一种面向DDR3接口的新型Chipkill编码

J4 ›› 2014, Vol. 36 ›› Issue (12): 2386-2393.

一种面向DDR3接口的新型Chipkill编码

王谛，许勇

（江南计算技术研究所，江苏无锡 214083）

收稿日期:2014-07-03 修回日期:2014-09-21 出版日期:2014-12-25 发布日期:2014-12-25
基金资助:
上海教委科研创新重点基础研究资助项目（12ZZ182）

A novel Chipkill code for DDR3 interface

WANG Di，XU Yong

（Jiangnan Institute of Computer Technology,Wuxi 214083,China）

Received:2014-07-03 Revised:2014-09-21 Online:2014-12-25 Published:2014-12-25

摘要/Abstract

摘要：

Chipkill是动态随机存储器系统中先进的容错手段，ReedSolomon（RS）码是实现Chipkill技术的良好编码。以18片DDR3×8存储器芯片为研究对象，首先提出了一种快速构造RS码生成矩阵并对其优化的算法，获得了仅有1 728个“1”的生成矩阵；然后设计了一种纠单符号错RS码高效译码电路，实现了用纯组合逻辑完成检错、纠错操作。将所提出的RS码与传统的SEC/DED汉明码进行对比。开销方面，译码电路面积小19%~27%，延迟仅高出6%~27%。检错纠错能力方面，减少39.76%的可检不可纠错误，并且在实验中未出现不可检错误。

关键词: Chipkill, ReedSolomon码, 生成矩阵优化, 高效译码

Abstract:

Chipkill is an advanced fault tolerance approach in dynamic random access memory system and Reed-Solomon (RS) code is a well errorcorrecting code to achieve chipkill.The memory system of eighteen DDR3 X8 chips is taken as an example.Firstly,an algorithm for fast producing and optimizing the generation matrix of RS code is proposed,and a generation matrix with only 1 728 ones is obtained.Secondly, an efficient decoding circuit for singlesymbol correction RS code is proposed,which completes error detection and correction just by using combination logic.Compared with the traditional SEC/DED Hamming code,the proposed decoding circuit of RS code reduces the area by 19%~27% with only 6%~7% latency increase and cuts down errors,which can be detected but not corrected, by 39.76%.And there are no undetected errors in the experiment.

Key words: Chipkill;Reed-Solomon code;generation matrix optimization;efficient decoding

王谛，许勇. 一种面向DDR3接口的新型Chipkill编码[J]. J4, 2014, 36(12): 2386-2393.

WANG Di，XU Yong. A novel Chipkill code for DDR3 interface [J]. J4, 2014, 36(12): 2386-2393.

参考文献［15］

［1］	Dell T J. A white paper on the benefits of Chipkillcorrect ECC for PC server main memory［Z］. New York:IBM Microelectronics Division, 1997.
［2］	Schroeder B, Pinheiro E, Weber W D. DRAM errors in the wild:A largescale field study［C］∥Proc of ACM SIGMETRICS Performance Evaluation Review, 2009, 37(1):193204.
［3］	Sridharan V, Liberty D. A study of DRAM failures in the field［C］∥Proc of IEEE International Conference for High Performance Computing, Networking, Storage and Analysis, 2012:111.
［4］	Neuberger G, de Lima Kastensmidt F G, Reis R. An automatic technique for optimizing ReedSolomon codes to improve fault tolerance in memories［J］. IEEE Design & Test, 2005, 22(1):5058.
［5］	Sridharan V, Liberty D. A field study of DRAM errors［J］. Studies, 2012, 3(5):1.
［6］	Hwang A A,Stefanovici I A,Schroeder B.Cosmic rays don’t strike twice:Understanding the nature of DRAM errors and the implications for system design［J］. ACM SIGARCH Computer Architecture News, 2012, 40(1):111122.
［7］	IBM Chipkill memory—advanced ECC memory for the IBM Netfinity 7000 M10 ［Z］. New York:IBM Personal Computer Company, 1999.
［8］	Locklear D. Chipkill correct memory architecture［Z］. Texas, USA:Dell Enterprise Systems Group, 2000.
［9］	Abdoo D G, Cabello J D. Error correction system for N bits using error correcting code designed for fewer than N bits:USA,U.S. Patent 5 490 155［P］. 19960206.
［10］	Reed I S, Solomon G. Polynomial codes over certain finite fields［J］. Journal of the Society for Industrial & Applied Mathematics, 1960, 8(2):300304.
［11］	Sarwate D V, Shanbhag N R. Highspeed architectures for ReedSolomon decoders［J］. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2001, 9(5):641655.
［12］	Song M A, Kuo S Y. A low complexity design of Reed Solomon code algorithm for advanced RAID system［J］. IEEE Transactions on Consumer Electronics, 2007, 53(2):265273.
［13］	Lee H.An areaefficient Euclidean algorithm block for ReedSolomon decoder［C］∥Proc of IEEE Annual Symposium on VLSI, 2003:209210.
［14］	Raghupathy A, Liu K J R. Algorithmbased lowpower/highspeed ReedSolomon decoder design［J］. IEEE Transactions on Circuits and Systems II:Analog and Digital Signal Processing, 2000, 47(11):12541270.
［15］	Cormen T H, Leiserson C E, Rivest R L, et al. Introduction to algorithms［M］. Cambridge:MIT press, 2001.