Exploiting residue number system for power-efficient digital signal processing in embedded processors

Rooju Chokshi, Krzysztof S. Berezowski, Aviral Shrivastava, Stanislaw J. Piestrak

Research output: Chapter in Book/Report/Conference proceedingConference contribution

26 Citations (Scopus)

Abstract

2's complement number system imposes a fundamental limitation on the power and performance of arithmetic circuits, due to the fundamental need of cross-datapath carry propagation. Residue Number System (RNS) breaks free of these bonds by decomposing a number into parts and performing arithmetic operations in parallel, significantly reducing the breadth of carry propagation. Consequently, RNS arithmetic has been proposed as a solution to improve the power-efficiency of arithmetic hardware. However, limitations of the expressiveness of RNS in terms of arithmetic operations together with overheads related to interaction with 2's complement arithmetic make programmable processor design that takes advantage of these benefits challenging. In this paper we meet this challenge by multi-tier synergistic co-design of architecture, micro-architecture, hardware components, as well as compilation techniques. Our experiments not only demonstrate simultaneous improvement of up to 30% in performance and 57% reduction in functional unit power consumption, but also that most of these benefits can be exploited with automatically generated code.

Original languageEnglish (US)
Title of host publicationEmbedded Systems Week 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09
Pages19-27
Number of pages9
DOIs
StatePublished - 2009
EventEmbedded Systems Week 2009, ESWEEK 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09 - Grenoble, France
Duration: Oct 11 2009Oct 16 2009

Other

OtherEmbedded Systems Week 2009, ESWEEK 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09
CountryFrance
CityGrenoble
Period10/11/0910/16/09

Fingerprint

Numbering systems
Digital signal processing
Hardware
Electric power utilization
Networks (circuits)
Experiments

Keywords

  • Compiler
  • Performance
  • Power
  • Processor
  • Residue Number System

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software
  • Electrical and Electronic Engineering

Cite this

Chokshi, R., Berezowski, K. S., Shrivastava, A., & Piestrak, S. J. (2009). Exploiting residue number system for power-efficient digital signal processing in embedded processors. In Embedded Systems Week 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09 (pp. 19-27) https://doi.org/10.1145/1629395.1629401

Exploiting residue number system for power-efficient digital signal processing in embedded processors. / Chokshi, Rooju; Berezowski, Krzysztof S.; Shrivastava, Aviral; Piestrak, Stanislaw J.

Embedded Systems Week 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09. 2009. p. 19-27.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Chokshi, R, Berezowski, KS, Shrivastava, A & Piestrak, SJ 2009, Exploiting residue number system for power-efficient digital signal processing in embedded processors. in Embedded Systems Week 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09. pp. 19-27, Embedded Systems Week 2009, ESWEEK 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09, Grenoble, France, 10/11/09. https://doi.org/10.1145/1629395.1629401
Chokshi R, Berezowski KS, Shrivastava A, Piestrak SJ. Exploiting residue number system for power-efficient digital signal processing in embedded processors. In Embedded Systems Week 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09. 2009. p. 19-27 https://doi.org/10.1145/1629395.1629401
Chokshi, Rooju ; Berezowski, Krzysztof S. ; Shrivastava, Aviral ; Piestrak, Stanislaw J. / Exploiting residue number system for power-efficient digital signal processing in embedded processors. Embedded Systems Week 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09. 2009. pp. 19-27
@inproceedings{3d039eec49634b4d8b0e2ed3dc56ff13,
title = "Exploiting residue number system for power-efficient digital signal processing in embedded processors",
abstract = "2's complement number system imposes a fundamental limitation on the power and performance of arithmetic circuits, due to the fundamental need of cross-datapath carry propagation. Residue Number System (RNS) breaks free of these bonds by decomposing a number into parts and performing arithmetic operations in parallel, significantly reducing the breadth of carry propagation. Consequently, RNS arithmetic has been proposed as a solution to improve the power-efficiency of arithmetic hardware. However, limitations of the expressiveness of RNS in terms of arithmetic operations together with overheads related to interaction with 2's complement arithmetic make programmable processor design that takes advantage of these benefits challenging. In this paper we meet this challenge by multi-tier synergistic co-design of architecture, micro-architecture, hardware components, as well as compilation techniques. Our experiments not only demonstrate simultaneous improvement of up to 30{\%} in performance and 57{\%} reduction in functional unit power consumption, but also that most of these benefits can be exploited with automatically generated code.",
keywords = "Compiler, Performance, Power, Processor, Residue Number System",
author = "Rooju Chokshi and Berezowski, {Krzysztof S.} and Aviral Shrivastava and Piestrak, {Stanislaw J.}",
year = "2009",
doi = "10.1145/1629395.1629401",
language = "English (US)",
isbn = "9781605586267",
pages = "19--27",
booktitle = "Embedded Systems Week 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09",

}

TY - GEN

T1 - Exploiting residue number system for power-efficient digital signal processing in embedded processors

AU - Chokshi, Rooju

AU - Berezowski, Krzysztof S.

AU - Shrivastava, Aviral

AU - Piestrak, Stanislaw J.

PY - 2009

Y1 - 2009

N2 - 2's complement number system imposes a fundamental limitation on the power and performance of arithmetic circuits, due to the fundamental need of cross-datapath carry propagation. Residue Number System (RNS) breaks free of these bonds by decomposing a number into parts and performing arithmetic operations in parallel, significantly reducing the breadth of carry propagation. Consequently, RNS arithmetic has been proposed as a solution to improve the power-efficiency of arithmetic hardware. However, limitations of the expressiveness of RNS in terms of arithmetic operations together with overheads related to interaction with 2's complement arithmetic make programmable processor design that takes advantage of these benefits challenging. In this paper we meet this challenge by multi-tier synergistic co-design of architecture, micro-architecture, hardware components, as well as compilation techniques. Our experiments not only demonstrate simultaneous improvement of up to 30% in performance and 57% reduction in functional unit power consumption, but also that most of these benefits can be exploited with automatically generated code.

AB - 2's complement number system imposes a fundamental limitation on the power and performance of arithmetic circuits, due to the fundamental need of cross-datapath carry propagation. Residue Number System (RNS) breaks free of these bonds by decomposing a number into parts and performing arithmetic operations in parallel, significantly reducing the breadth of carry propagation. Consequently, RNS arithmetic has been proposed as a solution to improve the power-efficiency of arithmetic hardware. However, limitations of the expressiveness of RNS in terms of arithmetic operations together with overheads related to interaction with 2's complement arithmetic make programmable processor design that takes advantage of these benefits challenging. In this paper we meet this challenge by multi-tier synergistic co-design of architecture, micro-architecture, hardware components, as well as compilation techniques. Our experiments not only demonstrate simultaneous improvement of up to 30% in performance and 57% reduction in functional unit power consumption, but also that most of these benefits can be exploited with automatically generated code.

KW - Compiler

KW - Performance

KW - Power

KW - Processor

KW - Residue Number System

UR - http://www.scopus.com/inward/record.url?scp=72049113126&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=72049113126&partnerID=8YFLogxK

U2 - 10.1145/1629395.1629401

DO - 10.1145/1629395.1629401

M3 - Conference contribution

AN - SCOPUS:72049113126

SN - 9781605586267

SP - 19

EP - 27

BT - Embedded Systems Week 2009 - 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems, CASES'09

ER -