Reliability-aware core partitioning in chip multiprocessors

Isil Oz, Haluk Rahmi Topcuoglu, Mahmut Kandemir, Oguz Tosun

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Executing multiple applications concurrently is an important way of utilizing the computational power provided by emerging chip multiprocessor (CMP) architectures. However, this multiprogramming brings a resource management and partitioning problem, for which one can find numerous examples in the literature. Most of the resource partitioning schemes proposed to date focus on performance or energy centric strategies. In contrast, this paper explores reliability-aware core partitioning strategies targeting CMPs. One of our schemes considers both performance and reliability objectives by maximizing a novel combined metric called the vulnerability-delay product (VDP). The vulnerability component in this metric is represented with Thread Vulnerability Factor (TVF), a recently proposed metric for quantifying thread vulnerability for multicores. Execution time of the given application represents the delay component of the VDP metric. As part of our experimental analysis, proposed core partitioning schemes are compared with respect to normalized weighted speedup, normalized weighted reliability loss and normalized weighted vulnerability delay product gain metrics for various workloads of benchmark applications.

Original languageEnglish (US)
Pages (from-to)160-176
Number of pages17
JournalJournal of Systems Architecture
Volume58
Issue number3-4
DOIs
StatePublished - Mar 1 2012

Fingerprint

Multiprogramming
Core analysis

All Science Journal Classification (ASJC) codes

  • Software
  • Hardware and Architecture

Cite this

Oz, Isil ; Topcuoglu, Haluk Rahmi ; Kandemir, Mahmut ; Tosun, Oguz. / Reliability-aware core partitioning in chip multiprocessors. In: Journal of Systems Architecture. 2012 ; Vol. 58, No. 3-4. pp. 160-176.
@article{9e8d8a86d1c046ad83de32c4d24e6999,
title = "Reliability-aware core partitioning in chip multiprocessors",
abstract = "Executing multiple applications concurrently is an important way of utilizing the computational power provided by emerging chip multiprocessor (CMP) architectures. However, this multiprogramming brings a resource management and partitioning problem, for which one can find numerous examples in the literature. Most of the resource partitioning schemes proposed to date focus on performance or energy centric strategies. In contrast, this paper explores reliability-aware core partitioning strategies targeting CMPs. One of our schemes considers both performance and reliability objectives by maximizing a novel combined metric called the vulnerability-delay product (VDP). The vulnerability component in this metric is represented with Thread Vulnerability Factor (TVF), a recently proposed metric for quantifying thread vulnerability for multicores. Execution time of the given application represents the delay component of the VDP metric. As part of our experimental analysis, proposed core partitioning schemes are compared with respect to normalized weighted speedup, normalized weighted reliability loss and normalized weighted vulnerability delay product gain metrics for various workloads of benchmark applications.",
author = "Isil Oz and Topcuoglu, {Haluk Rahmi} and Mahmut Kandemir and Oguz Tosun",
year = "2012",
month = "3",
day = "1",
doi = "10.1016/j.sysarc.2012.02.005",
language = "English (US)",
volume = "58",
pages = "160--176",
journal = "Journal of Systems Architecture",
issn = "1383-7621",
publisher = "Elsevier",
number = "3-4",

}

Reliability-aware core partitioning in chip multiprocessors. / Oz, Isil; Topcuoglu, Haluk Rahmi; Kandemir, Mahmut; Tosun, Oguz.

In: Journal of Systems Architecture, Vol. 58, No. 3-4, 01.03.2012, p. 160-176.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Reliability-aware core partitioning in chip multiprocessors

AU - Oz, Isil

AU - Topcuoglu, Haluk Rahmi

AU - Kandemir, Mahmut

AU - Tosun, Oguz

PY - 2012/3/1

Y1 - 2012/3/1

N2 - Executing multiple applications concurrently is an important way of utilizing the computational power provided by emerging chip multiprocessor (CMP) architectures. However, this multiprogramming brings a resource management and partitioning problem, for which one can find numerous examples in the literature. Most of the resource partitioning schemes proposed to date focus on performance or energy centric strategies. In contrast, this paper explores reliability-aware core partitioning strategies targeting CMPs. One of our schemes considers both performance and reliability objectives by maximizing a novel combined metric called the vulnerability-delay product (VDP). The vulnerability component in this metric is represented with Thread Vulnerability Factor (TVF), a recently proposed metric for quantifying thread vulnerability for multicores. Execution time of the given application represents the delay component of the VDP metric. As part of our experimental analysis, proposed core partitioning schemes are compared with respect to normalized weighted speedup, normalized weighted reliability loss and normalized weighted vulnerability delay product gain metrics for various workloads of benchmark applications.

AB - Executing multiple applications concurrently is an important way of utilizing the computational power provided by emerging chip multiprocessor (CMP) architectures. However, this multiprogramming brings a resource management and partitioning problem, for which one can find numerous examples in the literature. Most of the resource partitioning schemes proposed to date focus on performance or energy centric strategies. In contrast, this paper explores reliability-aware core partitioning strategies targeting CMPs. One of our schemes considers both performance and reliability objectives by maximizing a novel combined metric called the vulnerability-delay product (VDP). The vulnerability component in this metric is represented with Thread Vulnerability Factor (TVF), a recently proposed metric for quantifying thread vulnerability for multicores. Execution time of the given application represents the delay component of the VDP metric. As part of our experimental analysis, proposed core partitioning schemes are compared with respect to normalized weighted speedup, normalized weighted reliability loss and normalized weighted vulnerability delay product gain metrics for various workloads of benchmark applications.

UR - http://www.scopus.com/inward/record.url?scp=84859418353&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84859418353&partnerID=8YFLogxK

U2 - 10.1016/j.sysarc.2012.02.005

DO - 10.1016/j.sysarc.2012.02.005

M3 - Article

AN - SCOPUS:84859418353

VL - 58

SP - 160

EP - 176

JO - Journal of Systems Architecture

JF - Journal of Systems Architecture

SN - 1383-7621

IS - 3-4

ER -