Exploiting Resistance Drift Characteristics to Improve Reliability of LDPC-Assisted Phase-Change Memory

被引:3
作者
Zhang, Meng [1 ,2 ,3 ]
Wu, Fei [1 ,2 ,3 ]
Yu, Qin [1 ,2 ,3 ]
Liu, Weihua [1 ,2 ,3 ]
Ma, Ruixiang [1 ,2 ,3 ]
Xie, Changsheng [1 ,2 ,3 ]
机构
[1] Minist Educ China, Key Lab Informat Storage Syst, Wuhan Natl Lab Optoelect, Beijing, Peoples R China
[2] Huazhong Univ Sci & Technol, Engn Res Ctr Data Storage Syst & Technol, Wuhan 430074, Peoples R China
[3] Huazhong Univ Sci & Technol, Shenzhen Res Inst, Shenzhen 518000, Peoples R China
基金
中国国家自然科学基金; 中国博士后科学基金;
关键词
Resistance; Phase change materials; Iterative decoding; Programming; Memory; Electrodes; Simulation; Phase change memory; resistance drift characteristics; raw bit error rates; LDPC codes; decoding iterations;
D O I
10.1109/TDMR.2021.3087807
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Phase-change memory (PCM) as emerging non-volatile memory has attracted more attention and considered as the promising replacement of the main memory. PCM has shown good scalability and high storage density, but data storage reliability has become a challenge and concern. When data are written into PCM cells by a phase transition between amorphous and crystalline, the resistance of each state drifts as the increase of storage time due to structural relaxation. As a result, raw bit error rates (RBER) become higher and higher, severely degrading the data storage reliability (i.e., an important problem in PCM). In this paper, we first find that high error percentage exists between the full amorphous and amorphous states in multilevel cell (MLC) PCM via a preliminary experiment, which is the main factor leading to high RBER. Then, we analyze why this phenomenon exists in PCM from the view of the resistance drift. Finally, by exploiting the resistance drift characteristics, we propose drift compensation and drift-aware LDPC decoding schemes to improve reliability of PCM. Simulation results show that the proposed schemes can significantly reduce the RBER and LDPC decoding iterations.
引用
收藏
页码:324 / 330
页数:7
相关论文
共 19 条
  • [1] Chien W. C., 2016, IEEE INT EL DEV M IE, P552
  • [2] Huang C. H., 2020, POLYM REV, P1, DOI DOI 10.1080/15583724.2019.1688830
  • [3] A Reliable 3D MLC PCM Architecture with Resistance Drift Predictor
    Jalili, Majid
    Arjomand, Mohammad
    Sarbazi-Azad, Hamid
    [J]. 2014 44TH ANNUAL IEEE/IFIP INTERNATIONAL CONFERENCE ON DEPENDABLE SYSTEMS AND NETWORKS (DSN), 2014, : 204 - 215
  • [4] Krebs D., 2015, IEEE INT REL PHYS S, pMY51
  • [5] Kwon T, 2018, DES AUT TEST EUROPE, P1610, DOI 10.23919/DATE.2018.8342272
  • [6] Lee C., 2019, P IEEE ACM INT C COM, P1
  • [7] Improved Co-Simulation with Event Detection for Stochastic Behaviors of CPSs
    Liu, Jufu
    Jiang, Kaiqiang
    Wang, Xiao
    Cheng, Bei
    Du, Dehui
    [J]. PROCEEDINGS 2016 IEEE 40TH ANNUAL COMPUTER SOFTWARE AND APPLICATIONS CONFERENCE WORKSHOPS, VOL 1, 2016, : 209 - 214
  • [8] Phase Change Memory Reliability: A Signal Processing and Coding Perspective
    Pozidis, Haralampos
    Mittelholzer, Thomas
    Papandreou, Nikolaos
    Parnell, Thomas
    Stanisavljevic, Milos
    [J]. IEEE TRANSACTIONS ON MAGNETICS, 2015, 51 (04)
  • [9] Improving MLC PCM Performance through Relaxed Write and Read for Intermediate Resistance Levels
    Rashidi, Saeed
    Jalili, Majid
    Sarbazi-Azad, Hamid
    [J]. ACM TRANSACTIONS ON ARCHITECTURE AND CODE OPTIMIZATION, 2018, 15 (01)
  • [10] Shihao Song, 2020, ISMM 2020: Proceedings of the 2020 ACM SIGPLAN International Symposium on Memory Management, P30, DOI 10.1145/3381898.3397210