Exploiting Variable Precision Computation Array for Scalable Neural Network Accelerators

被引：0
|
作者：
Yang, Shaofei [1 ]
Liu, Longjun [1 ]
Li, Baoting [1 ]
Sun, Hongbin [1 ]
Zheng, Nanning [1 ]
机构：
[1] Xi An Jiao Tong Univ, Inst Artificial Intelligence & Robot, Xian 710049, Peoples R China
来源：
2020 2ND IEEE INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE CIRCUITS AND SYSTEMS (AICAS 2020) | 2020年
基金：
中国国家自然科学基金;
关键词：
Deep Neural Networks; Accelerator; Energy Efficiency Computing Array; Dynamic Quantization; Resiliency;
D O I：
暂无
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
In this paper, we present a flexible Variable Precision Computation Array (VPCA) component for different accelerators, which leverages a sparsification scheme for activations and a low bits serial-parallel combination computation unit for improving the efficiency and resiliency of accelerators. The VPCA can dynamically decompose the width of activation/weights (from 32bit to 3bit in different accelerators) into 2-bits serial computation units while the 2bits computing units can be combined in parallel computing for high throughput. We propose an on-the-fly compressing and calculating strategy SLE-CLC (single lane encoding, cross lane calculation), which could further improve performance of 2-bit parallel computing. The experiments results on image classification datasets show VPCA can outperforms DaDianNao, Stripes, Loom-2bit by 4.67x, 2.42x, 1.52x without other overhead on convolution layers.
引用
收藏
页码：315 / 319
页数：5
相关论文
共 50 条
  • [21] A TACTILE SENSOR ARRAY WITH A MONOLITHICALLY INTEGRATED NEURAL NETWORK FOR EDGE COMPUTATION
    Lei, Tengteng
    Hu, Yushen
    Wong, Man
    2023 IEEE 36TH INTERNATIONAL CONFERENCE ON MICRO ELECTRO MECHANICAL SYSTEMS, MEMS, 2023, : 13 - 16
  • [22] Fault-Tolerant Systolic Array Based Accelerators for Deep Neural Network Execution
    Zhang, Jeff
    Basu, Kanad
    Garg, Siddharth
    IEEE DESIGN & TEST, 2019, 36 (05) : 44 - 53
  • [23] A Precision-Scalable Energy-Efficient Convolutional Neural Network Accelerator
    Liu, Wenjian
    Lin, Jun
    Wang, Zhongfeng
    IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS, 2020, 67 (10) : 3484 - 3497
  • [24] A Precision-Scalable Deep Neural Network Accelerator With Activation Sparsity Exploitation
    Li, Wenjie
    Hu, Aokun
    Xu, Ningyi
    He, Guanghui
    IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS, 2024, 43 (01) : 263 - 276
  • [25] Array Aware Training/Pruning: Methods for Efficient Forward Propagation on Array-based Neural Network Accelerators
    Chitty-Venkata, Krishna Teja
    Somani, Arun K.
    2020 IEEE 31ST INTERNATIONAL CONFERENCE ON APPLICATION-SPECIFIC SYSTEMS, ARCHITECTURES AND PROCESSORS (ASAP 2020), 2020, : 37 - 44
  • [26] Scalable Computation of Isochrones with Network Expiration
    Gamper, Johann
    Bohlen, Michael
    Innerebner, Markus
    SCIENTIFIC AND STATISTICAL DATABASE MANAGEMENT, SSDBM 2012, 2012, 7338 : 526 - 543
  • [27] Research on High-Precision Stochastic Computing VLSI Structures for Deep Neural Network Accelerators
    WU Jingguo
    ZHU Jingwei
    XIONG Xiankui
    YAO Haidong
    WANG Chengchen
    CHEN Yun
    ZTE Communications, 2024, 22 (04) : 9 - 17
  • [28] Scalable Communication Architecture for Network-Attached Accelerators
    Neuwirth, Sarah
    Frey, Dirk
    Nuessle, Mondrian
    Bruening, Ulrich
    2015 IEEE 21ST INTERNATIONAL SYMPOSIUM ON HIGH PERFORMANCE COMPUTER ARCHITECTURE (HPCA), 2015, : 627 - 638
  • [29] Scalable end-to-end recurrent neural network for variable star classification
    Becker, I
    Pichara, K.
    Catelan, M.
    Protopapas, P.
    Aguirre, C.
    Nikzat, F.
    MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY, 2020, 493 (02) : 2981 - 2995
  • [30] Compiling Optimization for Neural Network Accelerators
    Song, Jin
    Zhuang, Yimin
    Chen, Xiaobing
    Zhi, Tian
    Liu, Shaoli
    ADVANCED PARALLEL PROCESSING TECHNOLOGIES (APPT 2019), 2019, 11719 : 15 - 26
12345