基于推测机制异构多核处理器容错方法与仿真

doi:10.16182/j.issn1004731x.joss.19-FZ0346

系统仿真学报 ›› 2019, Vol. 31 ›› Issue (12): 2685-2695.doi: 10.16182/j.issn1004731x.joss.19-FZ0346

基于推测机制异构多核处理器容错方法与仿真

余世干^1,2,3,4, 唐志敏^1,2,3, 叶笑春^1,2, 范东睿^1,2

1. 中国科学院计算技术研究所计算机体系结构国家重点实验室,北京 100190;
2. 中国科学院大学计算机控制与工程学院,北京 100049;
3. 先进微处理器技术国家工程实验室,四川成都 610218;
4. 阜阳师范大学信息工程学院,安徽阜阳 236041

收稿日期:2019-04-25 修回日期:2019-07-19 发布日期:2019-12-13
作者简介:余世干(1982-),男,安徽定远,硕士,副教授,研究方向为计算机体系结构与容错计算。
基金资助:
国家重点研发计划(2018YFB1003500), 国家自然科学基金(61732018, 61872335, 61802367), 安徽高校人才支持与科研项目(gxyq2019175, KJ2018A0669)

Fault-tolerant Method and Simulation of Heterogeneous Multi-core Processor Based on Speculative Mechanism

Yu Shigan^1,2,3,4, Tang Zhimin^1,2,3, Ye Xiaochun^1,2, Fan Dongrui^1,2

1. State Key Laboratory of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China;
2. School of Computer and Control Engineering, University of Chinese Academy of Sciences, Beijing 100049, China;
3.National Engineering Laboratory for Advanced Processor Technology, Chengdu 610218, China;
4.Information Engineering college, Fuyang Normal University, Fuyang 236041, China

Received:2019-04-25 Revised:2019-07-19 Published:2019-12-13

摘要/Abstract

摘要： 异构多核是处理器重要方向之一,却面临着瞬态故障频发问题,传统TMR(三模冗余)是主要解决办法,但有效率低,功耗高特点,提出基于推测机制高性能容错调度算法FTSAS。各异构核独立执行任务,记录最先完成核的状态值,采用前向推测法继续执行下一任务,采用多数一致原则,由落后的核完成结果比较,保障系统可靠性。仿真实验表明,FTSAS比当前容错方法平均性能提高了12.9%,注入200个错误时,具有相近的容错效果,但FTSAS平均执行性能提高了11.4%,平均功耗降低了15.8%。

关键词: 异构多核, 处理器, 推测机制, 容错, 调度

Abstract: Heterogeneous multicore is one of the important branches of processors,but they are still faced with frequent transient failures. TMR(Triple mode redundancy) is the main method to solve transient faults, which has the characteristics of low efficiency and high power consumption, a high-performance Fault-Tolerant Scheduling Algorithm with Speculative mechanism(FTSAS) is proposed. Each heterogeneous core can execute tasks independently, the state values of the first completed core are recorded, and the first completed core continues to perform the next task with forward speculative method. The results are compared by backward core, the majority consensus principle is adopted to ensure the reliability of the system. The simulation results show the average performance of the FTSAS is improved by 12.9% compared with state-of-the-art methods. When 200 errors are injected, the FTSAS has a similar fault-tolerant effect, however, FTSAS can achieve 11.4% average efficiency improvement and 15.8% average power consumption decrement.

Key words: Heterogeneous multicore, Processor, Speculative mechanism, fault tolerance, Scheduling

中图分类号:

TP391.7

余世干, 唐志敏, 叶笑春, 范东睿. 基于推测机制异构多核处理器容错方法与仿真[J]. 系统仿真学报, 2019, 31(12): 2685-2695.

Yu Shigan, Tang Zhimin, Ye Xiaochun, Fan Dongrui. Fault-tolerant Method and Simulation of Heterogeneous Multi-core Processor Based on Speculative Mechanism[J]. Journal of System Simulation, 2019, 31(12): 2685-2695.

参考文献

[1] McNairy C. Exascale fault tolerance challenge and approaches[C]. IEEE International Reliability Physics Symposium. Burlingame,USA: IEEE, 2018: 3C.4.1-10.
[2] Chen J, John L K.Efficient Program Scheduling for Heterogeneous Multicore Processors[C]. 46^th ACM/ IEEE Design Automation Conference. San Francisco, USA: IEEE, 2009: 927-930.
[3] Baital K, Chakrabarti A.Dynamic Scheduling of Real- Time Tasks in Heterogeneous Multicore Systems[J]. IEEE Embedded Systems Letters (S1943-0671), 2019, 11(1): 29-32.
[4] Greenhalgh P Big. LITTLE processing with ARM Cortex -A15&Cortex-A7. [EB/OL]. [2018-09]. http://www. arm.com.
[5] NVIDIA Corporation. Variable SMP-a multi-core CPU architecture for low power and high performance[EB/OL]. http://www.Nvidia.com. 2018.
[6] Chitlur N, Srinivasa G, Hahn S, et al.QuickIA: Exploring heterogeneous architectures on real prototypes[C]. IEEE International Symposium on High-Performance Computer Architecture (HPCA). New Orleans, USA: IEEE, 2012: 1-8.
[7] 海思麒麟980参数[EB/OL]. [2019-04-01].https: // baike. baidu.com.
Parameters of the kirin 980[EB/OL]. [2019-04-01].https: // baike.baidu. com.
[8] Naithani A, Eyerman S, Eeckhout L.Optimizing Soft Error Reliability Through Scheduling on Heterogeneous Multicore Processors[J]. IEEE Transactions on Computers (S1557-9956), 2018, 67(6): 830-846.
[9] Raparti V Y, Kapadia N, Pasricha S.CHARM: A check point-based resource manage-ment framework for reliable multicore computing in the dark silicon era[C]. IEEE International Conference on Computer Design (ICCD). Scottsdale, USA: IEEE, 2016: 201-208.
[10] Karlsson J, Liden P, Dahlgren P, et al.Using Heavy-ion Radiation to Validate Fault Handling Mechanisms[J]. IEEE Micro (S1937-4143), 1994, 14(1): 8-23.
[11] Lim H, Kim T, Lee D, et al.LARECD: Low area over- head and reliable error correction DMR architecture[C]. International SoC Design Conference. Seoul, South Korea: IEEE, 2018: 27-28.
[12] Zheng P, Zheng Q, Zeng Z.The Signal Design and Simulation of Triple Modular Redundant (TMR) Computer[C]. International Conference on Cybernetics and Intelligent Systems (CIS) and Conference on Robotics, Auto-mation and Mechatronics (RAM). Ningbo, China: IEEE, 2017: 758-762.
[13] 孙晓星. 面向方面的软件容错模型设计与分析技术 [D]. 上海: 华东理工大学, 2012.
Sun Xiaoxing.Aspect-oriented Modeling and Analysis Techniques for Software Fault Tolerance[D]. Shanghai: East China University of Science and Technology, 2012.
[14] Abate F, Sterpone L, Lisboa C A, et al.New Techniques for Improving the Performance of the Lockstep Architecture for SEEs Mitigation in FPGA Embedded Processors[J]. IEEE Transactions on Nuclear Science (S1558-1578), 2009, 56(4): 1992-2000.
[15] Roy A, Aydin H.Energy-efficient primary/backup scheduling techniques for heterogeneous multicore systems[C]. International Green and Sustainable Computing Conference. Orlando, USA: IEEE, 2017: 1-8.
[16] Ning L, Yao W, Ni J, et al.Fault Tolerance CMP Architecture based on SMT Technology[C]. International Multi-Symposiums on Computer and Computational Sciences. Iowa City, USA: IEEE, 2007: 425-429.
[17] 贾佳, 杨学军, 马亚青. 静态分析面向异构系统的应用级Checkpoint设置问题[J]. 软件学报, 2013, 24(6): 1361-1375.
Jia Jia, Yang Xuejun, Ma Yaqing.Static analysis for placement of application-level Checkpoints on heterogeneous system[J]. Chinese Journal of Software, 2013, 24(6): 1361-1375.
[18] 李祖松, 许先超, 胡伟武, 等. 龙芯2号处理器的同时多线程设计[J]. 计算机学报, 2009, 32(11): 2265-2273.
Li Zusong, Xu Xianchao, Hu Weiwu, et al.Design of the Simultaneous Multithreading Godson-2 Processor[J]. Chinese Journal of Computers, 2009, 32(11): 2265-2273.
[19] 贾佳, 杨学军, 李志凌. 一种基于冗余线程的GPU多副本容错技术[J]. 计算机研究与发展, 2013, 50(7): 1551-1562.
Jia Jia, Yang Xuejun, Li Zhiling.A Redundancy- Multithread-Based Multiple GPU Copies Fault- Tolerance Technique[J]. Chinese Journal of Computer Research and Development, 2013, 50(7): 1551-1562.
[20] 曹英. 大数据环境下Hadoop性能优化的研究[D]. 大连: 大连海事大学, 2013.
Cao Ying.Research of Performance Optimization of Hadoop in BigData[D]. Dalian: Dalian Maritime University, 2013.
[21] 陆志伟. 航空器无冲突4D航迹推测与生成[D]. 南京: 南京航空航天大学, 2012.
Lu Zhiwei.The Conflict-free 4D Trajectory Prediction and Generation for Aircrafts[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2012.
[22] 王丽华. 计算机容错系统的体系结构与安全性研究[D]. 成都: 西南交通大学, 2002.
Wang Lihua.The Research About the Architecture and Safety of the Computer Fault-tolerant System[D]. Chengdu: Southwest Jiaotong University, 2002.
[23] 谢国琪, 李仁发, 刘琳, 等. 异构分布式DAG可靠性模型与容错算法[J]. 计算机学报, 2013, 36(10): 2020-2033.
Xie Guoqi, Li Renfa, Liu Lin, et al.DAG Reliability Model and Fault-Tolerant Algorithm for Heterogeneous Distributed Systems[J]. Chinese Journal of computers, 2013, 36(10): 2020-2033.
[24] Benoit A, Hakem M, Robert Y.Fault tolerant scheduling of precedence task graphs on heterogeneous platforms[C]. IEEE International Symposium on Parallel and Distributed Processing. Miami, USA: IEEE, 2008: 1-8.
[25] Benoit A, Hakem M, Robert Y.Contention Awareness and Fault-tolerant Scheduling for Precedence Constrained Tasks in Heterogeneous Systems[J]. Parallel Computing (S0167-8191), 2009, 35(2): 83-108.
[26] 付剑. 星载计算机硬件容错设计与可靠性分析[D]. 长沙: 国防科技大学, 2009.
Fu Jian.The hardware Fault-tolerant Design and Reliability Analysis of On Board Computer[D]. Changsha: National University of Defense Technology, 2009.
[27] Xu J, Zhu Y, Ni J.A Simulator for Multicore Processor Micro-architecture Featuring Intercore Communication, Power and Thermal Behavior[C]. International Conference on Embedded Software and Systems Symposium. Sichuan China: IEEE, 2008: 237-242.
[28] Simplescalar LLC to serve and project[EB/OL]. [2018-09-20]. http:// www.simple scalar.com.
[29] 张福新, 章隆兵, 胡伟武. 基于Simplescalar的龙芯CPU模拟器Sim-Godson[J]. 计算机学报, 2007, 30(1): 68-73.
Zhang Fuxin, Zhang Longbin, Hu Weiwu.Sim-Godson: A Godson Processor Simulator Based on SimpleScalar[J]. Chinese Journal of Computers, 2007, 30(1): 68-73.
[30] 陈文智, 姜振宇, 吴帆. 基于MIPS体系的扩展指令融合技术[J]. 计算机学报, 2008, 31(11): 1888-1897.
Chen Wenzhi, Jiang Zhenyu, Wu Fan.Instruction Fusion Technology for the MIPS[J]. Chinese Journal of Computers, 2008, 31(11): 1888-1897.
[31] Luo Y, Packirisamy V, Hsu W.Energy efficient speculative threads: Dynamic thread allocation in same-ISA heterogeneous multicore systems[C]. Inter-national Conference on Parallel Architectures and Compilation Techniques. Vienna, Austria: IEEE, 2010: 453-464.
[32] 王健, 孙建伶, 王新宇, 等. 容错多处理机中一种高效的实时调度算法[J]. 软件学报, 2009, 20(10): 2628-2636.
Wang Jian, Sun Jianling, Wang Xinyu, et al.Efficient Scheduling Algorithm for Hard Real-Time Tasks in Primary-backup Based Multiprocessor Systems[J]. Chinese Journal of Software, 2009, 20(10): 2628-2636.
[33] Brooks D, Tiwari V, Martonosi M.Wattch: A framework for Architectural-Level Power Analysis and Optimi- zation[C]. Proceedings of the 27^th Annual International Symposium on Computer Architecture. Vancouver, BC, Canada: IEEE, 2000: 83-94.
[34] 倪俊杰. 基于SimpleScalar的性能和功耗分析多核模拟器[D]. 上海: 上海交通大学, 2008.
Ni Junjie.A Multiprocessor Simulator with Power Analysis Based on SimpleScalar[D]. Shanghai: Shanghai Jiao Tong University, 2008.
[35] 郑凯. 对数据在异构多核处理器模拟器中进行任务划分的研究[D]. 上海: 上海交通大学, 2008.
Zheng Kai.Evaluation of Partitioning Methods for Application on a Heterogeneous Multicore Processor Simulator[D]. Shanghai: Shanghai Jiao Tong University, 2008.

基于推测机制异构多核处理器容错方法与仿真

Fault-tolerant Method and Simulation of Heterogeneous Multi-core Processor Based on Speculative Mechanism

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	赵也践, 王艳红, 张俊, 于洪霞, 田中大. 改进Q学习算法在作业车间调度问题中的应用[J]. 系统仿真学报, 2022, 34(6): 1247-1258.
[2]	张其文, 张斌. 基于教学优化算法求解置换流水车间调度问题[J]. 系统仿真学报, 2022, 34(5): 1054-1063.
[3]	刘永奎, 曾鸣, 张霖, 郭金维, 原思阳, 平垚垚. 基于微服务架构的云制造调度仿真系统设计与开发[J]. 系统仿真学报, 2022, 34(4): 700-711.
[4]	薛乃阳, 丁丹, 王红敏, 樊怡乐, 刘仲谦. 引入微元法思想的混合测控资源联合调度方法[J]. 系统仿真学报, 2022, 34(4): 826-835.
[5]	叶飞, 李梓晴, 赖李媛君. 分布式车间生产维修联合调度仿真优化方法[J]. 系统仿真学报, 2022, 34(4): 688-699.
[6]	冯开团, 袁杰. 基于改进注水算法的离散车间任务分配问题研究[J]. 系统仿真学报, 2022, 34(4): 768-776.
[7]	陈魁, 毕利, 王文雅. 柔性作业车间AGV与机器双资源集成调度研究[J]. 系统仿真学报, 2022, 34(3): 461-469.
[8]	何欣航, 李而康, 张宏超. FlexRay总线静态段网络资源利用率优化研究[J]. 系统仿真学报, 2022, 34(3): 564-572.
[9]	吴定会, 张桐瑞, 张秀丽. 扰动累积下基于数字孪生的车间重调度[J]. 系统仿真学报, 2022, 34(3): 573-583.
[10]	赵梓安, 周泓, 雷颖健. 基于改进狼群算法与仿真的单元调度优化[J]. 系统仿真学报, 2022, 34(2): 201-211.
[11]	李启锐, 彭心怡. 基于深度强化学习的云作业调度及仿真研究[J]. 系统仿真学报, 2022, 34(2): 258-268.
[12]	赵静, 王鹏, 丁筱茜, 蒋国平, 徐丰羽, 孙雁飞. 基于观测器的四旋翼容错控制及仿真研究[J]. 系统仿真学报, 2022, 34(1): 1-10.
[13]	闫秀英, 党苗苗. 基于改进多目标粒子群算法的家庭用电时段优化[J]. 系统仿真学报, 2022, 34(1): 70-78.
[14]	李华珍, 柳有权, 朱家伟, 廖强. 基于改进粒子群算法的居民需求响应调度优化与仿真[J]. 系统仿真学报, 2021, 33(8): 1969-1979.
[15]	胡蓉, 董钰明, 钱斌. 基于探路者算法的绿色有限缓冲区流水线调度[J]. 系统仿真学报, 2021, 33(6): 1384-1396.