Argus: An End-to-End Framework for Accelerating CNNs on FPGAs

Yongming Shen; Tianchu Ji; Michael Ferdman; Peter Milder

doi:10.1109/MM.2019.2930607

Argus: An End-to-End Framework for Accelerating CNNs on FPGAs

Yongming Shen
, Tianchu Ji
, Michael Ferdman
, Peter Milder

Stony Brook University

Research output: Contribution to journal › Article › peer-review

12 Scopus citations

Abstract

In this article, we present Argus, an end-to-end framework for accelerating convolutional neural networks (CNNs) on field-programmable gate arrays (FPGAs) with minimum user effort. Argus uses state-of-the-art methods to auto-generate highly optimized CNN accelerator designs for FPGAs, and includes software for running an FPGA-backed CNN inference microservice.

Original language	English
Article number	8769906
Pages (from-to)	17-25
Number of pages	9
Journal	IEEE Micro
Volume	39
Issue number	5
DOIs	https://doi.org/10.1109/MM.2019.2930607
State	Published - Sep 1 2019

Keywords

Convolutional Neural Network
FPGA
Hardware Accelerator

Access to Document

10.1109/MM.2019.2930607

Cite this

@article{a07e046369884c238d85d96758db07f0,

title = "Argus: An End-to-End Framework for Accelerating CNNs on FPGAs",

abstract = "In this article, we present Argus, an end-to-end framework for accelerating convolutional neural networks (CNNs) on field-programmable gate arrays (FPGAs) with minimum user effort. Argus uses state-of-the-art methods to auto-generate highly optimized CNN accelerator designs for FPGAs, and includes software for running an FPGA-backed CNN inference microservice.",

keywords = "Convolutional Neural Network, FPGA, Hardware Accelerator",

author = "Yongming Shen and Tianchu Ji and Michael Ferdman and Peter Milder",

note = "Publisher Copyright: {\textcopyright} 1981-2012 IEEE.",

year = "2019",

month = sep,

day = "1",

doi = "10.1109/MM.2019.2930607",

language = "English",

volume = "39",

pages = "17--25",

journal = "IEEE Micro",

issn = "0272-1732",

number = "5",

}

TY - JOUR

T1 - Argus

T2 - An End-to-End Framework for Accelerating CNNs on FPGAs

AU - Shen, Yongming

AU - Ji, Tianchu

AU - Ferdman, Michael

AU - Milder, Peter

PY - 2019/9/1

Y1 - 2019/9/1

N2 - In this article, we present Argus, an end-to-end framework for accelerating convolutional neural networks (CNNs) on field-programmable gate arrays (FPGAs) with minimum user effort. Argus uses state-of-the-art methods to auto-generate highly optimized CNN accelerator designs for FPGAs, and includes software for running an FPGA-backed CNN inference microservice.

AB - In this article, we present Argus, an end-to-end framework for accelerating convolutional neural networks (CNNs) on field-programmable gate arrays (FPGAs) with minimum user effort. Argus uses state-of-the-art methods to auto-generate highly optimized CNN accelerator designs for FPGAs, and includes software for running an FPGA-backed CNN inference microservice.

KW - Convolutional Neural Network

KW - FPGA

KW - Hardware Accelerator

UR - https://www.scopus.com/pages/publications/85072162303

U2 - 10.1109/MM.2019.2930607

DO - 10.1109/MM.2019.2930607

M3 - Article

AN - SCOPUS:85072162303

SN - 0272-1732

VL - 39

SP - 17

EP - 25

JO - IEEE Micro

JF - IEEE Micro

IS - 5

M1 - 8769906

ER -

Argus: An End-to-End Framework for Accelerating CNNs on FPGAs

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this