Heterogeneous dark silicon chip multi-processors

Design and run-time management

Siddharth Garg, Yatish Turakhia, Diana Marculescu

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

A fundamental research question given the dark silicon problem is how best to leverage the abundance of transistors on the chip. In this chapter, we describe two solutions to this problem. In the first, we exploit the inherent variations in process parameters that exist in scaled technologies to offer increased performance. Since process variations result in core-to-core variations in power and frequency, the idea is to cherry pick the best subset of cores for an application so as to maximize performance within the power budget. Second, we describe an approach for synthesis of micro-architecturally dark silicon chip multi-processors. The goal is to determine the optimal number of cores of each type to provision the processor with, such that the area and power budgets are met and the application performance is maximized.

Original languageEnglish (US)
Title of host publicationThe Dark Side of Silicon
Subtitle of host publicationEnergy Efficient Computing in the Dark Silicon Era
PublisherSpringer International Publishing
Pages95-122
Number of pages28
ISBN (Electronic)9783319315966
ISBN (Print)9783319315942
DOIs
StatePublished - Jan 1 2017

Fingerprint

Silicon
Transistors

ASJC Scopus subject areas

  • Engineering(all)
  • Computer Science(all)

Cite this

Garg, S., Turakhia, Y., & Marculescu, D. (2017). Heterogeneous dark silicon chip multi-processors: Design and run-time management. In The Dark Side of Silicon: Energy Efficient Computing in the Dark Silicon Era (pp. 95-122). Springer International Publishing. https://doi.org/10.1007/978-3-319-31596-6_4

Heterogeneous dark silicon chip multi-processors : Design and run-time management. / Garg, Siddharth; Turakhia, Yatish; Marculescu, Diana.

The Dark Side of Silicon: Energy Efficient Computing in the Dark Silicon Era. Springer International Publishing, 2017. p. 95-122.

Research output: Chapter in Book/Report/Conference proceedingChapter

Garg, S, Turakhia, Y & Marculescu, D 2017, Heterogeneous dark silicon chip multi-processors: Design and run-time management. in The Dark Side of Silicon: Energy Efficient Computing in the Dark Silicon Era. Springer International Publishing, pp. 95-122. https://doi.org/10.1007/978-3-319-31596-6_4
Garg S, Turakhia Y, Marculescu D. Heterogeneous dark silicon chip multi-processors: Design and run-time management. In The Dark Side of Silicon: Energy Efficient Computing in the Dark Silicon Era. Springer International Publishing. 2017. p. 95-122 https://doi.org/10.1007/978-3-319-31596-6_4
Garg, Siddharth ; Turakhia, Yatish ; Marculescu, Diana. / Heterogeneous dark silicon chip multi-processors : Design and run-time management. The Dark Side of Silicon: Energy Efficient Computing in the Dark Silicon Era. Springer International Publishing, 2017. pp. 95-122
@inbook{6e62095f4c7e41768c060dd2ea30e238,
title = "Heterogeneous dark silicon chip multi-processors: Design and run-time management",
abstract = "A fundamental research question given the dark silicon problem is how best to leverage the abundance of transistors on the chip. In this chapter, we describe two solutions to this problem. In the first, we exploit the inherent variations in process parameters that exist in scaled technologies to offer increased performance. Since process variations result in core-to-core variations in power and frequency, the idea is to cherry pick the best subset of cores for an application so as to maximize performance within the power budget. Second, we describe an approach for synthesis of micro-architecturally dark silicon chip multi-processors. The goal is to determine the optimal number of cores of each type to provision the processor with, such that the area and power budgets are met and the application performance is maximized.",
author = "Siddharth Garg and Yatish Turakhia and Diana Marculescu",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/978-3-319-31596-6_4",
language = "English (US)",
isbn = "9783319315942",
pages = "95--122",
booktitle = "The Dark Side of Silicon",
publisher = "Springer International Publishing",

}

TY - CHAP

T1 - Heterogeneous dark silicon chip multi-processors

T2 - Design and run-time management

AU - Garg, Siddharth

AU - Turakhia, Yatish

AU - Marculescu, Diana

PY - 2017/1/1

Y1 - 2017/1/1

N2 - A fundamental research question given the dark silicon problem is how best to leverage the abundance of transistors on the chip. In this chapter, we describe two solutions to this problem. In the first, we exploit the inherent variations in process parameters that exist in scaled technologies to offer increased performance. Since process variations result in core-to-core variations in power and frequency, the idea is to cherry pick the best subset of cores for an application so as to maximize performance within the power budget. Second, we describe an approach for synthesis of micro-architecturally dark silicon chip multi-processors. The goal is to determine the optimal number of cores of each type to provision the processor with, such that the area and power budgets are met and the application performance is maximized.

AB - A fundamental research question given the dark silicon problem is how best to leverage the abundance of transistors on the chip. In this chapter, we describe two solutions to this problem. In the first, we exploit the inherent variations in process parameters that exist in scaled technologies to offer increased performance. Since process variations result in core-to-core variations in power and frequency, the idea is to cherry pick the best subset of cores for an application so as to maximize performance within the power budget. Second, we describe an approach for synthesis of micro-architecturally dark silicon chip multi-processors. The goal is to determine the optimal number of cores of each type to provision the processor with, such that the area and power budgets are met and the application performance is maximized.

UR - http://www.scopus.com/inward/record.url?scp=85028836785&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85028836785&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-31596-6_4

DO - 10.1007/978-3-319-31596-6_4

M3 - Chapter

SN - 9783319315942

SP - 95

EP - 122

BT - The Dark Side of Silicon

PB - Springer International Publishing

ER -