基于PCM的GPU存储系统设计与优化

计算机科学 ›› 2013, Vol. 40 ›› Issue (10): 29-31.

基于PCM的GPU存储系统设计与优化

穆帅,单书畅,邓仰东,王志华

清华大学微电子所北京100084;中国科学院计算技术研究所北京100084;清华大学微电子所北京100084;清华大学微电子所北京100084

出版日期:2018-11-16 发布日期:2018-11-16
基金资助:
本文受国家自然科学基金(61272085)资助

Evaluating and Optimizing of PCM Based GPU Memory Architecture

MU Shuai,SHAN Shu-chang,DENG Yang-dong and WANG Zhi-hua

Online:2018-11-16 Published:2018-11-16

摘要/Abstract

摘要： 以相变存储器(PCM)为代表的新型非易失存储器,具有存储密度高和静态功耗低等传统动态随机存取存储器(DRAM)不具备的优势,但是过长的写操作延时会严重影响访存的性能。设计了基于PCM的图形处理器(GPU)中的存储系统。仿真结果显示,GPU程序中的内存写请求分布极不均匀,对少量的内存地址有非常高的访问频率。面向访存分布不均匀特点的专用缓冲单元设计,能够有效地存储频繁访问的内存数据,从而减少对PCM的访问次数,消除过长的写操作延时对系统性能的负面影响。GPU仿真器上的结果显示,基于缓冲单元的PCM存储系统能够有效地提高GPU的运算性能。

关键词: 相变存储器,图形处理器,缓冲单元,存储系统

Abstract: Recently,the emerging non-volatile memory(NVM) exemplified by Phase Change Memory(PCM) has been considered to take the place of conventional DRAM in the processors due to their characteristics of large capacity and little static power．However,the overhead of long write latency can cause severe performance degradation．We evaluated the feasibility of PCM based GPU memory architecture．Based on the analysis of unique memory access behaviors captured from GPU benchmarks,a dedicated buffer was designed to alleviate the pressure of frequently accessing PCM．Simulation results prove the efficiency of our proposed dedicated buffer and show its great potential for PCM based GPU architecture.

Key words: Phase change memory(PCM),Graphics processing unit(GPU),Buffer,Memory hierarchy

穆帅,单书畅,邓仰东,王志华. 基于PCM的GPU存储系统设计与优化[J]. 计算机科学, 2013, 40(10): 29-31. https://doi.org/

MU Shuai,SHAN Shu-chang,DENG Yang-dong and WANG Zhi-hua. Evaluating and Optimizing of PCM Based GPU Memory Architecture[J]. Computer Science, 2013, 40(10): 29-31. https://doi.org/

参考文献

[1] Hennessy J L,Patterson D A．Computer Architecture A Quantitative Approach [M].San Mateo:Morgan Kaufmann,2011
[2] Freitas R,Wilcke W．Storage-class memory:The next storage system technology [J]．IBM Journal of Research and Development,2008,2(4/5):439-447
[3] 刘金磊,李琼．新型非易失相变存储器PCM应用研究[J]．计算机研究与发展,2012,49(S1):90-93
[4] Lee B C,Ipek E,Mutlu O,et al.Architecting Phase ChangeMemory as a Scalable DRAM Architecture[J]．Proc．of Interna-tional Symposium on Computer Architecture,2009,37(3):2-13
[5] Qureshi M K,Srinivasan V,Rivers J A．Scalable High Perfor-mance Main Memory System Using Phase Change Memory Technology [J]．Proc．of International Symposium on Computer Architecture,2009,7(3):24-33
[6] Kim D,Lee S,Chung J,et al.Hybrid DRAM/PRAM-basedMain Memory for Single-Chip CPU/GPU [C]∥Proc．of Design Automation Conference．2012:888-896
[7] 郝秀蕊,安虹,李小强,等．流处理器的相变存储器主存性能优化[J]．计算机工程,2011,7(24):251-256
[8] Raoux R,Burr G W,Breitwisch M J,et al．Phase Change Random Access Memory A Scalable Technology [J]．IBM Journal of Research and Development,2008,2(4/5):465-479
[9] Dong X Y,Jouppi N P,Xie Y.PCRAMsim System Level Performance Energy and Area Modeling for Phase Change RAM [C]∥Proc．of the International Conference on Computer Aided Design．2009:269-275
[10] Bakhoda A,Yuan G L,Fung W L,et al.Analyzing CUDAWorkloads Using a Dedicated GPU Simulator [C]∥Proc．of International Symposium on Performance Analysis of Systems and Software．2009:163-171

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed