Variable pipeline structure for coarse grained reconfigurable array CMA

Naoki Ando; Koichiro Masuyama; Hayate Okuhara; Hideharu Amano

doi:10.1109/FPT.2016.7929537

Variable pipeline structure for coarse grained reconfigurable array CMA

Naoki Ando, Koichiro Masuyama, Hayate Okuhara, Hideharu Amano

情報工学科

研究成果: Conference contribution

3 被引用数 (Scopus)

抄録

Cool mega-array (CMA) is a kind of coarse grained reconfigurable architecture (CGRA) which has shown its ability of ultra low-power computation. However, as CMA completely eliminates clock trees and registers, the performance improvement has been limited. In this paper, we introduce a variable pipeline structure to CMA with the minimum essential registers to provide more wide trade-off between performance and energy. Comparing with the baseline CMA (non-pipelined structure), an average of 77% improvement for performance was achieved with a small power overhead. Moreover, the energy efficiency was 1461 MOPS / mW at most which was about 2× that of the baseline structure. The best pipeline depth for an arbitrary energy-performance trade-off became selectable with only 11% area overhead.

本文言語	English
ホスト出版物のタイトル	Proceedings of the 2016 International Conference on Field-Programmable Technology, FPT 2016
出版社	Institute of Electrical and Electronics Engineers Inc.
ページ	217-220
ページ数	4
ISBN（電子版）	9781509056026
DOI	https://doi.org/10.1109/FPT.2016.7929537
出版ステータス	Published - 2017 5 15
イベント	15th International Conference on Field-Programmable Technology, FPT 2016 - Xi'an, China 継続期間: 2016 12 7 → 2016 12 9

Other

Other	15th International Conference on Field-Programmable Technology, FPT 2016
Country	China
City	Xi'an
Period	16/12/7 → 16/12/9

ASJC Scopus subject areas

Computer Networks and Communications
Electrical and Electronic Engineering
Safety, Risk, Reliability and Quality
Instrumentation
Hardware and Architecture

文献へのアクセス

10.1109/FPT.2016.7929537

フィンガープリント「Variable pipeline structure for coarse grained reconfigurable array CMA」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル

Ando, N, Masuyama, K, Okuhara, H & Amano, H 2017, Variable pipeline structure for coarse grained reconfigurable array CMA. in Proceedings of the 2016 International Conference on Field-Programmable Technology, FPT 2016., 7929537, Institute of Electrical and Electronics Engineers Inc., pp. 217-220, 15th International Conference on Field-Programmable Technology, FPT 2016, Xi'an, China, 16/12/7. https://doi.org/10.1109/FPT.2016.7929537

@inproceedings{6827577f3aa340a5bcc66b77355085cc,

title = "Variable pipeline structure for coarse grained reconfigurable array CMA",

abstract = "Cool mega-array (CMA) is a kind of coarse grained reconfigurable architecture (CGRA) which has shown its ability of ultra low-power computation. However, as CMA completely eliminates clock trees and registers, the performance improvement has been limited. In this paper, we introduce a variable pipeline structure to CMA with the minimum essential registers to provide more wide trade-off between performance and energy. Comparing with the baseline CMA (non-pipelined structure), an average of 77% improvement for performance was achieved with a small power overhead. Moreover, the energy efficiency was 1461 MOPS / mW at most which was about 2× that of the baseline structure. The best pipeline depth for an arbitrary energy-performance trade-off became selectable with only 11% area overhead.",

author = "Naoki Ando and Koichiro Masuyama and Hayate Okuhara and Hideharu Amano",

year = "2017",

month = may,

day = "15",

doi = "10.1109/FPT.2016.7929537",

language = "English",

pages = "217--220",

booktitle = "Proceedings of the 2016 International Conference on Field-Programmable Technology, FPT 2016",

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

note = "15th International Conference on Field-Programmable Technology, FPT 2016 ; Conference date: 07-12-2016 Through 09-12-2016",

}

TY - GEN

T1 - Variable pipeline structure for coarse grained reconfigurable array CMA

AU - Ando, Naoki

AU - Masuyama, Koichiro

AU - Okuhara, Hayate

AU - Amano, Hideharu

PY - 2017/5/15

Y1 - 2017/5/15

N2 - Cool mega-array (CMA) is a kind of coarse grained reconfigurable architecture (CGRA) which has shown its ability of ultra low-power computation. However, as CMA completely eliminates clock trees and registers, the performance improvement has been limited. In this paper, we introduce a variable pipeline structure to CMA with the minimum essential registers to provide more wide trade-off between performance and energy. Comparing with the baseline CMA (non-pipelined structure), an average of 77% improvement for performance was achieved with a small power overhead. Moreover, the energy efficiency was 1461 MOPS / mW at most which was about 2× that of the baseline structure. The best pipeline depth for an arbitrary energy-performance trade-off became selectable with only 11% area overhead.

AB - Cool mega-array (CMA) is a kind of coarse grained reconfigurable architecture (CGRA) which has shown its ability of ultra low-power computation. However, as CMA completely eliminates clock trees and registers, the performance improvement has been limited. In this paper, we introduce a variable pipeline structure to CMA with the minimum essential registers to provide more wide trade-off between performance and energy. Comparing with the baseline CMA (non-pipelined structure), an average of 77% improvement for performance was achieved with a small power overhead. Moreover, the energy efficiency was 1461 MOPS / mW at most which was about 2× that of the baseline structure. The best pipeline depth for an arbitrary energy-performance trade-off became selectable with only 11% area overhead.

UR - http://www.scopus.com/inward/record.url?scp=85021410661&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85021410661&partnerID=8YFLogxK

U2 - 10.1109/FPT.2016.7929537

DO - 10.1109/FPT.2016.7929537

M3 - Conference contribution

AN - SCOPUS:85021410661

SP - 217

EP - 220

BT - Proceedings of the 2016 International Conference on Field-Programmable Technology, FPT 2016

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 15th International Conference on Field-Programmable Technology, FPT 2016

Y2 - 7 December 2016 through 9 December 2016

ER -