View in EDS PDF Full Text

Power Challenges May End the Multicore Era.

Saved in:
Bibliographic Details
Title: Power Challenges May End the Multicore Era.
Authors: Esmaeilzadeh, Hadi1 (AUTHOR) hadianeh@cs.washington.edu, Blem, Emily2 (AUTHOR) blem@cs.wisc.edu, St. Amant, Renée3 (AUTHOR) stamant@cs.utexas.edu, Sankaralingam, Karthikeyan2 (AUTHOR) karu@cs.wisc.edu, Burger, Doug4 (AUTHOR) dburger@microsoft.com
Source: Communications of the ACM. Feb2013, Vol. 56 Issue 2, p93-102. 10p. 1 Diagram, 2 Charts, 6 Graphs.
Subjects: Microprocessor performance, Multicore processors, Performance of transistors, Moore's law, Microprocessor energy consumption, Transistor circuits
Abstract: Starting in 2004, the microprocessor industry has shifted to multicore scaling--increasing the number of cores per die each generation--as its principal strategy for continuing performance growth. Many in the research community believe that this exponential core scaling will continue into the hundreds or thousands of cores per chip, auguring a parallelism revolution in hardware or software. However, while transistor count increases continue at traditional Moore's Law rates, the per-transistor speed and energy efficiency improvements have slowed dramatically. Under these conditions, more cores are only possible if the cores are slower, simpler, or less utilized with each additional technology generation. This paper brings together transistor technology, processor core, and application models to understand whether multicore scaling can sustain the historical exponential performance growth in this energy-limited era. As the number of cores increases, power constraints may prevent powering of all cores at their full speed, requiring a fraction of the cores to be powered off at all times. According to our models, the fraction of these chips that is "dark" may be as much as 50% within three process generations. The low utility of this "dark silicon" may prevent both scaling to higher core counts and ultimately the economic viability of continued silicon scaling. Our results show that core count scaling provides much less performance gain than conventional wisdom suggests. Under (highly) optimistic scaling assumptions--for parallel workloads--multicore scaling provides a 7.9× (23% per year) over ten years. Under more conservative (realistic) assumptions, multicore scaling provides a total performance gain of 3.7× (14% per year) over ten years, and obviously less when sufficiently parallel workloads are unavailable. Without a breakthrough in process technology or microarchitecture, other directions are needed to continue the historical rate of performance improvement. [ABSTRACT FROM AUTHOR]
Copyright of Communications of the ACM is the property of Association for Computing Machinery and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
FullText Links:
  – Type: pdflink
Text:
  Availability: 0
Header DbId: egs
DbLabel: Engineering Source
An: 88141630
AccessLevel: 6
PubType: Periodical
PubTypeId: serialPeriodical
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Power Challenges May End the Multicore Era.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Esmaeilzadeh%2C+Hadi%22">Esmaeilzadeh, Hadi</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> hadianeh@cs.washington.edu</i><br /><searchLink fieldCode="AR" term="%22Blem%2C+Emily%22">Blem, Emily</searchLink><relatesTo>2</relatesTo> (AUTHOR)<i> blem@cs.wisc.edu</i><br /><searchLink fieldCode="AR" term="%22St%2E+Amant%2C+Renée%22">St. Amant, Renée</searchLink><relatesTo>3</relatesTo> (AUTHOR)<i> stamant@cs.utexas.edu</i><br /><searchLink fieldCode="AR" term="%22Sankaralingam%2C+Karthikeyan%22">Sankaralingam, Karthikeyan</searchLink><relatesTo>2</relatesTo> (AUTHOR)<i> karu@cs.wisc.edu</i><br /><searchLink fieldCode="AR" term="%22Burger%2C+Doug%22">Burger, Doug</searchLink><relatesTo>4</relatesTo> (AUTHOR)<i> dburger@microsoft.com</i>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Communications+of+the+ACM%22">Communications of the ACM</searchLink>. Feb2013, Vol. 56 Issue 2, p93-102. 10p. 1 Diagram, 2 Charts, 6 Graphs.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Microprocessor+performance%22">Microprocessor performance</searchLink><br /><searchLink fieldCode="DE" term="%22Multicore+processors%22">Multicore processors</searchLink><br /><searchLink fieldCode="DE" term="%22Performance+of+transistors%22">Performance of transistors</searchLink><br /><searchLink fieldCode="DE" term="%22Moore's+law%22">Moore's law</searchLink><br /><searchLink fieldCode="DE" term="%22Microprocessor+energy+consumption%22">Microprocessor energy consumption</searchLink><br /><searchLink fieldCode="DE" term="%22Transistor+circuits%22">Transistor circuits</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Starting in 2004, the microprocessor industry has shifted to multicore scaling--increasing the number of cores per die each generation--as its principal strategy for continuing performance growth. Many in the research community believe that this exponential core scaling will continue into the hundreds or thousands of cores per chip, auguring a parallelism revolution in hardware or software. However, while transistor count increases continue at traditional Moore's Law rates, the per-transistor speed and energy efficiency improvements have slowed dramatically. Under these conditions, more cores are only possible if the cores are slower, simpler, or less utilized with each additional technology generation. This paper brings together transistor technology, processor core, and application models to understand whether multicore scaling can sustain the historical exponential performance growth in this energy-limited era. As the number of cores increases, power constraints may prevent powering of all cores at their full speed, requiring a fraction of the cores to be powered off at all times. According to our models, the fraction of these chips that is "dark" may be as much as 50% within three process generations. The low utility of this "dark silicon" may prevent both scaling to higher core counts and ultimately the economic viability of continued silicon scaling. Our results show that core count scaling provides much less performance gain than conventional wisdom suggests. Under (highly) optimistic scaling assumptions--for parallel workloads--multicore scaling provides a 7.9× (23% per year) over ten years. Under more conservative (realistic) assumptions, multicore scaling provides a total performance gain of 3.7× (14% per year) over ten years, and obviously less when sufficiently parallel workloads are unavailable. Without a breakthrough in process technology or microarchitecture, other directions are needed to continue the historical rate of performance improvement. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Communications of the ACM is the property of Association for Computing Machinery and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=88141630
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1145/2408776.2408797
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 10
        StartPage: 93
    Subjects:
      – SubjectFull: Microprocessor performance
        Type: general
      – SubjectFull: Multicore processors
        Type: general
      – SubjectFull: Performance of transistors
        Type: general
      – SubjectFull: Moore's law
        Type: general
      – SubjectFull: Microprocessor energy consumption
        Type: general
      – SubjectFull: Transistor circuits
        Type: general
    Titles:
      – TitleFull: Power Challenges May End the Multicore Era.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Esmaeilzadeh, Hadi
      – PersonEntity:
          Name:
            NameFull: Blem, Emily
      – PersonEntity:
          Name:
            NameFull: St. Amant, Renée
      – PersonEntity:
          Name:
            NameFull: Sankaralingam, Karthikeyan
      – PersonEntity:
          Name:
            NameFull: Burger, Doug
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 02
              Text: Feb2013
              Type: published
              Y: 2013
          Identifiers:
            – Type: issn-print
              Value: 00010782
          Numbering:
            – Type: volume
              Value: 56
            – Type: issue
              Value: 2
          Titles:
            – TitleFull: Communications of the ACM
              Type: main
ResultId 1