Standard

A new way to divide and conquer. / Tiskin, Alexandre.

In: Parallel Processing Letters, Vol. 11, No. 4, 01.01.2001, p. 409-422.

Research output: Contribution to journal › Article › peer-review

Harvard

Tiskin, A 2001, 'A new way to divide and conquer', Parallel Processing Letters, vol. 11, no. 4, pp. 409-422. https://doi.org/10.1142/s0129626401000695

APA

Tiskin, A. (2001). A new way to divide and conquer. Parallel Processing Letters, 11(4), 409-422. https://doi.org/10.1142/s0129626401000695

Vancouver

Tiskin A. A new way to divide and conquer. Parallel Processing Letters. 2001 Jan 1;11(4):409-422. https://doi.org/10.1142/s0129626401000695

Author

Tiskin, Alexandre. / A new way to divide and conquer. In: Parallel Processing Letters. 2001 ; Vol. 11, No. 4. pp. 409-422.

BibTeX

@article{c414843b14f84ce6b543d343af03c977,
title = "A new way to divide and conquer",
abstract = "Valiant's model of bulk-synchronous parallel (BSP) computation does not allow the programmer to synchronize a subset, rather than the complete set of a parallel computer's processors. This is perceived by many to be an obstacle to expressing divide-and-conquer algorithms in the BSP model. We argue that the divide-and-conquer paradigm fits naturally into the BSP model, without any need for subset synchronization. The proposed method of divide-and-conquer BSP programming is fully compliant with the BSP computation model. The method is based on sequentially interleaved threads of BSP computation, called superthreads.",
keywords = "BSP, Bulk-synchronous parallel, Divide-and-conquer, Subset synchronization, Superthreads",
author = "Alexandre Tiskin",
year = "2001",
month = jan,
day = "1",
doi = "10.1142/s0129626401000695",
language = "English",
volume = "11",
pages = "409--422",
journal = "Parallel Processing Letters",
issn = "0129-6264",
publisher = "WORLD SCIENTIFIC PUBL CO PTE LTD",
number = "4",

}

RIS

TY - JOUR

T1 - A new way to divide and conquer

AU - Tiskin, Alexandre

PY - 2001/1/1

Y1 - 2001/1/1

N2 - Valiant's model of bulk-synchronous parallel (BSP) computation does not allow the programmer to synchronize a subset, rather than the complete set of a parallel computer's processors. This is perceived by many to be an obstacle to expressing divide-and-conquer algorithms in the BSP model. We argue that the divide-and-conquer paradigm fits naturally into the BSP model, without any need for subset synchronization. The proposed method of divide-and-conquer BSP programming is fully compliant with the BSP computation model. The method is based on sequentially interleaved threads of BSP computation, called superthreads.

AB - Valiant's model of bulk-synchronous parallel (BSP) computation does not allow the programmer to synchronize a subset, rather than the complete set of a parallel computer's processors. This is perceived by many to be an obstacle to expressing divide-and-conquer algorithms in the BSP model. We argue that the divide-and-conquer paradigm fits naturally into the BSP model, without any need for subset synchronization. The proposed method of divide-and-conquer BSP programming is fully compliant with the BSP computation model. The method is based on sequentially interleaved threads of BSP computation, called superthreads.

KW - BSP

KW - Bulk-synchronous parallel

KW - Divide-and-conquer

KW - Subset synchronization

KW - Superthreads

UR - http://www.scopus.com/inward/record.url?scp=0035569737&partnerID=8YFLogxK

U2 - 10.1142/s0129626401000695

DO - 10.1142/s0129626401000695

M3 - Article

AN - SCOPUS:0035569737

VL - 11

SP - 409

EP - 422

JO - Parallel Processing Letters

JF - Parallel Processing Letters

SN - 0129-6264

IS - 4

ER -

ID: 127712960