logo
科言猫学术研究的AI总结
language
手机端手机端
  • Phone Feature微信小程序
简体中文简体中文
notice消息通知
博客博客
delete
arrow
返回

DepAsync: An Asynchronous SNN Accelerator Based on Core-Dependency

delete2025-10-24
delete0
PRE
AI
Z
Zhuo Chen
D
De Ma
X
Xiaofei Jin
Q
Qinghui Xing
O
Ouwen Jin
X
Xin Du
S
Shuibing He
潘
潘纲 (Gang Pan)
DOI：10.1109/TC.2025.3625156delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
脉冲神经网络（SNNs）被广泛应用于类脑计算和神经科学研究中。已构建了几种多核加速器以提高SNNs的运行速度和能效。然而，当前加速器通常需要在SNNs的每个时间步之后在各核心之间进行显式同步，这给整体效率带来了挑战。本文提出DepAsync，一种异步架构，通过消除核心间同步，实现了快速且能效高的SNN推理，并具有出色的可扩展性。其核心思想是利用在编译时预确定的神经形态核心之间的依赖关系。我们为每个核心设计了DepAsync调度器，以跟踪其依赖项的运行状态，并控制核心在不等待其他核心完成任务的情况下安全地进入下一个时间步。这种方法避免了全局同步的必要性，使DepAsync能够最小化核心等待时间，以应对SNN工作负载中固有的核心和时间不平衡问题。使用五种SNN工作负载的全面评估表明，与最先进的同步架构相比，DepAsync实现了2.47倍的速度提升和1.55倍的能效提升。
Keyword:
Accelerators
spiking neural networks
neuromorphic computing

期刊

IEEE Transactions on Computers 封面图
IEEE Transactions on Computers
IF:
3.8
论文数:
5.3K
被引数:
9.8K

机构

Z
zhejiang university
学者数:
17.7W
论文数: 12.1W
被引数: 152
引用论文

引用论文

A Programmable Event-driven Architecture for Evaluating Spiking Neural Networks
err2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED)
IF0
err2017-07-01
err0
PREAI
errArnab Roy; Swagath Venkataramani; Neel Gala; Sanchari Sen; Kamakoti Veezhinathan; Anand Raghunathan
err分享
err收藏
Predictive Technology Model for Robust Nanoelectronic Design
errIntegrated Circuits and Systems
IF0
err2011-01-01
err0
PREAI
errYu Cao
err分享
err收藏
NeuroSim: A Circuit-Level Macro Model for Benchmarking Neuro-Inspired Architectures in Online LearningNeuroSim: 在线学习中神经启发架构基准测试的电路级宏模型
errIEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
IF0
err2018-12-01
err0
errOAAI
errPai-Yu Chen; Xiaochen Peng; Shimeng Yu
err分享
err收藏
Converting Static Image Datasets to Spiking Neuromorphic Datasets Using Saccades使用扫视将静态图像数据集转换为尖峰神经形态数据集
errFrontiers in Neuroscience
IF0
err2015-11-16
err0
errOAAI
errGarrick Orchard; Ajinkya Jayawant; Gregory K. Cohen; Nitish Thakor
err分享
err收藏
Spiking Deep Residual Networks尖峰深度残差网络
errIEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
IF8.9
err2023-08-01
err77
PREAI
errHu, Yangfan; Tang, Huajin; Pan, Gang
err分享
err收藏
SPAIC: A Spike-Based Artificial Intelligence Computing Framework
errIEEE COMPUTATIONAL INTELLIGENCE MAGAZINE
IF11.2
err2024-02-01
err5
errOAAI
errHong, Chaofei; Yuan, Mengwen; Zhang, Mengxiao; Wang, Xiao; Zhang, Chengjun; Wang, Jiaxin; Pan, Gang; Tang, Huajin
err分享
err收藏
Open-loop analog programmable electrochemical memory array开环模拟可编程电化学存储器阵列
errNATURE COMMUNICATIONS
IF15.7
err2023-10-04
err18
errOAAI
errChen, Peng; Liu, Fenghao; Lin, Peng; Li, Peihong; Xiao, Yu; Zhang, Bihua; Pan, Gang
err分享
err收藏
A Low Power, Fully Event-Based Gesture Recognition System低功耗，完全基于事件的手势识别系统
err2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
IF0
err2017-07-01
err0
PREAI
errArnon Amir; Brian Taba; David Berg; Timothy Melano; Jeffrey McKinstry; Carmelo Di Nolfo; Tapan Nayak; Alexander Andreopoulos; Guillaume Garreau; Marcela Mendoza; Jeff Kusnitz; Michael Debole; Steve Esser; Tobi Delbruck; Myron Flickner; Dharmendra Modha
err分享
err收藏
SpinalFlow: An Architecture and Dataflow Tailored for Spiking Neural Networks
err2020 ACM/IEEE 47th Annual International Symposium on Computer Architecture (ISCA)
IF0
err2020-05-01
err0
PREAI
errSurya Narayanan; Karl Taht; Rajeev Balasubramonian; Edouard Giacomin; Pierre-Emmanuel Gaillardon
err分享
err收藏
Mapping Large-Scale Spiking Neural Network on Arbitrary Meshed Neuromorphic Hardware在大规模脉冲神经网络与任意网格神经形态硬件之间的映射
errIEEE Transactions on Parallel and Distributed Systems
IF6
err
err0
PREAI
errOuwen Jin; Qinghui Xing; Zhuo Chen; Ming Zhang; De Ma; Ying Li; Xin Du; Shuibing He; Shuiguang Deng; Gang Pan
err分享
err收藏
学者 查看更多内容