IDWA: A Importance-Driven Weight Allocation Algorithm for Low Write–Verify Ratio RRAM-Based In-Memory Computing

Jingyuan Qu

doi:10.1109/TVLSI.2025.3578388

PRE

DOI：10.1109/TVLSI.2025.3578388

原文链接

原文求助

摘要

En 中文

基于电阻随机存取存储器（RRAM）的存内计算（IMC）架构目前正受到广泛关注。由于这种计算方法依赖于器件的模拟特性，RRAM的写入变化会在不同程度上影响计算精度。传统的写入-验证（W&V）流程针对所有权重参数执行，导致显著的时间开销。为解决此问题，我们提出了一种训练算法，能够在较低W&V开销成本下恢复受写入变化影响的离线IMC精度。我们在线性网络的训练过程中引入了基于重要性的权重分配（IDWA）算法。该算法约束次要权重值，以抑制变化干扰在该部分权重中的扩散，从而减少不必要的精度下降。此外，我们采用逐层优化算法识别神经网络中需进行W&V操作的重要权重。针对各类深度神经网络（DNN）架构和数据集的广泛测试表明，我们提出的选通W&V方法在精度保持和计算效率方面均优于当前最先进的选通W&V技术。在相同精度水平下，其速度提升相较于其他先进方法可达6倍至32倍。

Keyword:

Importance-driven weight allocation (IDWA)

in-memory computing (IMC)

resistive random access memory (RRAM)

variation

write–verify (W&V)

期刊

IEEE Transactions on Very Large Scale Integration (VLSI) Systems

IF:

3.1

论文数:

487

被引数:

7.3K

机构

harbin institute of technology

学者数:

8.0W

论文数: 6.6W

被引数: 66

引用论文

Multi-level switching of triple-layered TaOx RRAM with excellent reliability for storage class memory

2012 Symposium on VLSI Technology (VLSIT)

IF0

2012-06-01

PREAI

Seung Ryul Lee; Young-Bae Kim; Man Chang; Kyung Min Kim; Chang Bum Lee; Ji Hyun Hur; Gyeong-Su Park; Dongsoo Lee; Myoung-Jae Lee; Chang Jung Kim; U-In Chung; In-Kyeong Yoo; Kinam Kim

Vehicle Detection From UAV Imagery With Deep Learning: A Review 基于深度学习的无人机图像车辆检测研究进展

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS

IF8.9

2022-11-01

PREAI

Bouguettaya, Abdelmalek; Zarzour, Hafed; Kechida, Ahmed; Taberkit, Amine Mohammed

On the Reliability of RRAM-Based Neural Networks 基于RRAM的神经网络的可靠性

2023 IFIP/IEEE 31st International Conference on Very Large Scale Integration (VLSI-SoC)

IF0

2023-10-16

OAAI

Hassen Aziza; Cristian Zambelli; Said Hamdioui; Sumit Diware; Rajendra Bishnoi; Anteneh Gebregiorgis

Device Non-Ideality Effects and Architecture-Aware Training in RRAM In-Memory Computing Modules

2021 IEEE International Symposium on Circuits and Systems (ISCAS)

IF0

2021-05-01

PREAI

Qiwen Wang; Yongmo Park; Wei D. Lu

A crossbar array of magnetoresistive memory devices for in-memory computing 用于内存计算的磁阻存储器器件的交叉阵列

Nature

IF0

2022-01-12

PREAI

Seungchul Jung; Hyungwoo Lee; Sungmeen Myung; Hyunsoo Kim; Seung Keun Yoon; Soon-Wan Kwon; Yongmin Ju; Minje Kim; Wooseok Yi; Shinhee Han; Baeseong Kwon; Boyoung Seo; Kilho Lee; Gwan-Hyeob Koh; Kangho Lee; Yoonjong Song; Changkyu Choi; Donhee Ham; Sang Joon Kim

Improved Switching Stability and the Effect of an Internal Series Resistor in HfO₂/TiO_{<italic>x</italic>} Bilayer ReRAM Cells

IEEE Transactions on Electron Devices

IF0

2018-08-01

PREAI

Alexander Hardtdegen; Camilla La Torre; Felix Cuppers; Stephan Menzel; Rainer Waser; Susanne Hoffmann-Eifert

NEAT: Nonlinearity Aware Training for Accurate, Energy-Efficient, and Robust Implementation of Neural Networks on 1T-1R Crossbars

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

IF0

2022-08-01

OAAI

Abhiroop Bhattacharjee; Lakshya Bhatnagar; Youngeun Kim; Priyadarshini Panda

Two-step write–verify scheme and impact of the read noise in multilevel RRAM-based inference engine

Semiconductor Science and Technology

IF0

2020-10-10

PREAI

Wonbo Shim; Jae-sun Seo; Shimeng Yu

33.2 A Fully Integrated Analog ReRAM Based 78.4TOPS/W Compute-In-Memory Chip with Fully Parallel MAC Computing

2020 IEEE International Solid- State Circuits Conference - (ISSCC)

IF0

2020-02-01

PREAI

Qi Liu; Bin Gao; Peng Yao; Dong Wu; Junren Chen; Yachuan Pang; Wenqiang Zhang; Yan Liao; Cheng-Xin Xue; Wei-Hao Chen; Jianshi Tang; Yu Wang; Meng-Fan Chang; He Qian; Huaqiang Wu

A Compact Model for Metal–Oxide Resistive Random Access Memory With Experiment Verification

IEEE Transactions on Electron Devices

IF0

2016-05-01

OAAI

Zizhen Jiang; Yi Wu; Shimeng Yu; Lin Yang; Kay Song; Zia Karim; H.-S. Philip Wong

查看更多内容