A 22nm 16Mb Floating-Point ReRAM Compute-in-Memory Macro with 31.2TFLOPS/W for AI Edge Devices

📋 论文概要

本文提出了一款基于ReRAM的16Mb浮点存内计算宏，在22nm工艺下实现了31.2TFLOPS/W的能效，旨在满足AI边缘设备对高精度推理（如FP16/BF16）和高能效、低待机功耗的需求。该宏通过创新的电路和架构设计，解决了传统存内计算在浮点精度和能效之间的权衡问题。

💡 主要创新点

• 首次在22nm工艺下实现16Mb容量的浮点ReRAM存内计算宏，支持FP16/BF16精度。
• 通过优化的模拟计算电路和数字接口，实现了31.2TFLOPS/W的极高能效，显著优于传统数字加速器。
• 采用低待机功耗设计，延长了电池供电AI边缘设备的续航时间。

🏷 关键词

📄 原文摘要

Zhao-En Ke1, Yu-Hsiang Chin1, Hua-Jin Wen1, Yu-Chen Chang1, Wei-Ting Hsu1, Chung-Chuan Lo1, Ren-Shuo Liu1, Chih-Cheng Hsieh1, Kea-Tiong Tang1, Shih-Hsin Teng3, Chung-Cheng Chou3, Yu-Der Chih3, Tsung-Yung Jonathan Chang3, Meng-Fan Chang1,2 National Tsing Hua University, Hsinchu, Taiwan TSMC Corporate Research, Hsinchu, Taiwan 3 TSMC, Hsinchu, Taiwan *Equally Credited Authors (ECAs) 1 2 AI-edge devices demand high-precision computation (e.g. FP16 and BF16) for accurate inference in practical applications, while maintaining high energy efficiency (EF) and low standby power to prolong battery life. Thus, advanced nonvolatile AI-edge processors [1,2] require nonvolatile compute-in-memory (nvCIM) [3-5] with a large nonvolatile onchip memory, to store all of the neural network’s parameters (weight data) during power-off, and high-precision high-EF multiply-and-accumulate (MAC) operations during compute, to maximize battery life. Among nvCIMs, ReRAM-nvCIM stands out as a

👥 作者与机构

Tai-Hao Wen*1, Hung-Hsi Hsu*1,2, Win-San Khwa*2, Wei-Hsing Huang1,