Hardware-compiler co-design for adjustable data power savings

Hillery C. Hunter, Erik M. Nystrom, Daniel A. Connors, Wen-Mei W Hwu

Research output: Contribution to journalArticle

Abstract

To efficiently accommodate standards changes and algorithmic improvements, functional reconfigurability is increasingly desired for media processing. Such adaptability, however, generally comes at significant power cost. This work suggests that another dimension of adaptation can be beneficial - power adaptation. Through a unique compiler-hardware approach, we (1) demonstrate an extension to the state-of-the-art in data analyzability, yielding better control over scratchpad data management, and (2) combine this knowledge with an SRAM having variable latency and access properties, yielding adjustable power savings. Building upon the compiler techniques presented by [1], we evaluate the severity of the current on-chip storage power problem and detail how SRAM structures can be built to enable data power savings for media applications. We show how the implemented compiler techniques can be applied to other problems in the embedded/media processing domain, and present net data power savings results for a suite of media and telecommunication applications, including MPEG-2, MPEG-4, H.263, and JPEG-2000.

Original languageEnglish (US)
Pages (from-to)244-253
Number of pages10
JournalMicroprocessors and Microsystems
Volume33
Issue number4
DOIs
StatePublished - Jun 1 2009

Fingerprint

Static random access storage
Hardware
Processing
Information management
Telecommunication
Costs

Keywords

  • Adaptation
  • Architecture
  • Compilers
  • Data management
  • Power

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications
  • Artificial Intelligence

Cite this

Hardware-compiler co-design for adjustable data power savings. / Hunter, Hillery C.; Nystrom, Erik M.; Connors, Daniel A.; Hwu, Wen-Mei W.

In: Microprocessors and Microsystems, Vol. 33, No. 4, 01.06.2009, p. 244-253.

Research output: Contribution to journalArticle

Hunter, Hillery C. ; Nystrom, Erik M. ; Connors, Daniel A. ; Hwu, Wen-Mei W. / Hardware-compiler co-design for adjustable data power savings. In: Microprocessors and Microsystems. 2009 ; Vol. 33, No. 4. pp. 244-253.
@article{3574c5e231d84905aa3c73a0010c6d13,
title = "Hardware-compiler co-design for adjustable data power savings",
abstract = "To efficiently accommodate standards changes and algorithmic improvements, functional reconfigurability is increasingly desired for media processing. Such adaptability, however, generally comes at significant power cost. This work suggests that another dimension of adaptation can be beneficial - power adaptation. Through a unique compiler-hardware approach, we (1) demonstrate an extension to the state-of-the-art in data analyzability, yielding better control over scratchpad data management, and (2) combine this knowledge with an SRAM having variable latency and access properties, yielding adjustable power savings. Building upon the compiler techniques presented by [1], we evaluate the severity of the current on-chip storage power problem and detail how SRAM structures can be built to enable data power savings for media applications. We show how the implemented compiler techniques can be applied to other problems in the embedded/media processing domain, and present net data power savings results for a suite of media and telecommunication applications, including MPEG-2, MPEG-4, H.263, and JPEG-2000.",
keywords = "Adaptation, Architecture, Compilers, Data management, Power",
author = "Hunter, {Hillery C.} and Nystrom, {Erik M.} and Connors, {Daniel A.} and Hwu, {Wen-Mei W}",
year = "2009",
month = "6",
day = "1",
doi = "10.1016/j.micpro.2009.02.003",
language = "English (US)",
volume = "33",
pages = "244--253",
journal = "Microprocessors and Microsystems",
issn = "0141-9331",
publisher = "Elsevier",
number = "4",

}

TY - JOUR

T1 - Hardware-compiler co-design for adjustable data power savings

AU - Hunter, Hillery C.

AU - Nystrom, Erik M.

AU - Connors, Daniel A.

AU - Hwu, Wen-Mei W

PY - 2009/6/1

Y1 - 2009/6/1

N2 - To efficiently accommodate standards changes and algorithmic improvements, functional reconfigurability is increasingly desired for media processing. Such adaptability, however, generally comes at significant power cost. This work suggests that another dimension of adaptation can be beneficial - power adaptation. Through a unique compiler-hardware approach, we (1) demonstrate an extension to the state-of-the-art in data analyzability, yielding better control over scratchpad data management, and (2) combine this knowledge with an SRAM having variable latency and access properties, yielding adjustable power savings. Building upon the compiler techniques presented by [1], we evaluate the severity of the current on-chip storage power problem and detail how SRAM structures can be built to enable data power savings for media applications. We show how the implemented compiler techniques can be applied to other problems in the embedded/media processing domain, and present net data power savings results for a suite of media and telecommunication applications, including MPEG-2, MPEG-4, H.263, and JPEG-2000.

AB - To efficiently accommodate standards changes and algorithmic improvements, functional reconfigurability is increasingly desired for media processing. Such adaptability, however, generally comes at significant power cost. This work suggests that another dimension of adaptation can be beneficial - power adaptation. Through a unique compiler-hardware approach, we (1) demonstrate an extension to the state-of-the-art in data analyzability, yielding better control over scratchpad data management, and (2) combine this knowledge with an SRAM having variable latency and access properties, yielding adjustable power savings. Building upon the compiler techniques presented by [1], we evaluate the severity of the current on-chip storage power problem and detail how SRAM structures can be built to enable data power savings for media applications. We show how the implemented compiler techniques can be applied to other problems in the embedded/media processing domain, and present net data power savings results for a suite of media and telecommunication applications, including MPEG-2, MPEG-4, H.263, and JPEG-2000.

KW - Adaptation

KW - Architecture

KW - Compilers

KW - Data management

KW - Power

UR - http://www.scopus.com/inward/record.url?scp=65649085268&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=65649085268&partnerID=8YFLogxK

U2 - 10.1016/j.micpro.2009.02.003

DO - 10.1016/j.micpro.2009.02.003

M3 - Article

AN - SCOPUS:65649085268

VL - 33

SP - 244

EP - 253

JO - Microprocessors and Microsystems

JF - Microprocessors and Microsystems

SN - 0141-9331

IS - 4

ER -