Timing driven partition for multi-FPGA systems with TDM awareness

Sin Hong Liou; Sean Liu; Richard Sun; Hung Ming Chen

doi:10.1145/3372780.3375558

Timing driven partition for multi-FPGA systems with TDM awareness

Sin Hong Liou, Sean Liu, Richard Sun, Hung Ming Chen

Institute of Electronics

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

Abstract

Multi-FPGA system is a popular approach to achieve hardware acceleration with the scalability to accommodate large designs. To overcome the connectivity constraint between each pair of FPGAs, Time-division multiplexing (TDM) is adopted with the expense of additional delay that dominates the performance on multi-FPGA system based emulator. To the best of our knowledge, there is no prior work on partitioning for multi-FPGA system considering hardware configuration and the impact of TDM. This work proposes a partition methodology to improve timing performance for multi-FPGA system. Delay introduced by TDM is estimated and optimized using look-up table for better efficiency. Our experimental result shows 43% improvement in maximum delay while considering both hardware configuration and impact of TDM compared with cut driven partition approach.

Original language	English
Title of host publication	ISPD 2020 - Proceedings of the 2020 International Symposium on Physical Design
Publisher	Association for Computing Machinery
Pages	111-118
Number of pages	8
ISBN (Electronic)	9781450370912
DOIs	https://doi.org/10.1145/3372780.3375558
State	Published - 20 Sep 2020
Event	23rd International Symposium on Physical Design, ISPD 2020 - Taipei, Taiwan Duration: 29 Mar 2020 → 1 Apr 2020

Publication series

Name	Proceedings of the International Symposium on Physical Design

Conference

Conference	23rd International Symposium on Physical Design, ISPD 2020
Country	Taiwan
City	Taipei
Period	29/03/20 → 1/04/20

Keywords

Field programmable gate array
Partitioning
Physical design
Time-division-multiplexing

Access to Document

10.1145/3372780.3375558

Link to publication in Scopus

Fingerprint Dive into the research topics of 'Timing driven partition for multi-FPGA systems with TDM awareness'. Together they form a unique fingerprint.

Cite this

Liou, S. H., Liu, S., Sun, R., & Chen, H. M. (2020). Timing driven partition for multi-FPGA systems with TDM awareness. In ISPD 2020 - Proceedings of the 2020 International Symposium on Physical Design (pp. 111-118). (Proceedings of the International Symposium on Physical Design). Association for Computing Machinery. https://doi.org/10.1145/3372780.3375558

@inproceedings{e466c7b40c264c80b55f0f3e658b4565,

title = "Timing driven partition for multi-FPGA systems with TDM awareness",

abstract = "Multi-FPGA system is a popular approach to achieve hardware acceleration with the scalability to accommodate large designs. To overcome the connectivity constraint between each pair of FPGAs, Time-division multiplexing (TDM) is adopted with the expense of additional delay that dominates the performance on multi-FPGA system based emulator. To the best of our knowledge, there is no prior work on partitioning for multi-FPGA system considering hardware configuration and the impact of TDM. This work proposes a partition methodology to improve timing performance for multi-FPGA system. Delay introduced by TDM is estimated and optimized using look-up table for better efficiency. Our experimental result shows 43% improvement in maximum delay while considering both hardware configuration and impact of TDM compared with cut driven partition approach.",

keywords = "Field programmable gate array, Partitioning, Physical design, Time-division-multiplexing",

author = "Liou, {Sin Hong} and Sean Liu and Richard Sun and Chen, {Hung Ming}",

year = "2020",

month = sep,

day = "20",

doi = "10.1145/3372780.3375558",

language = "English",

series = "Proceedings of the International Symposium on Physical Design",

publisher = "Association for Computing Machinery",

pages = "111--118",

booktitle = "ISPD 2020 - Proceedings of the 2020 International Symposium on Physical Design",

note = "null ; Conference date: 29-03-2020 Through 01-04-2020",

}

Liou, SH, Liu, S, Sun, R & Chen, HM 2020, Timing driven partition for multi-FPGA systems with TDM awareness. in ISPD 2020 - Proceedings of the 2020 International Symposium on Physical Design. Proceedings of the International Symposium on Physical Design, Association for Computing Machinery, pp. 111-118, 23rd International Symposium on Physical Design, ISPD 2020, Taipei, Taiwan, 29/03/20. https://doi.org/10.1145/3372780.3375558

Timing driven partition for multi-FPGA systems with TDM awareness. / Liou, Sin Hong; Liu, Sean; Sun, Richard; Chen, Hung Ming.

ISPD 2020 - Proceedings of the 2020 International Symposium on Physical Design. Association for Computing Machinery, 2020. p. 111-118 (Proceedings of the International Symposium on Physical Design).

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

TY - GEN

T1 - Timing driven partition for multi-FPGA systems with TDM awareness

AU - Liou, Sin Hong

AU - Liu, Sean

AU - Sun, Richard

AU - Chen, Hung Ming

PY - 2020/9/20

Y1 - 2020/9/20

N2 - Multi-FPGA system is a popular approach to achieve hardware acceleration with the scalability to accommodate large designs. To overcome the connectivity constraint between each pair of FPGAs, Time-division multiplexing (TDM) is adopted with the expense of additional delay that dominates the performance on multi-FPGA system based emulator. To the best of our knowledge, there is no prior work on partitioning for multi-FPGA system considering hardware configuration and the impact of TDM. This work proposes a partition methodology to improve timing performance for multi-FPGA system. Delay introduced by TDM is estimated and optimized using look-up table for better efficiency. Our experimental result shows 43% improvement in maximum delay while considering both hardware configuration and impact of TDM compared with cut driven partition approach.

AB - Multi-FPGA system is a popular approach to achieve hardware acceleration with the scalability to accommodate large designs. To overcome the connectivity constraint between each pair of FPGAs, Time-division multiplexing (TDM) is adopted with the expense of additional delay that dominates the performance on multi-FPGA system based emulator. To the best of our knowledge, there is no prior work on partitioning for multi-FPGA system considering hardware configuration and the impact of TDM. This work proposes a partition methodology to improve timing performance for multi-FPGA system. Delay introduced by TDM is estimated and optimized using look-up table for better efficiency. Our experimental result shows 43% improvement in maximum delay while considering both hardware configuration and impact of TDM compared with cut driven partition approach.

KW - Field programmable gate array

KW - Partitioning

KW - Physical design

KW - Time-division-multiplexing

UR - http://www.scopus.com/inward/record.url?scp=85082557378&partnerID=8YFLogxK

U2 - 10.1145/3372780.3375558

DO - 10.1145/3372780.3375558

M3 - Conference contribution

AN - SCOPUS:85082557378

T3 - Proceedings of the International Symposium on Physical Design

SP - 111

EP - 118

BT - ISPD 2020 - Proceedings of the 2020 International Symposium on Physical Design

PB - Association for Computing Machinery

Y2 - 29 March 2020 through 1 April 2020

ER -