多核处理器共享Cache的划分算法

doi:10.16180/j.cnki.issn1007-7820.2024.09.005

电子科技 ›› 2024, Vol. 37 ›› Issue (9): 27-33.doi: 10.16180/j.cnki.issn1007-7820.2024.09.005

多核处理器共享Cache的划分算法

吕海玉, 罗广, 朱嘉炜, 张凤登

上海理工大学光电信息与计算机工程学院,上海 200093

收稿日期:2023-03-08 出版日期:2024-09-15 发布日期:2024-09-20
作者简介:吕海玉(1999-),男,硕士研究生。研究方向:分布式实时系统、多核处理器。
罗广(1999-),男,硕士研究生。研究方向:分布式实时系统。
张凤登(1963-),男,博士,教授。研究方向:汽车电子与现场总线、分布式实时系统。
基金资助:
国家自然科学基金(71840003);上海市自然科学基金(15ZR1429300)

Research on Shared Cache Partition Algorithm Based on Multi-Core Processors

LÜ Haiyu, LUO Guang, ZHU Jiawei, ZHANG Fengdeng

School of Optical-Electrical and Computer Engineering,University of Shanghai for Science & Technology,Shanghai 200093,China

Received:2023-03-08 Online:2024-09-15 Published:2024-09-20
Supported by:
National Natural Science Foundation of China(71840003);Shanghai Municipal Natural Science Fund Project(15ZR1429300)

摘要/Abstract

摘要：

针对多核处理器性能优化问题,文中深入研究多核处理器上共享Cache的管理策略,提出了基于缓存时间公平性与吞吐率的共享Cache划分算法MT-FTP(Memory Time based Fair and Throughput Partitioning)。以公平性和吞吐率两个评价性指标建立数学模型,并分析了算法的划分流程。仿真实验结果表明,MT-FTP算法在系统吞吐率方面表现较好,其平均IPC(Instructions Per Cycles)值比UCP (Use Case Point)算法高1.3%,比LRU(Least Recently Used)算法高11.6%。MT-FTP算法对应的系统平均公平性比LRU算法的系统平均公平性高17%,比UCP算法的平均公平性高16.5%。该算法实现了共享Cache划分公平性并兼顾了系统的吞吐率。

关键词: 片上多核处理器, 内存墙, 划分, 公平性, 吞吐率, 共享Cache, 缓存时间, 集成计算机

Abstract:

In order to optimize the performance of multi-core processors, this study deeply investigates the management strategy of shared Cache on multi-core processors, and proposes a shared Cache partitioning algorithm MT-FTP(Memory Time based Fair and Throughput Partitioning) based on the fairness of cache time and throughput rate. A mathematical model based on the fairness and throughput index is established, and the partitioning flow of the algorithm is analyzed in the proposed study. The simulation results show that the MT-FTP algorithm has excellent performance in system throughput, and its average IPC (Instructions Per Cycles) value is 1.3% higher than that of UCP(Use Case Point) algorithm and 11.6% higher than that of LRU (Least Recently Used) algorithm. The average fairness of MT-FTP algorithm is 17% higher than that of LRU algorithm, and 16.5% higher than that of UCP algorithm. This algorithm realizes the fairness of shared Cache partition and takes into account the throughput of the system.

Key words: chip multi-processor, memory wall, partition, fairness, throughput rate, shared Cache, cache time, integrated computer

中图分类号:

TP316.2

吕海玉, 罗广, 朱嘉炜, 张凤登. 多核处理器共享Cache的划分算法[J]. 电子科技, 2024, 37(9): 27-33.

LÜ Haiyu, LUO Guang, ZHU Jiawei, ZHANG Fengdeng. Research on Shared Cache Partition Algorithm Based on Multi-Core Processors[J]. Electronic Science and Technology, 2024, 37(9): 27-33.

图/表 7

图1

图2

图3

图4

图5

表1

图6

参考文献 18

[1]	Suo G, Yang X, Liu G, et al. IPC-based cache partitioning:An IPC-oriented dynamic shared cache partitioning mechanism[C]. Daejeon: International Conference on Convergence and Hybrid Information Technology, 2018:399-406.
[2]	Kim S, Chandra D, Solihin D. Fair cache sharing and partitioning in a chip multiprocessor architecture[C]. Antibes: Proceedings of the Thirteenth International Conference on Parallel Architectures and Compilation Techniques, 2014:111-122.
[3]	Chen X, Xiao H, Wardi Y, et al. Throughput regulation in shared memory multicore processors[C]. Bengaluru: IEEE the Twenty-second International Conference on High Performance Computing, 2015:12-20.
[4]	Qiu T, Liu X, Han M, et al. SRTS:A self-recoverable ti-me synchronization for sensor networks of healthcare IoT[J]. Computer Networks, 2017, 12(9):481-492.
[5]	王天一, 高博. 分布式实时系统数据分发服务DDS技术研究[J]. 电子科技, 2020, 33(8):40-45.
	Wang Tianyi, Gao Bo. Research on data distribution service of distributed real-time system[J]. Electronic Science and Technology, 2020, 33(8):40-45.
[6]	Gao C, Zhao G R, Lu J H, et al. Decentralised moving-horizon state estimation for a class of networked spatial-navigation systems with random parametric uncer-tainties and communication link failures[J]. IET Control Theory and Applications, 2015, 9(18):2666-2677.
[7]	赵龙飞, 李晓婷, 丁振兰, 等. 基于SPEC CPU2017的CPU性能对比分析[J]. 电子产品可靠性与环境试验, 2021, 39(S1):55-59.
	Zhao Longfei, Li Xiaoting, Ding Zhenlan, et al. A comparative analysis of CPU performance based on SPECCPU2017[J]. Electronic Product Reliability and Environmental Testing, 2021, 39(S1):55-59.
[8]	Liu C, Sivasubramaniam A, Kandemir M. Organizing thelast line of defense before hitting the memory wall for cmps[C]. Madrid: IEEE the Tenth International Symposium on High Performance Computer Architecture, 2014:175-179.
[9]	Qureshi M K, Patt Y N. Utility-based cache partitioning:A low-overhead,high-performance,runtime mechanismto partition shared caches[C]. Orlando: Proceedings of the Thirty-ninth Annual IEEE/ACM International Symposium on Microarchitecture, 2016:423-432.
[10]	刘让, 张凤登. FlexRay时钟同步拜占庭故障容错算法研究[J]. 软件导刊, 2020, 19(1):68-74.
	Liu Rang, Zhang Fengdeng. Research on fault tolerant algorithm of Byzantine fault in FlexRay clock synchronization[J]. Software Guide, 2020, 19(1):68-74.
[11]	康立毅, 陈晓, 王劲林. 择优同步分布式时钟同步系统设计[J]. 网络新媒体技术, 2019, 8(1):53-57.
	Kang Liyi, Chen Xiao, Wang Jinlin. Design of distributed clock synchronization system by selecting execllent network nodes to synchronize with[J]. Network New Media Technology, 2019, 8(1):53-57.
[12]	Xie Y, Loh G H. Pipp:Promotion/insertion pseudo-partit-ioning of multicore shared caches[J]. ACM SIGARCH Computer Architecture News, 2019, 37(3):174-183.
[13]	Liu L, Qi D. An independent task scheduling algorithmin heterogeneous multicore processor environment[C]. Chongqing: IEEE the Third Advanced Information Technology,Electronic and Automation Control Conference, 2018:142-146.
[14]	方娟, 李成艳, 王帅, 等. 一种基于频率的多核共享Cache替换算法[J]. 电子与信息学报, 2014, 36(5):1229-1234.
	Fang Juan, Li Chengyan, Wang Shuai, et al. A frequency based Cache replacement algorithm with partition of C-MPs[J]. Journal of Electronics and Information Technology, 2014, 36(5):1229-1234.
[15]	Xu W, Sun H, Wang X, et al. Design of last-level on-chip Cache using spin-torque transfer RAM[J]. IEEE Transactions on Very Large-Scale Integration Systems, 2011, 19(3):483-493.
[16]	Gao C, Zhao G R, Lu J H, et al. Decentralised moving-horizon state estimation for a class of networked spatial-navigation systems with random parametric uncertainties and communication link failures[J]. IET Control Theory and Applications, 2015, 9(18):2666-2677.
[17]	Austin T, Larson E, Dan E. Simple scalar:An infrastructure for computer system modeling[J]. Computer, 2012, 35(2):59-67.
[18]	Wulf W A, Mckee S A. Hitting the memory wall:Implications of the obvious[J]. ACM Sigarch Computer Architecture News, 1995, 23(1):20-24.

编号	测试用例
MIX₁	applu, mcf, gap, wvpwise
MIX₂	mcf, gap, gzip, perlbmk
MIX₃	ammp, twlf, apsi, parser
MIX₄	ammp, twlf, applu, mcf
MIX₅	ammp, parser, gzip, perlbmk
MIX₆	gzip, perlbmk, apsi, parser
MIX₇	ammp, twolf, crafty, mesa
MIX₈	gzip, perrbmk, crafty, mesa
MIX₉	gzip, ammp, applu, perlbmk
MIX₁₀	mesa, crafty, perbmk, applu

多核处理器共享Cache的划分算法

Research on Shared Cache Partition Algorithm Based on Multi-Core Processors

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 7

参考文献 18

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	罗广, 冒航, 朱扬烁, 张凤登. 嵌入式多核系统中的实时混合任务调度算法[J]. 电子科技, 2024, 37(8): 84-91.
[2]	朱嘉炜, 冒航, 张凤登. 多核处理器中混合关键级任务可调度及半分区划分算法[J]. 电子科技, 2024, 37(3): 1-9.
[3]	张铭坤,蔡文郁,张帅. 基于LiDAR扫描角度修正的障碍目标定位方法[J]. 电子科技, 2024, 37(1): 41-47.
[4]	刘伟,曾祥义,肖昊. 基于轻量级分组密码算法的SoC安全存储器设计[J]. 电子科技, 2023, 36(9): 15-20.
[5]	程甜甜,宋宇鲲. 基于FPGA的2底指数函数算法优化与实现[J]. 电子科技, 2023, 36(9): 66-72.
[6]	张海涛,张通,张宇辉,管银凤,张凤登. 基于MrsP协议的任务划分优化算法[J]. 电子科技, 2023, 36(3): 36-41.
[7]	林宇红,肖昊. 基于AES算法的eFlash控制器安全设计[J]. 电子科技, 2023, 36(10): 62-67.
[8]	王晓蕾,戴吴骏,杜高明,李桢旻,张多利. 一种高性能极化码SC译码器设计[J]. 电子科技, 2022, 35(8): 14-20.
[9]	薛亮,王然. 一种移动传感网络中K-栅栏寿命的优化方法[J]. 电子科技, 2022, 35(8): 34-40.
[10]	王春华,李斌,杜高明,李桢旻. 一种高性能RLWE加密处理器的设计与实现[J]. 电子科技, 2022, 35(11): 13-20.
[11]	罗小龙,刘子龙. 基于快速编码单元划分和自适应模式决策的HEVC效率改进算法[J]. 电子科技, 2020, 33(10): 21-25.
[12]	潘迪. 基于DBSCAN与梯度划分的Kinect障碍物轮廓检测算法[J]. 电子科技, 2019, 32(1): 86-90.
[13]	史雨雨，郁进明，李雪. 基于能量和密度的WSNs动态分区成簇路由算法[J]. , 2016, 29(6): 47-.
[14]	周菡, 于文斌. 视频码流的数值比特位划分[J]. , 2011, 24(12): 22-.
[15]	周菡, 于文斌. 视频码流的层次划分研究[J]. , 12, 25(9): 154-.