A multi-granularity parallelism object recognition processor with content-aware fine-grained task scheduling

Author

Junyoung Park ; Injoon Hong ; Gyeonghoon Kim ; Youchang Kim ; Kyuho Lee ; Seongwook Park ; Kyeongryeol Bong ; Hoi-Jun Yoo

Author_Institution

Dept. of Electr. Eng., Korea Adv. Inst. of Sci. & Technol. (KAIST), Daejeon, South Korea

fYear

2013

fDate

17-19 April 2013

Firstpage

1

Lastpage

3

Abstract

Multiple granularity parallel core architecture is proposed to accelerate object recognition with low area and energy consumption. By adopting task-level optimized cores with different parallelism and complexity, the proposed processor achieves real-time object recognition with 271.4 GOPS peak performance. In addition, content-aware fine-grained task scheduling is proposed to enable low power real-time object recognition on 30fps 720p HD video streams. As a result, the object recognition processor achieves 9.4nJ/pixel energy efficiency and 25.8 GOPS/W·mm² power-area efficiency in O.13um CMOS technology.

Keywords

CMOS integrated circuits; low-power electronics; microprocessor chips; object recognition; CMOS technology; GOPS peak performance; content-aware fine-grained task scheduling; low area; low energy consumption; multigranularity parallelism; object recognition processor; Computer architecture; Feature extraction; High definition video; Object recognition; Processor scheduling; Real-time systems; computer architecture; multicore processor; object recognition; task scheduling;

fLanguage

English

Publisher

ieee

Conference_Titel

Cool Chips XVI (COOL Chips), 2013 IEEE

Conference_Location

Yokohama

Print_ISBN

978-1-4673-5780-7

Electronic_ISBN

978-1-4673-5781-4

Type

conf

DOI

10.1109/CoolChips.2013.6547917

Filename

6547917