Thread Progress Equalization: Dynamically Adaptive Power-Constrained Performance Optimization of Multi-Threaded Applications

Yatish Turakhia, Guangshuo Liu, Siddharth Garg, Diana Marculescu

Research output: Contribution to journalArticle

Abstract

Dynamically adaptive multi-core architectures have been proposed as an effective solution to optimize performance for peak power constrained processors. In processors, the micro-architectural parameters or voltage/frequency of each core can be changed at run-time, thus providing a range of power/performance operating points for each core. In this paper, we propose Thread Progress Equalization (TPEq), a run-time mechanism for power constrained performance maximization of multithreaded applications running on dynamically adaptive multicore processors. Compared to existing approaches, TPEq (i) identifies and addresses two primary sources of inter-thread heterogeneity in multithreaded applications, (ii) determines the optimal core configurations in polynomial time with respect to the number of cores and configurations, and (iii) requires no modifications in the user-level source code. Our experimental evaluations demonstrate that TPEq outperforms state-of-the-art run-time power/performance optimization techniques proposed in literature for dynamically adaptive multicores by up to 23 percent.

Original languageEnglish (US)
Article number7565594
Pages (from-to)731-744
Number of pages14
JournalIEEE Transactions on Computers
Volume66
Issue number4
DOIs
StatePublished - Apr 1 2017

Fingerprint

Performance Optimization
Equalization
Constrained Optimization
Thread
Polynomials
Configuration
Multi-core Processor
Electric potential
Experimental Evaluation
Optimization Techniques
Percent
Polynomial time
Voltage
Optimise
Range of data
Demonstrate
Architecture

Keywords

  • Multi-threaded applications
  • power-constrained performance maximization
  • thread progress

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Software
  • Hardware and Architecture
  • Computational Theory and Mathematics

Cite this

Thread Progress Equalization : Dynamically Adaptive Power-Constrained Performance Optimization of Multi-Threaded Applications. / Turakhia, Yatish; Liu, Guangshuo; Garg, Siddharth; Marculescu, Diana.

In: IEEE Transactions on Computers, Vol. 66, No. 4, 7565594, 01.04.2017, p. 731-744.

Research output: Contribution to journalArticle

@article{02a06c4238f44695ae62f4850220074e,
title = "Thread Progress Equalization: Dynamically Adaptive Power-Constrained Performance Optimization of Multi-Threaded Applications",
abstract = "Dynamically adaptive multi-core architectures have been proposed as an effective solution to optimize performance for peak power constrained processors. In processors, the micro-architectural parameters or voltage/frequency of each core can be changed at run-time, thus providing a range of power/performance operating points for each core. In this paper, we propose Thread Progress Equalization (TPEq), a run-time mechanism for power constrained performance maximization of multithreaded applications running on dynamically adaptive multicore processors. Compared to existing approaches, TPEq (i) identifies and addresses two primary sources of inter-thread heterogeneity in multithreaded applications, (ii) determines the optimal core configurations in polynomial time with respect to the number of cores and configurations, and (iii) requires no modifications in the user-level source code. Our experimental evaluations demonstrate that TPEq outperforms state-of-the-art run-time power/performance optimization techniques proposed in literature for dynamically adaptive multicores by up to 23 percent.",
keywords = "Multi-threaded applications, power-constrained performance maximization, thread progress",
author = "Yatish Turakhia and Guangshuo Liu and Siddharth Garg and Diana Marculescu",
year = "2017",
month = "4",
day = "1",
doi = "10.1109/TC.2016.2608951",
language = "English (US)",
volume = "66",
pages = "731--744",
journal = "IEEE Transactions on Computers",
issn = "0018-9340",
publisher = "IEEE Computer Society",
number = "4",

}

TY - JOUR

T1 - Thread Progress Equalization

T2 - Dynamically Adaptive Power-Constrained Performance Optimization of Multi-Threaded Applications

AU - Turakhia, Yatish

AU - Liu, Guangshuo

AU - Garg, Siddharth

AU - Marculescu, Diana

PY - 2017/4/1

Y1 - 2017/4/1

N2 - Dynamically adaptive multi-core architectures have been proposed as an effective solution to optimize performance for peak power constrained processors. In processors, the micro-architectural parameters or voltage/frequency of each core can be changed at run-time, thus providing a range of power/performance operating points for each core. In this paper, we propose Thread Progress Equalization (TPEq), a run-time mechanism for power constrained performance maximization of multithreaded applications running on dynamically adaptive multicore processors. Compared to existing approaches, TPEq (i) identifies and addresses two primary sources of inter-thread heterogeneity in multithreaded applications, (ii) determines the optimal core configurations in polynomial time with respect to the number of cores and configurations, and (iii) requires no modifications in the user-level source code. Our experimental evaluations demonstrate that TPEq outperforms state-of-the-art run-time power/performance optimization techniques proposed in literature for dynamically adaptive multicores by up to 23 percent.

AB - Dynamically adaptive multi-core architectures have been proposed as an effective solution to optimize performance for peak power constrained processors. In processors, the micro-architectural parameters or voltage/frequency of each core can be changed at run-time, thus providing a range of power/performance operating points for each core. In this paper, we propose Thread Progress Equalization (TPEq), a run-time mechanism for power constrained performance maximization of multithreaded applications running on dynamically adaptive multicore processors. Compared to existing approaches, TPEq (i) identifies and addresses two primary sources of inter-thread heterogeneity in multithreaded applications, (ii) determines the optimal core configurations in polynomial time with respect to the number of cores and configurations, and (iii) requires no modifications in the user-level source code. Our experimental evaluations demonstrate that TPEq outperforms state-of-the-art run-time power/performance optimization techniques proposed in literature for dynamically adaptive multicores by up to 23 percent.

KW - Multi-threaded applications

KW - power-constrained performance maximization

KW - thread progress

UR - http://www.scopus.com/inward/record.url?scp=85015679605&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015679605&partnerID=8YFLogxK

U2 - 10.1109/TC.2016.2608951

DO - 10.1109/TC.2016.2608951

M3 - Article

AN - SCOPUS:85015679605

VL - 66

SP - 731

EP - 744

JO - IEEE Transactions on Computers

JF - IEEE Transactions on Computers

SN - 0018-9340

IS - 4

M1 - 7565594

ER -