计算机科学 ›› 2020, Vol. 47 ›› Issue (10): 130-135.doi: 10.11896/jsjkx.190800115

• 数据库&大数据&数据科学 • 上一篇    下一篇

面向访问模式的混合内存缓存替换策略

刘伟1,2,3, 孙童心1, 杜薇1,2   

  1. 1 武汉理工大学计算机科学与技术学院 武汉430070
    2 交通物联网技术湖北省重点实验室(武汉理工大学) 武汉430070
    3 嵌入式系统与服务计算教育部重点实验室(同济大学) 上海201804
  • 收稿日期:2019-08-23 修回日期:2020-02-04 出版日期:2020-10-15 发布日期:2020-10-16
  • 通讯作者: 刘伟(wliu@whut.edu.cn)
  • 基金资助:
    国家自然科学基金面上项目(61672384);湖北省自然科学面上项目(2020CFB749);教育部人文社科基金项目(16YJCZH014);中央高校基本科研业务费(WUT:2016III028,2017III028-005);嵌入式系统与服务计算教育部重点实验室(同济大学)开放基金(ESSCKF2018-05)

Access Pattern-oriented Cache Replacement Strategy for Hybrid Memory Architecture

LIU Wei1,2,3, SUN Tong-xin1, DU Wei1,2   

  1. 1 Department of Computer Science and Technology,Wuhan University of Technology,Wuhan 430070,China
    2 Hubei Key Laboratory of Transportation Internet of Things,Wuhan University of Technology,Wuhan 430070,China
    3 Key Laboratory of Embedded Systems and Service Computing,Tongji University,Shanghai 201804,China
  • Received:2019-08-23 Revised:2020-02-04 Online:2020-10-15 Published:2020-10-16
  • About author:LIU Wei,born in 1978,Ph.D,associate professor,is a member of China Computer Federation.His main research interests include in-memory computing and edge computing.
  • Supported by:
    General Program of National Natural Science Foundation of China(61672384),Natural Science Foundation of Hubei Province(2020CFB749),Humanities and Social Sciences Program of Ministry of Education(16YJCZH014),Fundamental Research Funds for the Central Universities (WUT:2016III028,2017III028-005) and Open Fund of Key Laboratory of Embedded Systems and Service Computing of Ministry of Education,Tongji University(ESSCKF2018-05)

摘要: 大数据时代催生了很多以数据为中心的技术和应用,这对计算机主存的速度、容量、能耗提出了更高的要求。为了解决传统DRAM(Dynamic Random Access Memory)内存遇到的瓶颈,由DRAM和非易失性存储NVM(Non-Volatile Memory)组成的混合内存技术受到了广泛的关注。在混合内存环境下,缓存的性能至关重要。针对混合内存环境,已有的缓存替换策略研究都是对LRU2思想的改进,虽然考虑了DRAM数据和NVM数据缺失惩罚不对称的现象,但是在面对LRU(Least Recently Used)性能差的负载时也会存在缓存抖动和污染问题,仍然存在优化空间。文中针对不同类型的负载特点,考虑了不同访问模式下DRAM与NVM数据的竞争关系,提出了一种动态可调整的缓存替换策略DLRP(Dynamic Level Replacement Policy)。该策略在面对不同类型的负载时能动态地选择最优的替换策略,在保持整体命中率较好的同时降低了NVM的缺失和写回。实验结果表明,相比WBAR策略,DLRP不仅在IPC上有平均16.5%的提升,而且在能耗和写操作数量上分别降低了5.2%和5.1%。

关键词: NVM, 访问模式, 缓存替换, 混合内存

Abstract: With the increasing demand on memory capacity and energy consumption,current DRAM based memory systems face the scalability challenges in terms of storage density and power.Hybrid memory architecture,a promising approach to large-capacity and energy-efficient main memory composed of emerging Non-Volatile Memory (NVM) and DRAM has received extensive attention.Cache plays an important role and highly affects the number of write and read to NVM and DRAM blocks.However,existing cache policies based on LRU failed to fully address the significant asymmetry between NVM operations and DRAM ope-rations under different type of workloads.Cache trashing and scans problems can still seriously affect the performance of the system.By analyzing characteristics of different types of load and the competition between DRAM and NVM data under different access patterns,this paper proposes a dynamically adjusted level cache replacement strategy (DLRP).Experiment results show that proposed strategy improves the performance by 16.5% on average compared with a state-of-the-art cache policy (WBAR).DLRP also reduces energy consumption and NVM writes by 5.1%,5.2% against WBAR.

Key words: Aaccess pattern, Cache replacement, Hybrid memory, NVM

中图分类号: 

  • TP333
[1]MAO W,LIU J N,TONG W,et al.A Review of Storage Tech-nology Research Based on Phase Change Memory[J].Chinese Journal of Computers,2015,38(5):944-960.
[2]LEE S,B H,NOH S H.CLOCK-DWF:A Write-History-AwarePage Replacement Algorithm for Hybrid PCM and DRAM Memory Architectures[J].IEEE Transactions on Computers,2014,63(9):2187-2200.
[3]SALKHORDEH R,ASADI H.An operating system level datamigration scheme in hybrid DRAM-NVM memory architecture[C]// Proceedings of the 2016 Conference on Design.Automation & Test in Europe,2016:936-941.
[4]POURSHIRAZI B,ZHU Z.Refree:A Refresh-Free HybridDRAM/PCM Main Memory System[C]//2016 IEEE International Parallel and Distributed Processing Symposium (IPDPS).2016:566-575.
[5]CHEN J,LIU H K,WANG X Y,et al.Large pages Supported Hierarchical DRAM/NVM Hybrid Memory Systems[J].Journal of Computer Research and Development,2018,55(9):2050-2065.
[6]LIU C M,YANG X,JIA G Y,et al.Hybrid Memory Page Management Strategy of Avoiding Page Migrations[J].Journal of Chinese Computer Systems,2019,40(6):1318-1323.
[7]LI Q,ZHONG J,LI X,et al.Memory Management Mechanism for Hybrid Memory Architecture Based on New Non-volatile Memory[J].ACTA ELECTRONICA SINICA,2019,47(3):664-670.
[8]CHEN D,JIN H,LIAO X,et al.MALRU:Miss-penalty aware LRU-based cache replacement for hybrid memory systems[C]//Proceedings of the Conference on Design.Automation & Test in Europe,2017:1086-1091.
[9]ZHANG D,JU L,ZHAO M,et al.Write-back aware sharedlast-level cache management for hybrid main memory[C]//Proceedings of the 53rd Annual Design Automation Conference.2016:1-6.
[10]HUANG Z B,ZHOU F,MA H D.A Cache Replacement Policy Adapting to the Request Access Pattern[J].Journal of Beijing University of Posts and Telecommunications,2016,39(3):44-48,53.
[11]QURESHI M K,JALEEL A,PATT Y N,et al.Adaptive insertion policies for high performance caching[J].ACM SIGARCH Computer Architecture News,2007,35(2):381-391.
[12]WEI W,JIANG D,XIONG J,et al.HAP:Hybrid-memory-aware partition in shared last-level cache[J].ACM Transactions on Architecture and Code Optimization,2017,14(3):24.
[13]JALEEL A,THEOBALD K B,STEELY J S C,et al.High performance cache replacement using re-reference interval prediction (RRIP)[C]//ACM SIGARCH Computer Architecture News.2010:60-71.
[14]BINKERT N,BECKMANN B,BLACK G,et al.The gem5 simulator[J].ACM Sigarch Computer Architecture News,2011,39(2):1-7.
[15]POREMBA M,XIE Y.Nvmain:An architectural-level mainmemory simulator for emerging non-volatile memories[C]//2012 IEEE Computer Society Annual Symposium on VLSI.2012:392-397.
[1] 刘高聪, 罗永平, 金培权.
基于热点数据的持久性内存索引查询加速
Accelerating Persistent Memory-based Indices Based on Hotspot Data
计算机科学, 2022, 49(8): 26-32. https://doi.org/10.11896/jsjkx.210700176
[2] 郦睿翔, 毛莺池, 郝帅.
基于近似匹配的移动边缘计算缓存管理方法
Cache Management Method in Mobile Edge Computing Based on Approximate Matching
计算机科学, 2021, 48(1): 96-102. https://doi.org/10.11896/jsjkx.200800215
[3] 杨冬菊,冯凯.
基于缓存的分布式统一身份认证优化机制研究
Distributed and Unified Authentication Optimization Mechanism Based on Cache
计算机科学, 2018, 45(3): 300-304. https://doi.org/10.11896/j.issn.1002-137X.2018.03.049
[4] 吴炀,付印金,陈卫卫,倪桂强.
一种高效的混合内存布局机制与编码技术
Efficient Mechanism of Hybrid Memory Placement and Erasure Code
计算机科学, 2017, 44(6): 57-62. https://doi.org/10.11896/j.issn.1002-137X.2017.06.009
[5] 田铭,邬江兴,兰巨龙.
信息中心网络中基于局部内容活跃度的自适应缓存算法
Adaptive Cache Algorithm Based on Local Content Activity in Information-centric Networks
计算机科学, 2016, 43(11): 164-171. https://doi.org/10.11896/j.issn.1002-137X.2016.11.032
[6] 曹旻,刘文中.
混合架构下多请求模式的缓存替换模型研究
Research on Cache Replacement Model Based on Multi-request Mode under Hybrid Architecture Model
计算机科学, 2015, 42(6): 175-180. https://doi.org/10.11896/j.issn.1002-137X.2015.06.038
[7] 雷娟,朱珠,符云清,石亮.
一种高效的闪存存储系统偏好访问模式识别技术研究
Research of Efficient Recognition Technique for Preferred Access Pattern of Flash Storage System
计算机科学, 2015, 42(2): 86-89. https://doi.org/10.11896/j.issn.1002-137X.2015.02.018
[8] 叶剑虹,叶 双.
基于混合模式的流媒体缓存调度算法
Dynamic Scheduling Algorithms for Streaming Media Based on Hybrid Content Delivery Network
计算机科学, 2013, 40(2): 61-64.
[9] 林文龙 刘业政 姜元春.
Web浏览预测的Markov模型综述

计算机科学, 2008, 35(1): 9-14.
[10] 安静斌 贾焰 王志英 韩伟红.
持久对象框架中基于多级访问模式的对象预取技术研究与实现

计算机科学, 2007, 34(6): 139-141.
[11] 黄敏 蔡志刚.
缓存替换算法研究综述

计算机科学, 2006, 33(B12): 191-193.
[12] 李文中 顾铁成 周俊 陆桑璐 陈道蓄.
适应性Web缓存的研究

计算机科学, 2005, 32(4): 11-15.
[13] 侯志强 刘东华 何戈 徐志伟.
Internet 3A访问模式及相关技术的研究

计算机科学, 2005, 32(2): 36-39.
[14] 郭岩 白硕 于满泉.
Web使用信息挖掘综述

计算机科学, 2005, 32(1): 1-7.
[15] 马力 焦李成 刘国营.
一种基于路径聚类的Web用户访问模式发现算法

计算机科学, 2004, 31(8): 140-141.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!