DocumentCode
722430
Title
A dual-issue embedded processor for low power devices
Author
Lozano, Hanni ; Ito, Mabo
Author_Institution
Dept. of Electr. & Comput. Eng., Univ. of British Columbia, Vancouver, BC, Canada
fYear
2015
fDate
2-4 March 2015
Firstpage
53
Lastpage
61
Abstract
While energy efficiency is essential to extend the battery life of embedded devices, performance cannot be ignored. High performance superscalar embedded processors are more energy efficient than low performance scalar processors, however, they consume more power which is very limited in battery operated or self powered embedded industrial devices. In this paper we propose an energy efficient dual-issue embedded processor that can deliver up to 60% improvement in IPC (instruction-per-cycle) performance with less than 20% increase in power consumption compared to a single-issue scalar processor. In contrast to traditional multi-issue embedded processors that use power intensive superscalar techniques to extract instruction-level parallelism from applications, the proposed processor uses simple hardware techniques to resolve instruction scheduling conflicts. The processor is optimized for implementation on low cost FPGA which makes it a suitable candidate for cost sensitive embedded industrial applications.
Keywords
embedded systems; field programmable gate arrays; microprocessor chips; performance evaluation; IPC performance improvement; battery life extension; battery operated embedded industrial devices; dual-issue embedded processor; energy efficiency; high performance superscalar embedded processors; instruction scheduling conflicts; instruction-per-cycle performance improvement; low cost FPGA; low power devices; self powered embedded industrial devices; Benchmark testing; Field programmable gate arrays; Hardware; Performance evaluation; Pipelines; Random access memory; Registers; embedded processors; low power;
fLanguage
English
Publisher
ieee
Conference_Titel
Industrial Networks and Intelligent Systems (INISCom), 2015 1st International Conference on
Conference_Location
Tokyo
Type
conf
DOI
10.4108/icst.iniscom.2015.258414
Filename
7157822
Link To Document