Cool mega-array: A highly energy efficient reconfigurable accelerator

N. Ozaki, Y. Yoshihiro, Y. Saito, D. Ikebuchi, M. Kimura, H. Amano, H. Nakamura, Kimiyoshi Usami, M. Namiki, M. Kondo

研究成果: Conference contribution

13 引用 (Scopus)

抄録

A highly energy efficient reconfigurable accelerator called CMA (Cool Mega-Array) is proposed. It consists of a large Processing Element (PE) array without memory elements for maintain result of ALU and configuration data, a small simple programmable micro controller for data management, and the data memory. Unlike traditional coarse grained reconfigurable processors, the power consumption for hardware context switching, storing intermediate data in registers, and clock distribution for them are eliminated from PE array which occupies large area of a chip. Configuration registers are collected to small area of micro controller. The data flow graph mapped on the PE array is static during execution. Various application programs can be implemented by making the best use of flexible data management instructions with the micro controller. When the delay time in the PE array is longer than the data handling time with the micro controller, the supply voltage for the PE array is scaled to reduce the power consumption without degrading the performance. In the opposite case, wave pipelining is applied to enhance PE array performance. A prototype chip CMA-1 with 8 × 8 PE array with 24-bit data width was fabricated in 2.1 × 4.2mm 2 65-nm CMOS technology, and achieves 2.4-GOPS/11.2-mW sustained performance. This energy efficiency is comparable to that of the most energy efficient accelerators that have been reported.

元の言語English
ホスト出版物のタイトル2011 International Conference on Field-Programmable Technology, FPT 2011
DOI
出版物ステータスPublished - 2011
イベント2011 International Conference on Field-Programmable Technology, FPT 2011 - New Delhi
継続期間: 2011 12 122011 12 14

Other

Other2011 International Conference on Field-Programmable Technology, FPT 2011
New Delhi
期間11/12/1211/12/14

Fingerprint

Particle accelerators
Processing
Controllers
Information management
Electric power utilization
Data flow graphs
Data storage equipment
Data handling
Application programs
Energy efficiency
Clocks
Time delay
Hardware
Electric potential

ASJC Scopus subject areas

  • Computational Mathematics

これを引用

Ozaki, N., Yoshihiro, Y., Saito, Y., Ikebuchi, D., Kimura, M., Amano, H., ... Kondo, M. (2011). Cool mega-array: A highly energy efficient reconfigurable accelerator. : 2011 International Conference on Field-Programmable Technology, FPT 2011 [6132668] https://doi.org/10.1109/FPT.2011.6132668

Cool mega-array : A highly energy efficient reconfigurable accelerator. / Ozaki, N.; Yoshihiro, Y.; Saito, Y.; Ikebuchi, D.; Kimura, M.; Amano, H.; Nakamura, H.; Usami, Kimiyoshi; Namiki, M.; Kondo, M.

2011 International Conference on Field-Programmable Technology, FPT 2011. 2011. 6132668.

研究成果: Conference contribution

Ozaki, N, Yoshihiro, Y, Saito, Y, Ikebuchi, D, Kimura, M, Amano, H, Nakamura, H, Usami, K, Namiki, M & Kondo, M 2011, Cool mega-array: A highly energy efficient reconfigurable accelerator. : 2011 International Conference on Field-Programmable Technology, FPT 2011., 6132668, 2011 International Conference on Field-Programmable Technology, FPT 2011, New Delhi, 11/12/12. https://doi.org/10.1109/FPT.2011.6132668
Ozaki N, Yoshihiro Y, Saito Y, Ikebuchi D, Kimura M, Amano H その他. Cool mega-array: A highly energy efficient reconfigurable accelerator. : 2011 International Conference on Field-Programmable Technology, FPT 2011. 2011. 6132668 https://doi.org/10.1109/FPT.2011.6132668
Ozaki, N. ; Yoshihiro, Y. ; Saito, Y. ; Ikebuchi, D. ; Kimura, M. ; Amano, H. ; Nakamura, H. ; Usami, Kimiyoshi ; Namiki, M. ; Kondo, M. / Cool mega-array : A highly energy efficient reconfigurable accelerator. 2011 International Conference on Field-Programmable Technology, FPT 2011. 2011.
@inproceedings{55fff4a5aa2141aa8512629a5c932ec7,
title = "Cool mega-array: A highly energy efficient reconfigurable accelerator",
abstract = "A highly energy efficient reconfigurable accelerator called CMA (Cool Mega-Array) is proposed. It consists of a large Processing Element (PE) array without memory elements for maintain result of ALU and configuration data, a small simple programmable micro controller for data management, and the data memory. Unlike traditional coarse grained reconfigurable processors, the power consumption for hardware context switching, storing intermediate data in registers, and clock distribution for them are eliminated from PE array which occupies large area of a chip. Configuration registers are collected to small area of micro controller. The data flow graph mapped on the PE array is static during execution. Various application programs can be implemented by making the best use of flexible data management instructions with the micro controller. When the delay time in the PE array is longer than the data handling time with the micro controller, the supply voltage for the PE array is scaled to reduce the power consumption without degrading the performance. In the opposite case, wave pipelining is applied to enhance PE array performance. A prototype chip CMA-1 with 8 × 8 PE array with 24-bit data width was fabricated in 2.1 × 4.2mm 2 65-nm CMOS technology, and achieves 2.4-GOPS/11.2-mW sustained performance. This energy efficiency is comparable to that of the most energy efficient accelerators that have been reported.",
author = "N. Ozaki and Y. Yoshihiro and Y. Saito and D. Ikebuchi and M. Kimura and H. Amano and H. Nakamura and Kimiyoshi Usami and M. Namiki and M. Kondo",
year = "2011",
doi = "10.1109/FPT.2011.6132668",
language = "English",
isbn = "9781457717406",
booktitle = "2011 International Conference on Field-Programmable Technology, FPT 2011",

}

TY - GEN

T1 - Cool mega-array

T2 - A highly energy efficient reconfigurable accelerator

AU - Ozaki, N.

AU - Yoshihiro, Y.

AU - Saito, Y.

AU - Ikebuchi, D.

AU - Kimura, M.

AU - Amano, H.

AU - Nakamura, H.

AU - Usami, Kimiyoshi

AU - Namiki, M.

AU - Kondo, M.

PY - 2011

Y1 - 2011

N2 - A highly energy efficient reconfigurable accelerator called CMA (Cool Mega-Array) is proposed. It consists of a large Processing Element (PE) array without memory elements for maintain result of ALU and configuration data, a small simple programmable micro controller for data management, and the data memory. Unlike traditional coarse grained reconfigurable processors, the power consumption for hardware context switching, storing intermediate data in registers, and clock distribution for them are eliminated from PE array which occupies large area of a chip. Configuration registers are collected to small area of micro controller. The data flow graph mapped on the PE array is static during execution. Various application programs can be implemented by making the best use of flexible data management instructions with the micro controller. When the delay time in the PE array is longer than the data handling time with the micro controller, the supply voltage for the PE array is scaled to reduce the power consumption without degrading the performance. In the opposite case, wave pipelining is applied to enhance PE array performance. A prototype chip CMA-1 with 8 × 8 PE array with 24-bit data width was fabricated in 2.1 × 4.2mm 2 65-nm CMOS technology, and achieves 2.4-GOPS/11.2-mW sustained performance. This energy efficiency is comparable to that of the most energy efficient accelerators that have been reported.

AB - A highly energy efficient reconfigurable accelerator called CMA (Cool Mega-Array) is proposed. It consists of a large Processing Element (PE) array without memory elements for maintain result of ALU and configuration data, a small simple programmable micro controller for data management, and the data memory. Unlike traditional coarse grained reconfigurable processors, the power consumption for hardware context switching, storing intermediate data in registers, and clock distribution for them are eliminated from PE array which occupies large area of a chip. Configuration registers are collected to small area of micro controller. The data flow graph mapped on the PE array is static during execution. Various application programs can be implemented by making the best use of flexible data management instructions with the micro controller. When the delay time in the PE array is longer than the data handling time with the micro controller, the supply voltage for the PE array is scaled to reduce the power consumption without degrading the performance. In the opposite case, wave pipelining is applied to enhance PE array performance. A prototype chip CMA-1 with 8 × 8 PE array with 24-bit data width was fabricated in 2.1 × 4.2mm 2 65-nm CMOS technology, and achieves 2.4-GOPS/11.2-mW sustained performance. This energy efficiency is comparable to that of the most energy efficient accelerators that have been reported.

UR - http://www.scopus.com/inward/record.url?scp=84857214326&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84857214326&partnerID=8YFLogxK

U2 - 10.1109/FPT.2011.6132668

DO - 10.1109/FPT.2011.6132668

M3 - Conference contribution

AN - SCOPUS:84857214326

SN - 9781457717406

BT - 2011 International Conference on Field-Programmable Technology, FPT 2011

ER -