General Reuse-Centric CNN Accelerator

被引：6

作者：

Cicek, Nihat Mert ^{[1
,2
]}

Ning, Lin ^{[3
]}

Ozturk, Ozcan ^{[4
]}

Shen, Xipeng ^{[3
]}

机构：

[1] Aselsan Inc, TR-06200 Yenimahalle Ankara, Turkey

[2] Bilkent Univ, Dept Comp Engn, TR-06800 Ankara, Turkey

[3] North Carolina State Univ, Dept Comp Sci, Raleigh, NC 27695 USA

[4] Bilkent Univ, Dept Comp Engn, TR-06800 Ankara, Turkey

来源：

IEEE TRANSACTIONS ON COMPUTERS | 2022年 / 71卷 / 04期

基金：

美国国家科学基金会;

关键词：

Neurons; Hardware; Convolution; Engines; Software; Acceleration; IEEE Senior Members; CNN; reuse-centric; accelerator;

D O I：

10.1109/TC.2021.3064608

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

This article introduces the first general reuse-centric accelerator for CNN inferences. Unlike prior work that exploits similarities only across consecutive video frames, general reuse-centric accelerator is able to discover similarities among arbitrary patches within an image or across independent images, and translate them into computation time and energy savings. Experiments show that the accelerator complements both prior software-based CNN and various CNN hardware accelerators, producing up to 14.96X speedups for similarity discovery, up to 2.70X speedups for overall inference.

引用

页码：880 / 891

页数：12

共 50 条

[31] Configurable CNN Accelerator Based on Tiling Dataflow
Li, Yihuang
Ma, Sheng
Guo, Yang
Xu, Rui
Chen, Guilin
PROCEEDINGS OF 2018 IEEE 9TH INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING AND SERVICE SCIENCE (ICSESS), 2018, : 309 - 313
[32] A Depthwise Separable Convolution Architecture for CNN Accelerator
Srivastava, Harsh
Sarawadekar, Kishor
PROCEEDINGS OF 2020 IEEE APPLIED SIGNAL PROCESSING CONFERENCE (ASPCON 2020), 2020, : 1 - 5
[33] Hardware-Software Codesign of a CNN Accelerator
Yi, Changjae
Kang, Donghyun
Ha, Soonhoi
2022 25TH EUROMICRO CONFERENCE ON DIGITAL SYSTEM DESIGN (DSD), 2022, : 348 - 356
[34] CaFPGA: An automatic generation model for CNN accelerator
Xu, Jinwei
Liu, Zhiqiang
Jiang, Jingfei
Dou, Yong
Li, Shijie
MICROPROCESSORS AND MICROSYSTEMS, 2018, 60 : 196 - 206
[35] The Sparsity and Activation Analysis of Compressed CNN Networks in a HW CNN Accelerator Model
Lee, Mi-Young
Lee, Joo-Hyun
Kim, Jin-Kyu
Kim, Byung-Jo
Kim, Ju-Yeob
2019 INTERNATIONAL SOC DESIGN CONFERENCE (ISOCC), 2019, : 255 - 256
[36] Architecture-centric software process for software reuse
Department of Computer Science and Technology, Xi'an Jiaotong University, Xi'an 710049, China
High Technol Letters, 2006, SUPPL. (85-89):
[37] ADS-CNN: Adaptive Dataflow Scheduling for lightweight CNN accelerator on FPGAs
Wan, Yi
Xie, Xianzhong
Chen, Junfan
Xie, Kunpeng
Yi, Dezhi
Lu, Ye
Gai, Keke
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE, 2024, 158 : 138 - 149
[38] Methods and Infrastructure in the Era of Accelerator-Centric Architectures
Reagen, Brandon
Shao, Yakun Sophia
Xi, Sam
Wei, Gu-Yeon
Brooks, David
2017 IEEE 60TH INTERNATIONAL MIDWEST SYMPOSIUM ON CIRCUITS AND SYSTEMS (MWSCAS), 2017, : 902 - 905
[39] FLIP: Data-centric Edge CGRA Accelerator
Wu, Dan
Chen, Peng
Bandara, Thilini Kaushalya
Li, Zhaoying
Mitra, Tulika
ACM TRANSACTIONS ON DESIGN AUTOMATION OF ELECTRONIC SYSTEMS, 2024, 29 (01)
[40] Supporting Address Translation for Accelerator-Centric Architectures
Cong, Jason
Fang, Zhenman
Hao, Yuchen
Reinman, Glenn
2017 23RD IEEE INTERNATIONAL SYMPOSIUM ON HIGH PERFORMANCE COMPUTER ARCHITECTURE (HPCA), 2017, : 37 - 48

← 1 2 3 4 5 →