Extending distributed shared memory for the cell broadband engine to a channel model

Kenneth Skovhede; Morten Nørgaard Larsen; Brian Vinter

doi:10.1007/978-3-642-28151-8_11

Extending distributed shared memory for the cell broadband engine to a channel model

Kenneth Skovhede^*, Morten Nørgaard Larsen, Brian Vinter

^*Corresponding author for this work

Abstract

As the performance gains from traditional processors decline, alternative processor designs are becoming available. One such processor is the CELL-BE processor, which theoretically can deliver a sustained performance close to 205 GFLOPS per processor[10]. Unfortunately, the high performance comes at the price of a quite complex programming model. In this paper we present an easy-to-use, CSP-like, communication method, which enables transfers of shared memory objects. The channel based communication method can significantly reduce the complexity of massively parallel programs. By implementing a few scientific computational cores we show that performance and scalability of the system is acceptable for most problems.

Original language	English
Title of host publication	Applied Parallel and Scientific Computing : 10th International Conference, PARA 2010, Reykjavík, Iceland, June 6-9, 2010, Revised Selected Papers, Part I
Editors	Kristján Jónasson
Number of pages	11
Volume	Part I
Publisher	Springer
Publication date	2012
Pages	108-118
ISBN (Print)	978-3-642-28150-1
ISBN (Electronic)	978-3-642-28151-8
DOIs	https://doi.org/10.1007/978-3-642-28151-8_11
Publication status	Published - 2012
Event	10th International Conference on Applied Parallel and Scientific Computing - Reykjavik, Iceland Duration: 6 Jun 2010 → 9 Jun 2010 Conference number: 10

Conference

Conference	10th International Conference on Applied Parallel and Scientific Computing
Number	10
Country/Territory	Iceland
City	Reykjavik
Period	06/06/2010 → 09/06/2010

Series	Lecture notes in computer science
Volume	7133
ISSN	0302-9743

Keywords

CELL-BE
channel communication
CSP
DSMCBE

Access to Document

10.1007/978-3-642-28151-8_11

Cite this

Skovhede, K., Larsen, M. N., & Vinter, B. (2012). Extending distributed shared memory for the cell broadband engine to a channel model. In K. Jónasson (Ed.), Applied Parallel and Scientific Computing: 10th International Conference, PARA 2010, Reykjavík, Iceland, June 6-9, 2010, Revised Selected Papers, Part I (Vol. Part I, pp. 108-118). Springer. Lecture notes in computer science Vol. 7133 https://doi.org/10.1007/978-3-642-28151-8_11

Extending distributed shared memory for the cell broadband engine to a channel model. / Skovhede, Kenneth; Larsen, Morten Nørgaard; Vinter, Brian.

Applied Parallel and Scientific Computing: 10th International Conference, PARA 2010, Reykjavík, Iceland, June 6-9, 2010, Revised Selected Papers, Part I. ed. / Kristján Jónasson. Vol. Part I Springer, 2012. p. 108-118 (Lecture notes in computer science, Vol. 7133).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Skovhede, K, Larsen, MN & Vinter, B 2012, Extending distributed shared memory for the cell broadband engine to a channel model. in K Jónasson (ed.), Applied Parallel and Scientific Computing: 10th International Conference, PARA 2010, Reykjavík, Iceland, June 6-9, 2010, Revised Selected Papers, Part I. vol. Part I, Springer, Lecture notes in computer science, vol. 7133, pp. 108-118, 10th International Conference on Applied Parallel and Scientific Computing, Reykjavik, Iceland, 06/06/2010. https://doi.org/10.1007/978-3-642-28151-8_11

Skovhede K, Larsen MN, Vinter B. Extending distributed shared memory for the cell broadband engine to a channel model. In Jónasson K, editor, Applied Parallel and Scientific Computing: 10th International Conference, PARA 2010, Reykjavík, Iceland, June 6-9, 2010, Revised Selected Papers, Part I. Vol. Part I. Springer. 2012. p. 108-118. (Lecture notes in computer science, Vol. 7133). doi: 10.1007/978-3-642-28151-8_11

Skovhede, Kenneth ; Larsen, Morten Nørgaard ; Vinter, Brian. / Extending distributed shared memory for the cell broadband engine to a channel model. Applied Parallel and Scientific Computing: 10th International Conference, PARA 2010, Reykjavík, Iceland, June 6-9, 2010, Revised Selected Papers, Part I. editor / Kristján Jónasson. Vol. Part I Springer, 2012. pp. 108-118 (Lecture notes in computer science, Vol. 7133).

@inproceedings{8c4f2dc77c094ea2b10654838f40ed4c,

title = "Extending distributed shared memory for the cell broadband engine to a channel model",

abstract = "As the performance gains from traditional processors decline, alternative processor designs are becoming available. One such processor is the CELL-BE processor, which theoretically can deliver a sustained performance close to 205 GFLOPS per processor[10]. Unfortunately, the high performance comes at the price of a quite complex programming model. In this paper we present an easy-to-use, CSP-like, communication method, which enables transfers of shared memory objects. The channel based communication method can significantly reduce the complexity of massively parallel programs. By implementing a few scientific computational cores we show that performance and scalability of the system is acceptable for most problems.",

keywords = "CELL-BE, channel communication, CSP, DSMCBE",

author = "Kenneth Skovhede and Larsen, {Morten N{\o}rgaard} and Brian Vinter",

year = "2012",

doi = "10.1007/978-3-642-28151-8_11",

language = "English",

isbn = "978-3-642-28150-1",

volume = "Part I",

series = "Lecture notes in computer science",

publisher = "Springer",

pages = "108--118",

editor = "Kristj{\'a}n J{\'o}nasson",

booktitle = "Applied Parallel and Scientific Computing",

note = "10th International Conference on Applied Parallel and Scientific Computing ; Conference date: 06-06-2010 Through 09-06-2010",

}

TY - GEN

T1 - Extending distributed shared memory for the cell broadband engine to a channel model

AU - Skovhede, Kenneth

AU - Larsen, Morten Nørgaard

AU - Vinter, Brian

N1 - Conference code: 10

PY - 2012

Y1 - 2012

N2 - As the performance gains from traditional processors decline, alternative processor designs are becoming available. One such processor is the CELL-BE processor, which theoretically can deliver a sustained performance close to 205 GFLOPS per processor[10]. Unfortunately, the high performance comes at the price of a quite complex programming model. In this paper we present an easy-to-use, CSP-like, communication method, which enables transfers of shared memory objects. The channel based communication method can significantly reduce the complexity of massively parallel programs. By implementing a few scientific computational cores we show that performance and scalability of the system is acceptable for most problems.

AB - As the performance gains from traditional processors decline, alternative processor designs are becoming available. One such processor is the CELL-BE processor, which theoretically can deliver a sustained performance close to 205 GFLOPS per processor[10]. Unfortunately, the high performance comes at the price of a quite complex programming model. In this paper we present an easy-to-use, CSP-like, communication method, which enables transfers of shared memory objects. The channel based communication method can significantly reduce the complexity of massively parallel programs. By implementing a few scientific computational cores we show that performance and scalability of the system is acceptable for most problems.

KW - CELL-BE

KW - channel communication

KW - CSP

KW - DSMCBE

UR - http://www.scopus.com/inward/record.url?scp=84857466990&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-28151-8_11

DO - 10.1007/978-3-642-28151-8_11

M3 - Article in proceedings

AN - SCOPUS:84857466990

SN - 978-3-642-28150-1

VL - Part I

T3 - Lecture notes in computer science

SP - 108

EP - 118

BT - Applied Parallel and Scientific Computing

A2 - Jónasson, Kristján

PB - Springer

T2 - 10th International Conference on Applied Parallel and Scientific Computing

Y2 - 6 June 2010 through 9 June 2010

ER -

Extending distributed shared memory for the cell broadband engine to a channel model

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this