期刊文献+

可配置及历史信息感知的多级缓存策略 被引量:1

Configurable and History Information-Aware Replacement Policy for a Multilevel Cache
下载PDF
导出
摘要 传统磁盘存储设备因其固有的机械特性,已不能满足当前的数据密集型应用程序的需求。基于闪存的固态存储设备(solid state drive,SSD)的出现改善了这种情况,并被广泛用作缓存以降低内存与磁盘之间的性能差距。针对由DRAM和SSD构成的多级缓存,提出了一种可配置的历史信息感知的多级缓存替换策略Charm.Charm允许用户配置应用的访问模式、读写模式等多项内容,并且还可以根据应用对文件的历史访问信息来判断访问模式,从而能够适应访问模式的变化.此外,Charm过滤掉那些只访问一次的数据,将多次访问的热数据缓存至SSD,减少对SSD的写入次数,提升SSD寿命.使用MCsim对Charm与现有的多级缓存替换算法进行了对比测试,在实际的工作负载下,Charm优于其它多级缓存算法. The traditional disk storage devices have been unable to meet the needs of current data intensive applications due to their inherent mechanical properties.The appearance of the solid state drive(SSD),which is based on flash memory,improves this situation,and it is widely used as a cache to reduce the performance gap between memory and disk.This paper proposed a configurable and history information aware replacement policy called Charm for a multi-level cache which consists of DRAM and SSD.Charm allows users to configure multiple content such as access patterns,read-write mode and so on for applications,and can also determine access pattern by watching historical access information.It is possible to adapt to changes in access patterns.In addition,Charm filters out the data which is only visited once,and caches the hot data which is visited once or more into the SSD.Thereby it reduces the number of write to the SSD,and extends the life time of SSD.The experiments with the real life workload on MCsim show that Charm is better than other multi-level cache replacement algorithms.
出处 《计算机研究与发展》 EI CSCD 北大核心 2015年第S2期163-170,共8页 Journal of Computer Research and Development
基金 国家"八六三"高技术研究发展计划基金项目(2013AA013203) 华中科技大学自主创新研究基金项目(HUST:2014QN010)
关键词 多级缓存 动态随机存取存储器 固态盘 缓存替换算法 历史信息感知 multi-level cache dynamic random access memory(DRAM) solid state drive(SSD) cache replacement algorithms history information-aware
  • 相关文献

参考文献9

  • 1Pei Cao,Edward W. Felten,Anna R. Karlin,Kai Li.??Implementation and performance of integrated application-controlled file caching, prefetching, and disk scheduling(J)ACM Transactions on Computer Systems (TOCS) . 1996 (4)
  • 2Huang Sai,Wei Qingsong,Chen Jianxi,et al.Improving flash-based disk cache with lazy adaptive replacement. Proc of the 29th IEEE Symp on Mass Storage Systems and Technologies (MSST) . 2013
  • 3Jiang Dejun,Che Yukun,Xiong Jin,et al.uCache:A utility-aware multilevel SSD cache management policy. Proc of the 10th IEEE Int Conf on High Performance Computing and Communications&20131EEE Int Conf on Embedded and Ubiquitous Computing (HPCC_EUC) . 2013
  • 4Zhang Y,Rajimwale A,Arpaci-Dusseau A. C,Arpaci-Dusseau R. H.End-to-end Data Integrity for File Systems:A ZFS Case Study. Proceedings of the 8th USENIX Conference on File and Storage Technologies (FAST‘10) . 2010
  • 5Theodore Johnson,Dennis Shasha.2Q: A Low Overhead High PerformanceBuffer Management Replacement Algorithm. Proceedings of the20thInternational Conference on Very Large Data Bases (VLDB ’’94) . 1994
  • 6O’Neil E,O’Neil P,Weikum G.The LRU-K Page Replacement Algorithm For Database Disk Buffering. Proceedings of the 1993 ACM SIGMOD International Conference on Management of Data . 1993
  • 7M. Factor,,A. Schuster.Know-it-all Replacement for a Multilevel Cache. 5th USENIX Conference on File and Storage Technologies . 2007
  • 8Song Jiang,Xiaodong Zhang.??LIRS(J)ACM SIGMETRICS Performance Evaluation Review . 2002 (1)
  • 9沈志荣,薛巍,舒继武.新型非易失存储研究[J].计算机研究与发展,2014,51(2):445-453. 被引量:14

二级参考文献36

  • 1Roberts I), Kgil T, Mudge T. Using non-volatile memory to save energy in servers[C] //Proc of the Conf on Design, Automation and Test in Europe. Belgium: European Design and Automation Association, 2009:743-748.
  • 2Qureshi M, Franceschini M, Lastras Montaflo L. Improving read performance of phase change memories via write cancellation and write pausing[C] //Proc of the 16th IEEE lnt Syrup on High Performance Computer Architecture. Piscataway, NJ: 1EEE, 2010:1 11.
  • 3Mishra A, Dong X, Sun G, et al. Architecting on chip interconnects for stacked 3D STT-RAM caches in CMPs [C] //Proc of the 38th lnt Syrup on Computer Architecture. New York: ACM, 2011:69-80.
  • 4Gun X, lpek E, Soyata T. Resistive computation: Avoiding the power wall with low-leakage, stt mram based computing [C] //Proc of the 37th Int Symp on Computer Architecture. New York: ACM, 2010:371 382.
  • 5Nigam A, Munira K, Ghosh A, et al. Model based study on energy and performance optimization for STT RAM [C/OL] //2011 Non-Volatile Memories Workshop. 2011. [2013 04 01]. http://nvmw, ucsd. edu/2011/.
  • 6Kong J, Zhou H. Improving privacy and lifetime of PCM based main memory [C] //Proc of the 40th Annual 1EEE/ IFIP Int Conf on Dependable Systems and Networks. Piscataway, NJ: IEEE, 2010:333-342.
  • 7Chhabra S, Solihin Y. i NVMM: A secure non volatile main memory system with incremental encryption [C] //Proc of the 38thlntSymponComputer Architecture. New York: ACM, 2011:177 188.
  • 8Sun (3, DongX, Xie Y, eta[. A novel architecture of the 3[) stacked MRAM L2 cache for CMPs [C] //Proc of the 15th IEEE Int Symp on High Performance Computer Architecture. Piscataway, N J: IEEE, 2009:239-249.
  • 9Kang H, Ryu K, Lee D, et al. Process variation tolerant all digital multiphase DLL for DDR3 interface [C] //Proc of IEEE Custom Integrated Circuits Conference. Piscataway, NJ: IEEE, 2010:1 4.
  • 10Zhou P, Zhao B, Yang J, et al. Energy reduction for STTRAM using early write termination [C] //Proe of IEEE/ ACM 2009 Int Conf on Computer Aided Design. Piseataway, NJ: IEEE, 2009:264-268.

共引文献13

同被引文献7

引证文献1

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