An 8-Mb DC-Current-Free Binary-to-8b Precision ReRAM Nonvolatile Computing-in-Memory Macro using Time-SpaceReadout with 1286.4 - 21.6TOPS/W for Edge-AI Devices

📋 论文概要

本文提出了一种8Mb无直流电流的ReRAM非易失性计算存储宏，采用时间-空间读出方法，支持二进制到8位精度的乘积累加操作。该设计解决了边缘AI设备中高精度、低功耗非易失性计算的需求。

💡 主要创新点

• 首次实现无直流电流的ReRAM计算存储宏，降低了静态功耗
• 提出时间-空间读出技术，支持二进制到8位精度的灵活配置，兼顾精度与能效
• 通过优化电路架构实现高能效和短计算延迟，适用于电池供电的边缘AI系统

🏷 关键词

📄 原文摘要

Tai-Hao Wen1, Chin-I Su2, Win-San Khwa2, Chung-Chuan Lo1, Ren-Shuo Liu1, Chih-Cheng Hsieh1, Kea-Tiong Tang1, Yu-Der Chih2, Tsung-Yung Jonathan Chang2, Meng-Fan Chang1,2 National Tsing Hua University, Hsinchu, Taiwan TSMC, Hsinchu, Taiwan 1 2 Battery-powered edge-AI devices require nonvolatile computing-in-memory (nvCIM) macros for nonvolatile data storage and multiply-and-accumulate (MAC) operations. High inference accuracy requires MAC operations with high input (IN), weight (W), and output (OUT) precisions. A high energy efficiency (EFMAC) and a short computing latency (tAC) are also required. Most existing silicon-verified nvCIM macros use current-mode signal generation; using current [1-3] or hybrid current-voltage readout schemes [4-5] for multibit MAC operations to compensate for the small BL-voltage swing and signal margin resulting from the low read-disturb-free voltage (VRD). As shown in Fig. 11.4.1, current-mode nvCIMs face various challenges: (1) a limited

👥 作者与机构

Je-Min Hung1, Yen-Hsiang Huang1, Sheng-Po Huang1, Fu-Chun Chang1,