Bounding cache miss costs of multithreaded computations under general schedulers

Richard Cole, Vijaya Ramachandran

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We analyze the caching overhead incurred by a class of multithreaded algorithms when scheduled by an arbitrary scheduler. We obtain bounds that match or improve upon the well-known O(Q + S · (M/B)) caching cost for the randomized work stealing (RWS) scheduler, where S is the number of steals, Q is the sequential caching cost, and M and B are the cache size and block (or cache line) size respectively.

Original languageEnglish (US)
Title of host publicationSPAA 2017 - Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures
PublisherAssociation for Computing Machinery
Pages351-362
Number of pages12
VolumePart F129316
ISBN (Electronic)9781450345934
DOIs
StatePublished - Jul 24 2017
Event29th ACM Symposium on Parallelism in Algorithms and Architectures, SPAA 2017 - Washington, United States
Duration: Jul 24 2017Jul 26 2017

Other

Other29th ACM Symposium on Parallelism in Algorithms and Architectures, SPAA 2017
CountryUnited States
CityWashington
Period7/24/177/26/17

Fingerprint

Caching
Scheduler
Cache
Costs
Line
Arbitrary

Keywords

  • Cache oblivious
  • Fork-Join algorithms
  • Schedulers

ASJC Scopus subject areas

  • Software
  • Theoretical Computer Science
  • Hardware and Architecture

Cite this

Cole, R., & Ramachandran, V. (2017). Bounding cache miss costs of multithreaded computations under general schedulers. In SPAA 2017 - Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures (Vol. Part F129316, pp. 351-362). Association for Computing Machinery. https://doi.org/10.1145/3087556.3087572

Bounding cache miss costs of multithreaded computations under general schedulers. / Cole, Richard; Ramachandran, Vijaya.

SPAA 2017 - Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures. Vol. Part F129316 Association for Computing Machinery, 2017. p. 351-362.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Cole, R & Ramachandran, V 2017, Bounding cache miss costs of multithreaded computations under general schedulers. in SPAA 2017 - Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures. vol. Part F129316, Association for Computing Machinery, pp. 351-362, 29th ACM Symposium on Parallelism in Algorithms and Architectures, SPAA 2017, Washington, United States, 7/24/17. https://doi.org/10.1145/3087556.3087572
Cole R, Ramachandran V. Bounding cache miss costs of multithreaded computations under general schedulers. In SPAA 2017 - Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures. Vol. Part F129316. Association for Computing Machinery. 2017. p. 351-362 https://doi.org/10.1145/3087556.3087572
Cole, Richard ; Ramachandran, Vijaya. / Bounding cache miss costs of multithreaded computations under general schedulers. SPAA 2017 - Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures. Vol. Part F129316 Association for Computing Machinery, 2017. pp. 351-362
@inproceedings{97bac5f6a1274a7394e4cc3ec210afc9,
title = "Bounding cache miss costs of multithreaded computations under general schedulers",
abstract = "We analyze the caching overhead incurred by a class of multithreaded algorithms when scheduled by an arbitrary scheduler. We obtain bounds that match or improve upon the well-known O(Q + S · (M/B)) caching cost for the randomized work stealing (RWS) scheduler, where S is the number of steals, Q is the sequential caching cost, and M and B are the cache size and block (or cache line) size respectively.",
keywords = "Cache oblivious, Fork-Join algorithms, Schedulers",
author = "Richard Cole and Vijaya Ramachandran",
year = "2017",
month = "7",
day = "24",
doi = "10.1145/3087556.3087572",
language = "English (US)",
volume = "Part F129316",
pages = "351--362",
booktitle = "SPAA 2017 - Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Bounding cache miss costs of multithreaded computations under general schedulers

AU - Cole, Richard

AU - Ramachandran, Vijaya

PY - 2017/7/24

Y1 - 2017/7/24

N2 - We analyze the caching overhead incurred by a class of multithreaded algorithms when scheduled by an arbitrary scheduler. We obtain bounds that match or improve upon the well-known O(Q + S · (M/B)) caching cost for the randomized work stealing (RWS) scheduler, where S is the number of steals, Q is the sequential caching cost, and M and B are the cache size and block (or cache line) size respectively.

AB - We analyze the caching overhead incurred by a class of multithreaded algorithms when scheduled by an arbitrary scheduler. We obtain bounds that match or improve upon the well-known O(Q + S · (M/B)) caching cost for the randomized work stealing (RWS) scheduler, where S is the number of steals, Q is the sequential caching cost, and M and B are the cache size and block (or cache line) size respectively.

KW - Cache oblivious

KW - Fork-Join algorithms

KW - Schedulers

UR - http://www.scopus.com/inward/record.url?scp=85027862195&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85027862195&partnerID=8YFLogxK

U2 - 10.1145/3087556.3087572

DO - 10.1145/3087556.3087572

M3 - Conference contribution

AN - SCOPUS:85027862195

VL - Part F129316

SP - 351

EP - 362

BT - SPAA 2017 - Proceedings of the 29th ACM Symposium on Parallelism in Algorithms and Architectures

PB - Association for Computing Machinery

ER -