基于多级磁自旋存储器的Cache调度策略的设计

摘要/Abstract

摘要： 多级磁自旋存储器(Multi-Level Cell Spin-Transfer Torque RAM,MLC STT-RAM)可在一个存储单元中存储多个比特位,有望取代SRAM用于构建大容量低功耗的最后一级Cache(Last Level Cache,LLC)。MLC STT-RAM的静态功耗在理论上为0,且拥有高密度和优秀的读操作特性,但它的缺陷在于低效的写操作。针对这一问题,在MLC STT-RAM Cache hard/soft逻辑分区结构的基础上,实现了MLC STT-RAM LLC写操作密集度预测技术以及相应Cache结构的设计。通过动态预测写操作密集度较高的Cache块,帮助MLC STT-RAM LLC减少执行写操作的代价。预测的基本思想是利用访存指令地址与相应Cache块行为特征的联系,根据预测结果决定数据在LLC中的放置位置。实验结果显示,在MLC STT-RAM LLC中应用写操作密集度预测技术,使得写操作动态功耗降低6.3%的同时,系统性能有所提升。

关键词: 低功耗, 多级磁自旋存储器, 预测机制, 最后一级高速缓存

Abstract: Multi-level cell (MLC) STT-RAM which can store multiple bits per cell,has been considered as a promising alternative to SRAM for the last-level Cache.MLC STT-RAM can reduce static power consumption significantly and has smaller cell size facilitates and better read performance.However,a major shortcoming of MLC STT-RAM Cache is its inefficient write operations.Based on hard/soft partition structure,this paper implemented write intensity prediction for energy-efficient MLC STT-RAM LLC.The objective of this architecture is to dynamically predict whether blocks will be written more than certain times thereby helping to reduce write latency and energy of MLC STT-RAM Cache.The key idea to solve this problem is to correlate write intensity with memory access instruction addresses.On top of that,this paper designed MLC STT-RAM LLC based on this predictor,in which prediction results are used to determine Cache line placement.Experimental results showed that this architecture reduces 6.3% of write energy consumption and improves system performance by 1.9% on average compared to the previous approach.

Key words: Energy-efficient, Last level high-speed cache, Multi-Level Cell Spin-Transfer Torque-RAM, Prediction mechanism

中图分类号:

TP332.3

朱艳娜,王党辉. 基于多级磁自旋存储器的Cache调度策略的设计[J]. 计算机科学, 2018, 45(6A): 513-517. https://doi.org/

ZHU Yan-na,WANG Dang-hui. Design of Cache Scheduling Policies Based on MLC STT-RAM[J]. Computer Science, 2018, 45(6A): 513-517. https://doi.org/

参考文献

[1]CHANDRAKASAN A P,SHENG S,et al.Low-Power CMOS Digital Design [J].IEEE Journal of Solid-State Circuits,1992,27(4):473-484.
[2]JOSHI M,ZHANG W Y,LI T.Mercury:A Fast and Energy-Efficient Multi-level Cell based Phase Change Memory System [C]∥High Performance Computer Architecture (HPCA).San Antonio:IEEE,2011:345-356.
[3]PARK S K,MAENG M K,PARK K W,et al.Adaptive Wear-Leveling Algorithm for PRAM Main Memory with a DRAM Buffer [C]∥ACM Transactions on Embedded Computing Systems.New York:ACM,2014.
[4]EKEN E,ZHANG Y,YAN B.Spin-hall assisted STT-RAM design and discussion [C]∥IEEE Magnatics Conference (INTERMAG).Austin:IEEE,2015:1.
[5]ZHANG Y J,LI Y,SUN Z Y.Read Performance:The Newest Barrier in Scaled STT-RAM [C]∥IEEE Transactions on Very Large Scale Integration (VLSI) Systems.IEEE,2015:1170-1174.
[6]CHEN Y R,WONG W F,LI H,et al.Processor Caches with Multi-level Spin-transfer Torque RAM Cells [C]∥IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED).Piscataway:IEEE,2011:73-78.
[7]BI X Y,MAO M J,WANG D H,et al.Unleashing the potential of MLC STT-RAM Caches [C]∥IEEE/ACM International Conference on Computer-Aided Design (ICCAD).San Jose:IEEE,2013:429-436.
[8]MURALIMANOLAR N,BALASUBRAMONIAN B,JOUPPI N P,et al.CACTI 6.0:A tool to model large Caches[R].HP Laboratories,Technical Report,2009.
[9]CHEN Y R,WANG X B,ZHU W Z.Access Scheme of Multi-Level Cell Spin-Transfer Torque Random Access Memory and its Optimization [C]∥IEEE International Midwest Symposium on Circuits and Systems (MWSCAS).Seattle:IEEE,2010:1109-1112.
[10]ZHOU P,ZHAO B,YANG J,et al.Energy Reduction for STT-RAM using Early Write Termination [C]∥IEEE/ACM International Conference on Computer-Aided Design of Technical Papers.San Jose:IEEE,2009:264-268.
[11]JIANG L,ZHAO B,ZHANG Y T,et al.Construction Large and Fast Multi-Level Cell STT-MRAM based Cache for Embedded Processors [C]∥Design Automation Conference.New York:ACM,2012:907-912.
[12]AHN J W,YOO S J,CHOI K.Write intensity prediction for ener- gy-efficient non-volatile Caches [C]∥International Sympo-sium on Low Power Electronics and Design (ISLPED).Beijing:IEEE,2013:223-228.
[13]HENNING J L.SPEC CPU2006 benchmark descriptions[J]. ACM SIGARCH Computer Architecture,2006,34(4):1-17.

相关文章 15

[1]	何权奇, 余飞鸿. 面向无线网络相机的低功耗架构研究综述 Review of Low Power Architecture for Wireless Network Cameras 计算机科学, 2021, 48(6A): 369-373. https://doi.org/10.11896/jsjkx.201100099
[2]	张英, 陶磊岩, 曹健, 王世会, 赵茜, 张兴. 实时低功耗飞行器神经网络 Real-time Low Power Consumption Aircraft Neural Network 计算机科学, 2021, 48(3): 196-200. https://doi.org/10.11896/jsjkx.191200142
[3]	范星冉, 宋国治, 李加正. 基于混合混沌大爆炸算法的三维片上网络低功耗映射 Low-power Mapping Method for Three-dimensional Network on Chip Based on Hybrid Chaotic Big Bang-big Crunch 计算机科学, 2019, 46(8): 100-105. https://doi.org/10.11896/j.issn.1002-137X.2019.08.016
[4]	杨瑞君, 祝可, 程燕. 基于SVM访问预测机制的Web缓存数据库级替换策略 Database-level Web Cache Replacement Strategy Based on SVM Access Prediction Mechanism 计算机科学, 2019, 46(6): 201-205. https://doi.org/10.11896/j.issn.1002-137X.2019.06.030
[5]	王华华, 周远文, 刘江兵. LLN中基于混合式的网络拥塞控制路由算法 Hybrid-based Network Congestion Control Routing Algorithm for LLN 计算机科学, 2019, 46(6): 107-111. https://doi.org/10.11896/j.issn.1002-137X.2019.06.015
[6]	吴伟男, 刘建明. 面向低功耗无线传感器网络的动态重传算法 Dynamic Retransmission Algorithm inLow-power Wireless Sensor Networks 计算机科学, 2018, 45(6): 96-99. https://doi.org/10.11896/j.issn.1002-137X.2018.06.016
[7]	谈恩民,范玉祥. 一种基于海明排序进行无关位填充的低功耗测试向量优化方法 Optimization Method of Low Power Test Vectors Based on Hamming Sorting for X Bits Padding 计算机科学, 2018, 45(2): 249-253. https://doi.org/10.11896/j.issn.1002-137X.2018.02.043
[8]	王耀兴,刘建军. 低功耗WSN节点及其接口协议的设计 Design of Low-power Node and Interface Protocol in WSN 计算机科学, 2014, 41(Z11): 228-231.
[9]	方娟,王帅,于璐. 面向低功耗共享Cache路适应划分算法研究 Research of Lower Power Oriented Way-adaptive Partition Algorithm in Shared Cache of CMP 计算机科学, 2014, 41(7): 36-39. https://doi.org/10.11896/j.issn.1002-137X.2014.07.006
[10]	冯永,姚龙海,张亮. 基于M/M/1/K排队模型的低功耗无线通信网络TDMA协议延迟评估及仿真 Delay Evaluation and Simulation of TDMA Through M/M/1/K Queue Model in Low-energy Wireless Communication Environment 计算机科学, 2013, 40(Z6): 262-264.
[11]	田泽,张骏,许宏杰,郭亮,黎小玉. 图形处理器低功耗设计技术研究 Low-power Design Techniques for GPU 计算机科学, 2013, 40(Z6): 210-216.
[12]	方娟,郭媚,杜文娟. WPP-L2:多核处理器中共享Cache低功耗路预测算法 WPP-L2:A Way-predicting Algorithm for Shared L2Cache of Multi-core Processors 计算机科学, 2013, 40(8): 34-37.
[13]	王剑锋,陈灿峰,刘嘉,郗闽军. 一种基于IPv6和低功耗蓝牙的物联网体系结构 Internet of Things Architecture Based on IPv6and Bluetooth Low Energy 计算机科学, 2013, 40(5): 97-102.
[14]	顾云丽,徐昕,侯荣涛,杜杰,钱焕延,梅园. 基于移动Sink的无线传感器网络任播路由协议 Anycast Routing Protocol for Mobile Sinks in Wireless Sensor Networks 计算机科学, 2013, 40(12): 166-168.
[15]	葛红美,徐超,陈念,廖希密. 一种面向嵌入式系统总线的低功耗优化方法 Low Power Optimization Method Oriented to Embedded System’s Bus 计算机科学, 2013, 40(12): 31-36.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed