A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference

Jinling Xu; Yonggui Wang; Wenbiao Zhouy

doi:10.1109/ISCAS48785.2022.9937592

A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference

Jinling Xu, Yonggui Wang, Wenbiao Zhouy^*

^*Corresponding author for this work

School of Integrated Circuits and Electronics

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

The neural networks on 3D data and applications have emerged in the past five years. However, there are only a few dedicated hardware designs were proposed for 3D data and algorithms. Meanwhile, they lack flexibility and adaptation for the fast evolvement of software algorithms. We propose a heterogeneous accelerator design on Xilinx Zynq and Zynq UltraScale+ platform. An innovative vector pipeline is designed in the accelerator that can reach the near limitation of BRAM frequency, and it gives the final design frequency closure at 550MHz with 100% DSP usage.

Original language	English
Title of host publication	IEEE International Symposium on Circuits and Systems, ISCAS 2022
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	2725-2729
Number of pages	5
ISBN (Electronic)	9781665484855
DOIs	https://doi.org/10.1109/ISCAS48785.2022.9937592
Publication status	Published - 2022
Event	2022 IEEE International Symposium on Circuits and Systems, ISCAS 2022 - Austin, United States Duration: 27 May 2022 → 1 Jun 2022

Publication series

Name	Proceedings - IEEE International Symposium on Circuits and Systems
Volume	2022-May
ISSN (Print)	0271-4310

Conference

Conference	2022 IEEE International Symposium on Circuits and Systems, ISCAS 2022
Country/Territory	United States
City	Austin
Period	27/05/22 → 1/06/22

Access to Document

10.1109/ISCAS48785.2022.9937592

Cite this

Xu, J., Wang, Y., & Zhouy, W. (2022). A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference. In IEEE International Symposium on Circuits and Systems, ISCAS 2022 (pp. 2725-2729). (Proceedings - IEEE International Symposium on Circuits and Systems; Vol. 2022-May). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ISCAS48785.2022.9937592

Xu, Jinling ; Wang, Yonggui ; Zhouy, Wenbiao. / A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference. IEEE International Symposium on Circuits and Systems, ISCAS 2022. Institute of Electrical and Electronics Engineers Inc., 2022. pp. 2725-2729 (Proceedings - IEEE International Symposium on Circuits and Systems).

@inproceedings{77dafd1c70234e6396aebc2ac5b2a098,

title = "A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference",

abstract = "The neural networks on 3D data and applications have emerged in the past five years. However, there are only a few dedicated hardware designs were proposed for 3D data and algorithms. Meanwhile, they lack flexibility and adaptation for the fast evolvement of software algorithms. We propose a heterogeneous accelerator design on Xilinx Zynq and Zynq UltraScale+ platform. An innovative vector pipeline is designed in the accelerator that can reach the near limitation of BRAM frequency, and it gives the final design frequency closure at 550MHz with 100% DSP usage.",

author = "Jinling Xu and Yonggui Wang and Wenbiao Zhouy",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 2022 IEEE International Symposium on Circuits and Systems, ISCAS 2022 ; Conference date: 27-05-2022 Through 01-06-2022",

year = "2022",

doi = "10.1109/ISCAS48785.2022.9937592",

language = "English",

series = "Proceedings - IEEE International Symposium on Circuits and Systems",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "2725--2729",

booktitle = "IEEE International Symposium on Circuits and Systems, ISCAS 2022",

address = "United States",

}

Xu, J, Wang, Y & Zhouy, W 2022, A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference. in IEEE International Symposium on Circuits and Systems, ISCAS 2022. Proceedings - IEEE International Symposium on Circuits and Systems, vol. 2022-May, Institute of Electrical and Electronics Engineers Inc., pp. 2725-2729, 2022 IEEE International Symposium on Circuits and Systems, ISCAS 2022, Austin, United States, 27/05/22. https://doi.org/10.1109/ISCAS48785.2022.9937592

A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference. / Xu, Jinling; Wang, Yonggui; Zhouy, Wenbiao.
IEEE International Symposium on Circuits and Systems, ISCAS 2022. Institute of Electrical and Electronics Engineers Inc., 2022. p. 2725-2729 (Proceedings - IEEE International Symposium on Circuits and Systems; Vol. 2022-May).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference

AU - Xu, Jinling

AU - Wang, Yonggui

AU - Zhouy, Wenbiao

PY - 2022

Y1 - 2022

N2 - The neural networks on 3D data and applications have emerged in the past five years. However, there are only a few dedicated hardware designs were proposed for 3D data and algorithms. Meanwhile, they lack flexibility and adaptation for the fast evolvement of software algorithms. We propose a heterogeneous accelerator design on Xilinx Zynq and Zynq UltraScale+ platform. An innovative vector pipeline is designed in the accelerator that can reach the near limitation of BRAM frequency, and it gives the final design frequency closure at 550MHz with 100% DSP usage.

AB - The neural networks on 3D data and applications have emerged in the past five years. However, there are only a few dedicated hardware designs were proposed for 3D data and algorithms. Meanwhile, they lack flexibility and adaptation for the fast evolvement of software algorithms. We propose a heterogeneous accelerator design on Xilinx Zynq and Zynq UltraScale+ platform. An innovative vector pipeline is designed in the accelerator that can reach the near limitation of BRAM frequency, and it gives the final design frequency closure at 550MHz with 100% DSP usage.

UR - http://www.scopus.com/inward/record.url?scp=85142495218&partnerID=8YFLogxK

U2 - 10.1109/ISCAS48785.2022.9937592

DO - 10.1109/ISCAS48785.2022.9937592

M3 - Conference contribution

AN - SCOPUS:85142495218

T3 - Proceedings - IEEE International Symposium on Circuits and Systems

SP - 2725

EP - 2729

BT - IEEE International Symposium on Circuits and Systems, ISCAS 2022

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2022 IEEE International Symposium on Circuits and Systems, ISCAS 2022

Y2 - 27 May 2022 through 1 June 2022

ER -

Xu J, Wang Y, Zhouy W. A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference. In IEEE International Symposium on Circuits and Systems, ISCAS 2022. Institute of Electrical and Electronics Engineers Inc. 2022. p. 2725-2729. (Proceedings - IEEE International Symposium on Circuits and Systems). doi: 10.1109/ISCAS48785.2022.9937592

A Heterogeneous FPGA-based Accelerator Design for Efficient and Low-cost Point Clouds Deep Learning Inference

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this