Exploiting application tunability for efficient, predictable parallel resource management

Fangzhe Chang, Vijay Karamcheti, Zvi Kedem

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Parallel computing is becoming increasing central and mainstream, driven both by the widespread availability of commodity SMP and high-performance cluster platforms, as well as the growing use of parallelism in general-purpose applications such as image recognition, virtual reality, and media processing. In addition to performance requirements, the latter computations impose soft real-time constraints, necessitating efficient, predictable parallel resource management. In this paper, we propose a novel approach for increasing parallel system utilization while meeting application soft real-time deadlines. Our approach exploits the application tunability found in several general-purpose computations. Tunability refers to an application's ability to trade off resource requirements over time, while maintaining a desired level of output quality. We first describe language extensions to support tunability in the Calypso system, then characterize the performance benefits of tunability, using a synthetic task system to systematically identify its benefits. Our results show that application tunability is convenient to express and can significantly improve parallel system utilization for computations with predictability requirements.

Original languageEnglish (US)
Title of host publicationProceedings of the International Parallel Processing Symposium, IPPS
PublisherIEEE
Pages749-758
Number of pages10
StatePublished - 1999
EventProceedings of the 1999 13th International Parallel Processing Symposium and 10th Symposium on Parallel and Distributed Processing - San Juan
Duration: Apr 12 1999Apr 16 1999

Other

OtherProceedings of the 1999 13th International Parallel Processing Symposium and 10th Symposium on Parallel and Distributed Processing
CitySan Juan
Period4/12/994/16/99

Fingerprint

Image recognition
Parallel processing systems
Virtual reality
Availability
Processing

ASJC Scopus subject areas

  • Hardware and Architecture

Cite this

Chang, F., Karamcheti, V., & Kedem, Z. (1999). Exploiting application tunability for efficient, predictable parallel resource management. In Proceedings of the International Parallel Processing Symposium, IPPS (pp. 749-758). IEEE.

Exploiting application tunability for efficient, predictable parallel resource management. / Chang, Fangzhe; Karamcheti, Vijay; Kedem, Zvi.

Proceedings of the International Parallel Processing Symposium, IPPS. IEEE, 1999. p. 749-758.

Research output: Chapter in Book/Report/Conference proceedingChapter

Chang, F, Karamcheti, V & Kedem, Z 1999, Exploiting application tunability for efficient, predictable parallel resource management. in Proceedings of the International Parallel Processing Symposium, IPPS. IEEE, pp. 749-758, Proceedings of the 1999 13th International Parallel Processing Symposium and 10th Symposium on Parallel and Distributed Processing, San Juan, 4/12/99.
Chang F, Karamcheti V, Kedem Z. Exploiting application tunability for efficient, predictable parallel resource management. In Proceedings of the International Parallel Processing Symposium, IPPS. IEEE. 1999. p. 749-758
Chang, Fangzhe ; Karamcheti, Vijay ; Kedem, Zvi. / Exploiting application tunability for efficient, predictable parallel resource management. Proceedings of the International Parallel Processing Symposium, IPPS. IEEE, 1999. pp. 749-758
@inbook{92bfaa86440349b6abd900146533502a,
title = "Exploiting application tunability for efficient, predictable parallel resource management",
abstract = "Parallel computing is becoming increasing central and mainstream, driven both by the widespread availability of commodity SMP and high-performance cluster platforms, as well as the growing use of parallelism in general-purpose applications such as image recognition, virtual reality, and media processing. In addition to performance requirements, the latter computations impose soft real-time constraints, necessitating efficient, predictable parallel resource management. In this paper, we propose a novel approach for increasing parallel system utilization while meeting application soft real-time deadlines. Our approach exploits the application tunability found in several general-purpose computations. Tunability refers to an application's ability to trade off resource requirements over time, while maintaining a desired level of output quality. We first describe language extensions to support tunability in the Calypso system, then characterize the performance benefits of tunability, using a synthetic task system to systematically identify its benefits. Our results show that application tunability is convenient to express and can significantly improve parallel system utilization for computations with predictability requirements.",
author = "Fangzhe Chang and Vijay Karamcheti and Zvi Kedem",
year = "1999",
language = "English (US)",
pages = "749--758",
booktitle = "Proceedings of the International Parallel Processing Symposium, IPPS",
publisher = "IEEE",

}

TY - CHAP

T1 - Exploiting application tunability for efficient, predictable parallel resource management

AU - Chang, Fangzhe

AU - Karamcheti, Vijay

AU - Kedem, Zvi

PY - 1999

Y1 - 1999

N2 - Parallel computing is becoming increasing central and mainstream, driven both by the widespread availability of commodity SMP and high-performance cluster platforms, as well as the growing use of parallelism in general-purpose applications such as image recognition, virtual reality, and media processing. In addition to performance requirements, the latter computations impose soft real-time constraints, necessitating efficient, predictable parallel resource management. In this paper, we propose a novel approach for increasing parallel system utilization while meeting application soft real-time deadlines. Our approach exploits the application tunability found in several general-purpose computations. Tunability refers to an application's ability to trade off resource requirements over time, while maintaining a desired level of output quality. We first describe language extensions to support tunability in the Calypso system, then characterize the performance benefits of tunability, using a synthetic task system to systematically identify its benefits. Our results show that application tunability is convenient to express and can significantly improve parallel system utilization for computations with predictability requirements.

AB - Parallel computing is becoming increasing central and mainstream, driven both by the widespread availability of commodity SMP and high-performance cluster platforms, as well as the growing use of parallelism in general-purpose applications such as image recognition, virtual reality, and media processing. In addition to performance requirements, the latter computations impose soft real-time constraints, necessitating efficient, predictable parallel resource management. In this paper, we propose a novel approach for increasing parallel system utilization while meeting application soft real-time deadlines. Our approach exploits the application tunability found in several general-purpose computations. Tunability refers to an application's ability to trade off resource requirements over time, while maintaining a desired level of output quality. We first describe language extensions to support tunability in the Calypso system, then characterize the performance benefits of tunability, using a synthetic task system to systematically identify its benefits. Our results show that application tunability is convenient to express and can significantly improve parallel system utilization for computations with predictability requirements.

UR - http://www.scopus.com/inward/record.url?scp=0032667358&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032667358&partnerID=8YFLogxK

M3 - Chapter

AN - SCOPUS:0032667358

SP - 749

EP - 758

BT - Proceedings of the International Parallel Processing Symposium, IPPS

PB - IEEE

ER -