In-memory Bulk Bitwise Logic Operation for Multi-level Cell Non-volatile Memories

被引:0
作者
Salehi, Sayed Ahmad [1 ]
机构
[1] Univ Kentucky, Elect & Comp Engn Dept, Lexington, KY 40506 USA
来源
PROCEEDINGS OF THE INTERNATIONAL SYMPOSIUM ON MEMORY SYSTEMS, MEMSYS 2022 | 2022年
关键词
Processing-in-memory (PIM); Memory architecture; Bulk bitwise logic operation; Multi-level cell; Non-volatile memories;
D O I
10.1145/3565053.3565058
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Multi-level cell (MLC) memories are of particular interest due to their high density and low per-bit cost. The significant contrast between high and low resistance states in phase change memory (PCM) technology supports its MLC capability. In this paper, we propose a processing-in-memory (PIM) structure for resistance-based, including PCM, MLC (2-bit/cell) memories. With slight modification of the existing peripheral circuitry of PCM, the proposed structure enables performing in-memory bulk logic operations. The modified peripheral circuitry activates two rows of PCM cells simultaneously, reads their bitlines, and converts them to desired outputs to compute the bitwise operation for pairs of bits stored in the activated cells. Unlike PIM structures based on single-level cell (SLC) that process rows of single bits in each memory access, the proposed MLC-based structure processes rows of 2 bits in each memory access. This leads to a significant saving in computational time and energy. Our evaluation for bitwise vector OR operations shows significant improvement in computational speed and energy compared to state-of-the-art work using SLC. Future development of software support for the proposed structure, will enable us to efficiently perform complex image processing and deep neural network algorithms in memory.
引用
收藏
页数:5
相关论文
共 18 条
[1]  
[Anonymous], 2019, Intel Optane Memory
[2]   Multilevel-Cell Phase-Change Memory: A Viable Technology [J].
Athmanathan, Aravinthan ;
Stanisavljevic, Milos ;
Papandreou, Nikolaos ;
Pozidis, Haralampos ;
Eleftheriou, Evangelos .
IEEE JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND SYSTEMS, 2016, 6 (01) :87-100
[3]  
Burr Geoffrey W., 2010, IEEE Globecom
[4]   PRIME: A Novel Processing-in-memory Architecture for Neural Network Computation in ReRAM-based Main Memory [J].
Chi, Ping ;
Li, Shuangchen ;
Xu, Cong ;
Zhang, Tao ;
Zhao, Jishen ;
Liu, Yongpan ;
Wang, Yu ;
Xie, Yuan .
2016 ACM/IEEE 43RD ANNUAL INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE (ISCA), 2016, :27-39
[5]   In-Memory Intelligence [J].
Finkbeiner, Tim ;
Hush, Glen ;
Larsen, Troy ;
Lea, Perry ;
Leidel, John ;
Manning, Troy .
IEEE MICRO, 2017, 37 (04) :30-38
[6]  
Hruska Joel, 2014, Western Digital's HGST Division Creates New Phase-Change SSD that's Orders of Magnitude Faster than any NAND Flash Drive on the Market
[7]   Memory- and time-efficient dense network for single-image super-resolution [J].
Imanpour, Nasrin ;
Naghsh-Nilchi, Ahmad R. ;
Monadjemi, Amirhassan ;
Karshenas, Hossein ;
Nasrollahi, Kamal ;
Moeslund, Thomas B. .
IET SIGNAL PROCESSING, 2021, 15 (02) :141-152
[8]   Programmable Resistance Switching in Nanoscale Two-Terminal Devices [J].
Jo, Sung Hyun ;
Kim, Kuk-Hwan ;
Lu, Wei .
NANO LETTERS, 2009, 9 (01) :496-500
[9]   DRISA: A DRAM-based Reconfigurable In-Situ Accelerator [J].
Li, Shuangchen ;
Niu, Dimin ;
Malladi, Krishna T. ;
Zheng, Hongzhong ;
Brennan, Bob ;
Xie, Yuan .
50TH ANNUAL IEEE/ACM INTERNATIONAL SYMPOSIUM ON MICROARCHITECTURE (MICRO), 2017, :288-301
[10]  
Li SC, 2016, DES AUT CON, DOI [10.1145/2897937.2898064, 10.1109/ICAUMS.2016.8479697]