利用多维分级Cache替换策略减少对PCM内存写回量

计算机工程与科学

利用多维分级Cache替换策略减少对PCM内存写回量

阮深沉，王海霞,汪东升

（清华大学信息技术与科学国家实验室，北京 100084）

收稿日期:2016-04-15 修回日期:2016-06-20 出版日期:2016-08-25 发布日期:2016-08-25
基金资助:
国家自然科学基金(61303002，61373025)

Using multi-dimensional hierarchical Cache replacement policy to reduce the write traffic to PCM memory

RUAN Shen-chen,WANG Hai-xia,WANG Dong-sheng

（National Laboratory for Information Science and Technology,Tsinghua University,Beijing 100084,China）

Received:2016-04-15 Revised:2016-06-20 Online:2016-08-25 Published:2016-08-25

摘要/Abstract

摘要：

寻找新型存储材料代替DRAM内存是当前的一个研究热点。相变存储PCM因其具有低功耗、高存储密度和非易失性的优点受到广泛的关注，然而PCM的可擦写次数有限，要用作内存必须考虑如何减少对其的写操作。针对该问题，一种有效的解决方法是优化Cache替换策略，减少Cache中脏块被替换出的数量。现有研究主要通过在插入和访问命中时给脏块设定较高的保护优先级来达到给脏块额外保护的目的，但是在降级过程中不再对脏块与干净块进行区分，这导致Cache可能在存在大量干净块的情况下仍然先替换脏块。提出一种新型的Cache替换策略MAC，它通过一个多维分级结构在脏块与干净块之间设置了不可逾越的界限，使得脏块能得到更有力的保护。模拟实验表明，相对LRU替换策略，MAC以较低的硬件开销代价平均减少约2512%的内存写，同时对程序运行性能几乎没有影响。

关键词: 相变存储, Cache替换策略, 写回量

Abstract:

Searching for novel storage materials is a current hotspot. The phase change memory (PCM) draws wide attention due to the advantages of low power consumption, high density and non-volatility, however, it incurs finite endurance. So it is necessary to consider how to reduce the write traffic to memory. As for this problem, a optimizing cache replacement policies to reduce the amount of dirty blocks evicted from cache is an efficient method. The existing works mainly achieve the goal of protecting dirty blocks by setting higher protection priority for them when they are inserted in or re-referenced, however, they make no distinction between dirty blocks and clean ones in the process of demotion, which may leads to the result that dirty blocks are preferentially evicted even though there are still many clean blocks. We propose a novel cache replacement policy called multilayer ark for cache (MAC), which sets insurmountable boundaries between the dirty cache blocks and the clean ones through a multidimensional hierarchy structure, and dirty blocks therefore can be protected more effectively. Simulation results show that compared with the LRU replacement policy, the MAC can averagely reduce the writes to memory by about 25.12% with low hardware cost while the program performance is hardly affected.

Key words: phase change memory, Cache replacement policy, write traffic

阮深沉，王海霞,汪东升. 利用多维分级Cache替换策略减少对PCM内存写回量[J]. 计算机工程与科学.

RUAN Shen-chen,WANG Hai-xia,WANG Dong-sheng. Using multi-dimensional hierarchical Cache replacement policy to reduce the write traffic to PCM memory [J]. Computer Engineering & Science.

参考文献

［1］Zhang X, Hu Q, Wang D, et al. A read-write aware replacement policy for phase change memory［M］∥Advanced Parallel Processing Technologies. Berlin: Springer Berlin Heidelberg, 2011:31-45.
［2］Lee B C,Zhou P,Yang J,et al.Phase-change technology and the future of main memory［J］.IEEE Micro,2010,30（1）: 143.
［3］Lee H G,Baek S,Nicopoulos C,et al.An energy-and performance-aware DRAM cache architecture for hybrid DRAM/PCM main memory systems［C］∥Proc of 2011 IEEE 29th International Conference onComputer Design,2011: 381-387.
［4］Rodríguez-Rodríguez R, Castro F, Chaver D, et al. Reducing writes in phase-change memory environments by using efficient cache replacement policies［C］∥Proc of the Conference on Design,Automation and Test in Europe,2013: 93-96.
［5］Zhou P,Zhao B,Yang J,et al.A durable and energy efficient main memory using phase change memory technology［C］∥Proc of International Symposium on Computer Architecture, 2009: 14-23.
［6］Cho S,Lee H.Flip-n-write: A simple deterministic technique to improve pram write performance,energy and endurance［C］∥Proc of the 42nd Annual IEEE/ACM International Symposium on Microarchitecture(MICRO-42),2009: 347-357.
［7］Fang Y,Li H,Li X.SoftPCM: Enhancing energy efficiency and lifetime of phase change memory in video applications via approximate write［C］∥Proc of 2012 IEEE 21st Asian Test Symposium （ATS）,2012: 131-136.
［8］Hu J,Xue C J,Zhuge Q,et al.Write activity reduction on non-volatile main memories for embedded chip multiprocessors［J］.ACM Transactions on Embedded Computing Systems （TECS）,2013,12（3）: 77.
［9］Qureshi M K,Srinivasan V,Rivers J A.Scalable high performance main memory system using phase-change memory technology［J］.ACM SIGARCH Computer Architecture News,2009,37（3）: 24-33.
［10］Lilly B P,Williams III G R,Sadoughi-yarandi M,et al.Least recently used mechanism for cache line eviction from a cache memory: U.S.Patent 20，150，026，404［P］.2015-01-22.
［11］Jaleel A,Theobald K B,Steely Jr S C,et al.High performance cache replacement using re-reference interval prediction （RRIP）［C］∥Proc of International Symposium on Computer Architecture, 2010: 60-71.
［12］Rodríguez-Rodríguez R,Castro F,Chaver D,et al.Reducing writes in phase-change memory environments by using efficient cache replacement policies［C］∥Proc of the Conference on Design,Automation and Test in Europe,2013: 93-96.
［13］Yoo S,Lee E,Bahn H.The least-dirty-first cache replacement policy for phase-change memory［C］∥Proc of the 29th Annual ACM Symposium on Applied Computing,2014: 1449-1454.
［14］Li Z, Jin P, Su X, et al.CCF-LRU: A new buffer replacement algorithm for flash memory［J］.Consumer Electronics,IEEE Transactions on,2009,55（3）: 1351-1359.

[1]	张震1,2，付印金1，胡谷雨1. 相变存储器写寿命延长关键技术研究进展[J]. 计算机工程与科学, 2018, 40(09): 1546-1555.
[2]	石文强，倪晓强，金作霖，张民选. Cache动态插入策略模型研究[J]. J4, 2013, 35(10): 159-165.