A Ferroelectric FET-Based Processing-in-Memory Architecture for DNN Acceleration

被引：42

作者：

Long, Yun ^{[1
]}

Kim, Daehyun ^{[1
]}

Lee, Edward ^{[1
]}

Saha, Priyabrata ^{[1
]}

Mudassar, Burhan Ahmad ^{[1
]}

She, Xueyuan ^{[1
]}

Khan, Asif Islam ^{[1
]}

Mukhopadhyay, Saibal ^{[1
]}

机构：

[1] Georgia Inst Technol, Sch Elect & Comp Engn, Atlanta, GA 30332 USA

来源：

IEEE JOURNAL ON EXPLORATORY SOLID-STATE COMPUTATIONAL DEVICES AND CIRCUITS | 2019年 / 5卷 / 02期

基金：

美国国家科学基金会;

关键词：

Virtual machine monitors; Engines; Clocks; Transistors; Logic gates; Computer architecture; Random access memory; Deep neural network (DNN); ferroelectric FET (FeFET); processing-in-memory (PIM);

D O I：

10.1109/JXCDC.2019.2923745

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

This paper presents a ferroelectric FET (FeFET)-based processing-in-memory (PIM) architecture to accelerate the inference of deep neural networks (DNNs). We propose a digital in-memory vector-matrix multiplication (VMM) engine design utilizing the FeFET crossbar to enable bit-parallel computation and eliminate analog-to-digital conversion in prior mixed-signal PIM designs. A dedicated hierarchical network-on-chip (H-NoC) is developed for input broadcasting and on-the-fly partial results processing, reducing the data transmission volume and latency. Simulations in 28-nm CMOS technology show 115xand 6.3xhigher computing efficiency (GOPs/W) over desktop GPU (NvidiaGTX1080Ti) and resistive random access memory (ReRAM)-based design, respectively.

引用

页码：113 / 122

页数：10

共 50 条

[1] FELIX: A Ferroelectric FET Based Low Power Mixed-Signal In-Memory Architecture for DNN Acceleration
Soliman, Taha
Laleni, Nellie
Kirchner, Tobias
Mueller, Franz
Shrivastava, Ashish
Kaempfe, Thomas
Guntoro, Andre
Wehn, Norbert
[J]. ACM TRANSACTIONS ON EMBEDDED COMPUTING SYSTEMS, 2022, 21 (06)
[2] RETRANSFORMER: ReRAM-based Processing-in-Memory Architecture for Transformer Acceleration
Yang, Xiaoxuan
Yan, Bonan
Li, Hai
Chen, Yiran
[J]. 2020 IEEE/ACM INTERNATIONAL CONFERENCE ON COMPUTER AIDED-DESIGN (ICCAD), 2020,
[3] ReRAM-Based Processing-in-Memory Architecture for Recurrent Neural Network Acceleration
Long, Yun
Na, Taesik
Mukhopadhyay, Saibal
[J]. IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, 2018, 26 (12) : 2781 - 2794
[4] VSPIM: SRAM Processing-in-Memory DNN Acceleration via Vector-Scalar Operations
Nie, Chen
Tang, Chenyu
Lin, Jie
Hu, Huan
Lv, Chenyang
Cao, Ting
Zhang, Weifeng
Jiang, Li
Liang, Xiaoyao
Qian, Weikang
Sun, Yanan
He, Zhezhi
[J]. IEEE TRANSACTIONS ON COMPUTERS, 2024, 73 (10) : 2378 - 2390
[5] Study on Processing-in-Memory Technology based on Dataflow Architecture
Choi, Kyu Hyun
Hwang, Taeho
[J]. 2022 INTERNATIONAL CONFERENCE ON ELECTRONICS, INFORMATION, AND COMMUNICATION (ICEIC), 2022,
[6] A DNN Compression Framework for SOT-MRAM-based Processing-In-Memory Engine
Yuan, Geng
Ma, Xiaolong
Lin, Sheng
Li, Zhengang
Deng, Jieren
Ding, Caiwen
[J]. 2020 IEEE 33RD INTERNATIONAL SYSTEM-ON-CHIP CONFERENCE (SOCC), 2020, : 37 - 42
[7] PIM-DH: Re RAM-based Processing-in-Memory Architecture for Deep Hashing Acceleration
Liu, Fangxin
Zhao, Wenbo
Chen, Yongbiao
Wang, Zongwu
He, Zhezhi
Yang, Rui
Tang, Qidong
Yang, Tao
Zhuo, Cheng
Jiang, Li
[J]. PROCEEDINGS OF THE 59TH ACM/IEEE DESIGN AUTOMATION CONFERENCE, DAC 2022, 2022, : 1087 - 1092
[8] ReRAM-based Processing-in-Memory Architecture for Blockchain Platforms
Wang, Fang
Shen, Zhaoyan
Han, Lei
Shao, Zili
[J]. 24TH ASIA AND SOUTH PACIFIC DESIGN AUTOMATION CONFERENCE (ASP-DAC 2019), 2019, : 615 - 620
[9] PIMCH: Cooperative Memory Prefetching in Processing-In-Memory Architecture
Xui, Sheng
Wang, Ying
Han, Yinhe
Li, Xiaowei
[J]. 2018 23RD ASIA AND SOUTH PACIFIC DESIGN AUTOMATION CONFERENCE (ASP-DAC), 2018, : 209 - 214
[10] Comparative Analysis and Energy-Efficient Write Scheme of Ferroelectric FET-Based Memory Cells
Ko, Dong Han
Oh, Tae Woo
Lim, Sehee
Kim, Se Keon
Jung, Seong-Ook
[J]. IEEE ACCESS, 2021, 9 : 127895 - 127905

← 1 2 3 4 5 →