基於混合式相變化儲存裝置之特徵感知寫入策略

A Pattern-Aware Write Strategy for Hybrid PCM Storage Devices

黃祥智(Xiang-Zhi Huang)

指導教授 ： 郭大維

共同指導教授 ： 張原豪

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2015年)

https://doi.org/10.6342/NTU.2015.00598

摘要

中文摘要

電腦架構演變至今,有許許多多的候選者被期待於取代目前的硬盤儲存裝置,在這些候選者當中相變化記憶體是非常備受期待的次世代儲存裝置,由於他有著高密度和極低的使用延遲所以非常適合作為現代架構下的儲存裝置。但是相變化記憶體的高耗能是一個非常嚴重的致命傷,而目前的寫回策略中並大多都是以減少頁缺失次數為目標,並沒有以最佳化儲存系統耗能為出發,另外相變化記憶體的耗能和他所存的資料特徵有關,因此本文提出一個考量儲存資料特徵並且以減少耗能為目標的寫回策略。
關鍵字: 相變化記憶體,儲存裝置系統, 耗能最小化, 寫回策略, 特徵感知

關鍵字

相變化記憶體儲 ； 存裝置系統 ； 耗能最小化 ； 寫回策略 ； 特徵感知

並列摘要

ABSTRACT
Phase change memory (PCM) is a potential candidate on the storage applications due to its nanosecond-level access latency and byte-addressability. In addition, with the help of multiple-level-per- cell (MLC) technology, PCM could provide comparable capacity to flash memory. However, adopting MLC PCM needs much larger power consumption than SLC PCM. Thus, in this paper, we exploit a SLC/MLC hybrid memory architecture with the proposed pattern-aware write back policy to minimize the energy consumption on the storage devices In addition, we also propose a counter buffer design to reduce the cost on manipulating data structures, and meanwhile, we design a data migration mechanism to migrate data to MLC PCM when the space of SLC PCM is exhausted. We conducted the experiments on the well-known benchmarks and for which the results are encourage.

並列關鍵字

Phase change memory ； Storage Device ； Minimize energy consumption ； Write back Policy ； Pattern-aware

參考文獻

[1] Mengying Zhao , Yuan Xue, Xue.C.j, Minimizing MLC PCM Write Energy for Free through Profiling-based State Remapping, ASP-DAC 2015.

[2] Jue Wang, Xiangyu Dong, Guangyu Sun, Dimin Niu, Yuan Xie, Energy-Efficient Multi-Level Cell Phase-Change Memory System with Data Encoding, ICCD 2011.

[4] Jianhui Yue, Yifeng Zhu, Exploiting Subarrays Inside a Bank to Improve Phase Change Memory Performance, DATE 2013.

[5] Ping Zhou, Bo Zhao, A Durable and Energy Efficient Main Memory Using Phase Change Memory Technology, ISCA 2009

[7] Z. Shao, Y. Liu, Y. Chen, and T. Li. Utilizing pcm for energy optimization in embedded systems. In VLSI (ISVLSI), 2012 IEEE Computer Society Annual Symposium on, pages 398–403, Aug 2012.