A reconfigurable in‐memory‐computation architecture with in‐situ update and shift capability-Reference-Cited by-同舟云学术

A reconfigurable in‐memory‐computation architecture with in‐situ update and shift capability

Published:2024-04 Issue:8 Volume:60 Page:
ISSN:0013-5194
Container-title:Electronics Letters
language:en
Short-container-title:Electronics Letters

Author:

Liu Yihe¹^ORCID,Wang Junjie¹,Liu Shuang¹,Zhang Xiaoyang²,Sun Mingyuan³,Chen Tupei⁴,Liu Yang¹

Affiliation:

1. State Key Laboratory of Electronic Thin Films and Integrated Devices University of Electronic Science and Technology of China Chengdu P. R. China

2. Science and Technology on Millimeter‐wave Laboratory Beijing Institute of Remote‐Sensing Equipment Beijing P. R. China

3. Chang Feng Mechanics and Electronics Technology Academy Beijing P. R. China

4. School of Electrical and Electronic Engineering Nanyang Technological University Singapore Singapore

Abstract

AbstractThis work proposes a storage element (SE) design for in‐memory computing (IMC). Using the proposed SE design, an IMC array has been constructed to enable in‐situ updates of stored weights. Compared with some existing related works which employ 2n bit cells for storing an n‐bit value, the proposed structure in this work exhibits a O(n) complexity of area overhead, which means that only n bit cells are needed to store n‐bit values, offering a significant improvement in silicon area usage. Compared to the purely digitally‐controlled weight update schemes, the approach proposed in this work demonstrates an approximate 1.47× increase in power consumption. Furthermore, it exhibits superior robustness compared to existing work. Additionally, the proposed SE design can achieve the necessary shift functionality in the cutting‐edge floating‐point (FP)‐IMC architecture through simple control mechanisms.

Publisher

Institution of Engineering and Technology (IET)

Reference7 articles.

1. Compute-in-Memory Technologies and Architectures for Deep Learning Workloads

2. SRAM-Based Computing-in-Memory Macro With Fully Parallel One-Step Multibit Computation

3. An Area- and Energy-Efficient Spiking Neural Network With Spike-Time-Dependent Plasticity Realized With SRAM Processing-in-Memory Macro and On-Chip Unsupervised Learning

4. A 65 nm thermometer‐encoded time/charge‐based compute‐in‐memory neural network accelerator at 0.735 pJ/MAC and 0.41 pJ/Update;Gong M.;IEEE Trans. Circuits Syst. II Express Briefs,2020

5. Wu P.‐C. et al.:A 22nm 832Kb hybrid‐domain floating‐point SRAM in‐memory‐compute macro with 16.2‐70.2 TFLOPS/W for high‐accuracy AI‐edge devices. In:2023 IEEE International Solid‐State Circuits Conference (ISSCC).IEEE(2023)