期刊文献+

一个软件可控Cache替换策略的行为分析

Behavior analysis of software-controlled Cache replacement policy
原文传递
导出
摘要 针对传统的仅由硬件控制的Cache替换策略不能得到令人满意Cache利用率的问题,提出了一个软件可控Cache替换策略:最优Cache划分(OCP),并利用Presburger公式给出了OCP策略下的Cache行为精确分析.OCP Cache替换策略简化了Cache行为和Cache失效分析方法.实验结果表明OCP Cache替换策略能有效地降低Cache失效率. Conventional hardware-only Cache replacement policies do not lead to satisfactory Cache utilization. A software-controlled Cache replacement policy is proposed, optimum Cache partition (OCP), which can be carried out through Cache hints and lest recently used (LRU) replacement policy. Presburger arithmetic is used to exactly model the behavior of loop nests under OCP policy. The OCP replacement policy results in plain Cache behaviors, and makes Cache misses analyzing and optimizing easily and efficiently. OCP replacement policy has been implemented in a compiler test-bed and evaluated on a set of scientific computing benchmarks. Initial results show that this approach is effective on reducing the Cache miss rate.
出处 《华中科技大学学报(自然科学版)》 EI CAS CSCD 北大核心 2009年第8期19-22,共4页 Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金 湖南省自然科学基金资助项目(07JJ3121)
关键词 Cache缓存 编译器 编译器优化 软件可控Cache Cache提示 Cache行为分析 Cache storage program compilers compiler optimization software-controlled Cache Cache hint Cache behavior analysis
  • 相关文献

参考文献9

  • 1Kathail V, Schlansker M S, Rau B R. HPL-PD architecture specification: Version 1. 1. Technical Re- port HPL-93-80[R]. [s. l.] : Hewlett-Packard Laboratories USA, 2000.
  • 2Intel. IA-64 application developerrs architecture guide [R]. Santa Clara, Calif: Intel Corp, 1999.
  • 3Beyls K, Hollander E D. Compile-time cache Hint generation for EPIC architectures[C] ,//Proceedings of the 2nd International Workshop on Explicitly Parallel Instruction Computing (EPIC) Architectures and Compiler Techniques. Los Alamitos: IEEE Computer Society, 2002: 19-29.
  • 4Yang H, Govindarajan R, Gao G R, et al. Compilerassisted cache replacement: problem formulation and performance evaluation[C] // Proceedings of the 16th International Workshop on Languages and Compilersfor Parallel Computing. College Sation.. Lecture Notes in Computer Science, 2003.
  • 5Pugh W. Counting solutions to presburger formulas: how and why[J]. ACM SIGPLAN Notices, 1994, 29(6) : 121-134.
  • 6Anup Gangwar. The trimaran compiler research infrastructure[EB/OL]. [2009-02-01]. http://www. trimaran, org.
  • 7Edler J, Hill M D. Dinero IV trace-driven uniprocessor cache simulator[EB/OL]. [2009-01-18]. http:// www. cs. wisc. edu/-markhill/DineroIV.
  • 8Dennis Gannon, William Jalby, Kyle Gallivan. Strat egies for Cache and local memory management by global programming transformation [J]. Journal of Parallel and Distributed Computing, 1988, 5 (5): 587-616.
  • 9Rivera G, Tseng C W. Eliminating conflict misses for high performance architectures [C]//ACM Internacional Conference on Supereomputing (1CS'98). Melbourne: ACM, 1998: 353-360.

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部