分析框架,对Cache行为分析过程进行了层次划分,提出了易于复用的Cache行为分析方法设计,能够针对不同架构的Cache机制分析方法进行建模,并以统一的分析框架对分析过程进行复用.案例实验表明,该框架可支持采用抽象解释对使用LRU策略的Cache行为进行建模分析,并能够得到Cache命中情况标记信息以支持后续WCET的估算过程.展开更多
The hit rate, a major metric for evaluating proxy caches, is mostly limited by the replacement strategy of proxy caches. However, in traditional proxy caches, the hit rate does not usually successfully predict how w...The hit rate, a major metric for evaluating proxy caches, is mostly limited by the replacement strategy of proxy caches. However, in traditional proxy caches, the hit rate does not usually successfully predict how well a proxy cache will perform because the proxy cache counts any hit in its caching space which has many pages without useful information, so its replacement strategy fails to determine which pages to keep and which to release. The proxy cache efficiency can be measured more accurately using the valid hit rate introduced in this paper. An efficient replacement strategy based on the Site Graph model for WWW (World Wide Web) documents is also discussed in this paper. The model analyzes user access behavior as a basis for the replacement strategy. Simulation results demonstrate that the replacement strategy improves proxy cache efficiency. 展开更多
基金Sponsored by the National Natural Science Foundation of China (No.60573141, 60773041)National 863 High Tech- nology Research Program of China (No.2007AA01Z404, 2007AA01Z478)+2 种基金High Technology Research Programme of Jiangsu Province (No.BG2006001)Key Laboratory of Information Technology Processing of Jiangsu Province (kjs06006)Project of NJUPT (NY207135)
文摘Cache performance tuning tools are conducive to develop program with good locality and fully use cache to decrease the influence caused by speed gap between processor and memory. This paper introduces the design and implementation of a cache performance tuning tool named CTuning, which employs a source level instrumentation method to gather program data access information, and uses a limited reuse distance model to analyze cache behavior. Experiments on 183.equake improve average performance more than 6% and show that CTuning is proficient not only in locating cache performance bottlenecks to guide manual code transformation, but also in analyzing cache behavior relationship among variables, thus to direct manual data reorganization.
文摘程序最坏执行时间(Worst Case Execution Time,WCET)是嵌入式实时系统时间属性验证的基础,在采用静态分析技术的WCET估算中需要分阶段对不同的执行环境约束条件进行分析,并整合所有约束信息、结合程序控制流结构估算全局最坏路径,因此各阶段分析的中间结果对最终的WCET估算性能具有较大影响.在现代嵌入式系统中,硬件平台中的Cache机制成为对执行时间影响较大的硬件体系结构,对其进行精确的行为分析在WCET估算中具有重要的现实意义.采用抽象解释理论对Cache行为进行分析已有较为成熟的技术成果和相关工具,但由于静态分析技术具有较难理解和使用的特点,对于技术没有覆盖、工具没有支持的硬件架构,针对这类硬件架构进行相关研究和验证工具开发都具有较大难度和挑战.该文以抽象解释为理论基础,以复用Cache分析过程为目标,提出了基于抽象解释的模块化Cache行为分析框架,对Cache行为分析过程进行了层次划分,提出了易于复用的Cache行为分析方法设计,能够针对不同架构的Cache机制分析方法进行建模,并以统一的分析框架对分析过程进行复用.案例实验表明,该框架可支持采用抽象解释对使用LRU策略的Cache行为进行建模分析,并能够得到Cache命中情况标记信息以支持后续WCET的估算过程.
基金the State High- Tech Developments Planof China (No.86 3- 30 6 - ZT0 1- 0 3- 1) IBM China Research Lab Huawei Enterprise Funding onScience and Technology
文摘The hit rate, a major metric for evaluating proxy caches, is mostly limited by the replacement strategy of proxy caches. However, in traditional proxy caches, the hit rate does not usually successfully predict how well a proxy cache will perform because the proxy cache counts any hit in its caching space which has many pages without useful information, so its replacement strategy fails to determine which pages to keep and which to release. The proxy cache efficiency can be measured more accurately using the valid hit rate introduced in this paper. An efficient replacement strategy based on the Site Graph model for WWW (World Wide Web) documents is also discussed in this paper. The model analyzes user access behavior as a basis for the replacement strategy. Simulation results demonstrate that the replacement strategy improves proxy cache efficiency.
