Towards an optimized multi FPGA architecture with STDM network: A preliminary study

Kazuei Hironaka, Ng Anh Vu Doan, Hideharu Amano

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this work, we propose a multi FPGA architecture with STDM network that aims to tackle compute-intensive applications such as neural networks training or pattern recognition in artificial intelligence while realizing high cost-performance and energy efficiency. To achieve this goal, optimizing different aspects of the system communication is a key challenge. In order to do this, a preliminary study on the application mapping for both the execution time and the number of slots for the STDM is carried out. An optimization based on a multi-criteria paradigm is implemented and the preliminary results show the possibility to optimize several parameters of the communication simultaneously alongside quantitative analyses of different architecture choices.

Original languageEnglish
Title of host publicationApplied Reconfigurable Computing
Subtitle of host publicationArchitectures, Tools, and Applications - 14th International Symposium, ARC 2018, Proceedings
PublisherSpringer Verlag
Pages142-150
Number of pages9
ISBN (Print)9783319788890
DOIs
Publication statusPublished - 2018 Jan 1
Event14th International Symposium on Applied Reconfigurable Computing, ARC 2018 - Santorini, Greece
Duration: 2018 May 22018 May 4

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10824 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other14th International Symposium on Applied Reconfigurable Computing, ARC 2018
CountryGreece
CitySantorini
Period18/5/218/5/4

Fingerprint

Field Programmable Gate Array
Field programmable gate arrays (FPGA)
Communication
Multi-criteria
Energy Efficiency
Execution Time
Pattern Recognition
Pattern recognition
Artificial intelligence
Communication Systems
Energy efficiency
Artificial Intelligence
Paradigm
Optimise
Neural Networks
Neural networks
Optimization
Costs
Architecture
Training

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Hironaka, K., Doan, N. A. V., & Amano, H. (2018). Towards an optimized multi FPGA architecture with STDM network: A preliminary study. In Applied Reconfigurable Computing: Architectures, Tools, and Applications - 14th International Symposium, ARC 2018, Proceedings (pp. 142-150). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10824 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-78890-6_12

Towards an optimized multi FPGA architecture with STDM network : A preliminary study. / Hironaka, Kazuei; Doan, Ng Anh Vu; Amano, Hideharu.

Applied Reconfigurable Computing: Architectures, Tools, and Applications - 14th International Symposium, ARC 2018, Proceedings. Springer Verlag, 2018. p. 142-150 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10824 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hironaka, K, Doan, NAV & Amano, H 2018, Towards an optimized multi FPGA architecture with STDM network: A preliminary study. in Applied Reconfigurable Computing: Architectures, Tools, and Applications - 14th International Symposium, ARC 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10824 LNCS, Springer Verlag, pp. 142-150, 14th International Symposium on Applied Reconfigurable Computing, ARC 2018, Santorini, Greece, 18/5/2. https://doi.org/10.1007/978-3-319-78890-6_12
Hironaka K, Doan NAV, Amano H. Towards an optimized multi FPGA architecture with STDM network: A preliminary study. In Applied Reconfigurable Computing: Architectures, Tools, and Applications - 14th International Symposium, ARC 2018, Proceedings. Springer Verlag. 2018. p. 142-150. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-78890-6_12
Hironaka, Kazuei ; Doan, Ng Anh Vu ; Amano, Hideharu. / Towards an optimized multi FPGA architecture with STDM network : A preliminary study. Applied Reconfigurable Computing: Architectures, Tools, and Applications - 14th International Symposium, ARC 2018, Proceedings. Springer Verlag, 2018. pp. 142-150 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{71453570a4b04edcbf5b82fd81d01467,
title = "Towards an optimized multi FPGA architecture with STDM network: A preliminary study",
abstract = "In this work, we propose a multi FPGA architecture with STDM network that aims to tackle compute-intensive applications such as neural networks training or pattern recognition in artificial intelligence while realizing high cost-performance and energy efficiency. To achieve this goal, optimizing different aspects of the system communication is a key challenge. In order to do this, a preliminary study on the application mapping for both the execution time and the number of slots for the STDM is carried out. An optimization based on a multi-criteria paradigm is implemented and the preliminary results show the possibility to optimize several parameters of the communication simultaneously alongside quantitative analyses of different architecture choices.",
author = "Kazuei Hironaka and Doan, {Ng Anh Vu} and Hideharu Amano",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-78890-6_12",
language = "English",
isbn = "9783319788890",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "142--150",
booktitle = "Applied Reconfigurable Computing",
address = "Germany",

}

TY - GEN

T1 - Towards an optimized multi FPGA architecture with STDM network

T2 - A preliminary study

AU - Hironaka, Kazuei

AU - Doan, Ng Anh Vu

AU - Amano, Hideharu

PY - 2018/1/1

Y1 - 2018/1/1

N2 - In this work, we propose a multi FPGA architecture with STDM network that aims to tackle compute-intensive applications such as neural networks training or pattern recognition in artificial intelligence while realizing high cost-performance and energy efficiency. To achieve this goal, optimizing different aspects of the system communication is a key challenge. In order to do this, a preliminary study on the application mapping for both the execution time and the number of slots for the STDM is carried out. An optimization based on a multi-criteria paradigm is implemented and the preliminary results show the possibility to optimize several parameters of the communication simultaneously alongside quantitative analyses of different architecture choices.

AB - In this work, we propose a multi FPGA architecture with STDM network that aims to tackle compute-intensive applications such as neural networks training or pattern recognition in artificial intelligence while realizing high cost-performance and energy efficiency. To achieve this goal, optimizing different aspects of the system communication is a key challenge. In order to do this, a preliminary study on the application mapping for both the execution time and the number of slots for the STDM is carried out. An optimization based on a multi-criteria paradigm is implemented and the preliminary results show the possibility to optimize several parameters of the communication simultaneously alongside quantitative analyses of different architecture choices.

UR - http://www.scopus.com/inward/record.url?scp=85046261161&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046261161&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-78890-6_12

DO - 10.1007/978-3-319-78890-6_12

M3 - Conference contribution

AN - SCOPUS:85046261161

SN - 9783319788890

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 142

EP - 150

BT - Applied Reconfigurable Computing

PB - Springer Verlag

ER -