A proposal of thread virtualization environment for cell broadband engine

Masahiro Yamada, Yuri Nishikawa, Masato Yoshimi, Hideharu Amano

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

2 Citations (Scopus)

Abstract

Effective parallel programming for a PC cluster integrating multi-core processors requires programmers two types of skill that have different nature; that is multi-thread programming to use multiple cores, and programming for inter-node communication by using libraries such as MPI. For the ease of programming development, we propose a Thread Virtualization Environment (TVE) which virtualizes multiple cores in multiple nodes connected by a network as if they are in one node. If we use this environment, we only need the knowledge of multi-thread programming techniques to effectively utilize the computing resources in multiple nodes. As long inter-node communication delay can severely downgrade performance in some applications, we implemented a caching mechanism on each node so that the number of inter-node communication can be reduced. As the result of executing Monte-Carlo method, whose algorithm requires few data transfers, on TVE, we confirmed that performance scaled well as number of cores increased. On the other hand, Levenshtein Distance computation with frequent data transfers, performance using 30 cores was 0.029 times compared to that of using 6 cores. However, by adopting cache mechanism, inter-node data transfer time was shortened to 5% using the same program.

Original languageEnglish
Title of host publicationProceedings of the IASTED International Conference on Parallel and Distributed Computing and Systems
Pages32-39
Number of pages8
DOIs
Publication statusPublished - 2010
EventIASTED International Conference on Parallel and Distributed Computing and Systems, PDCS 2010 - Marina Del Rey, CA, United States
Duration: 2010 Nov 82010 Nov 10

Other

OtherIASTED International Conference on Parallel and Distributed Computing and Systems, PDCS 2010
CountryUnited States
CityMarina Del Rey, CA
Period10/11/810/11/10

Fingerprint

Data transfer
Engines
Communication
Parallel programming
Computer programming
Monte Carlo methods
Virtualization

Keywords

  • Cell/B.E.
  • Multicore programming
  • Parallel programming
  • PC cluster
  • Playstation3

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

Yamada, M., Nishikawa, Y., Yoshimi, M., & Amano, H. (2010). A proposal of thread virtualization environment for cell broadband engine. In Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Systems (pp. 32-39) https://doi.org/10.2316/P.2010.724-027

A proposal of thread virtualization environment for cell broadband engine. / Yamada, Masahiro; Nishikawa, Yuri; Yoshimi, Masato; Amano, Hideharu.

Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Systems. 2010. p. 32-39.

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

Yamada, M, Nishikawa, Y, Yoshimi, M & Amano, H 2010, A proposal of thread virtualization environment for cell broadband engine. in Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Systems. pp. 32-39, IASTED International Conference on Parallel and Distributed Computing and Systems, PDCS 2010, Marina Del Rey, CA, United States, 10/11/8. https://doi.org/10.2316/P.2010.724-027
Yamada M, Nishikawa Y, Yoshimi M, Amano H. A proposal of thread virtualization environment for cell broadband engine. In Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Systems. 2010. p. 32-39 https://doi.org/10.2316/P.2010.724-027
Yamada, Masahiro ; Nishikawa, Yuri ; Yoshimi, Masato ; Amano, Hideharu. / A proposal of thread virtualization environment for cell broadband engine. Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Systems. 2010. pp. 32-39
@inproceedings{53e14f0324ba4ee59e095d46e11c1f3d,
title = "A proposal of thread virtualization environment for cell broadband engine",
abstract = "Effective parallel programming for a PC cluster integrating multi-core processors requires programmers two types of skill that have different nature; that is multi-thread programming to use multiple cores, and programming for inter-node communication by using libraries such as MPI. For the ease of programming development, we propose a Thread Virtualization Environment (TVE) which virtualizes multiple cores in multiple nodes connected by a network as if they are in one node. If we use this environment, we only need the knowledge of multi-thread programming techniques to effectively utilize the computing resources in multiple nodes. As long inter-node communication delay can severely downgrade performance in some applications, we implemented a caching mechanism on each node so that the number of inter-node communication can be reduced. As the result of executing Monte-Carlo method, whose algorithm requires few data transfers, on TVE, we confirmed that performance scaled well as number of cores increased. On the other hand, Levenshtein Distance computation with frequent data transfers, performance using 30 cores was 0.029 times compared to that of using 6 cores. However, by adopting cache mechanism, inter-node data transfer time was shortened to 5{\%} using the same program.",
keywords = "Cell/B.E., Multicore programming, Parallel programming, PC cluster, Playstation3",
author = "Masahiro Yamada and Yuri Nishikawa and Masato Yoshimi and Hideharu Amano",
year = "2010",
doi = "10.2316/P.2010.724-027",
language = "English",
isbn = "9780889868786",
pages = "32--39",
booktitle = "Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Systems",

}

TY - GEN

T1 - A proposal of thread virtualization environment for cell broadband engine

AU - Yamada, Masahiro

AU - Nishikawa, Yuri

AU - Yoshimi, Masato

AU - Amano, Hideharu

PY - 2010

Y1 - 2010

N2 - Effective parallel programming for a PC cluster integrating multi-core processors requires programmers two types of skill that have different nature; that is multi-thread programming to use multiple cores, and programming for inter-node communication by using libraries such as MPI. For the ease of programming development, we propose a Thread Virtualization Environment (TVE) which virtualizes multiple cores in multiple nodes connected by a network as if they are in one node. If we use this environment, we only need the knowledge of multi-thread programming techniques to effectively utilize the computing resources in multiple nodes. As long inter-node communication delay can severely downgrade performance in some applications, we implemented a caching mechanism on each node so that the number of inter-node communication can be reduced. As the result of executing Monte-Carlo method, whose algorithm requires few data transfers, on TVE, we confirmed that performance scaled well as number of cores increased. On the other hand, Levenshtein Distance computation with frequent data transfers, performance using 30 cores was 0.029 times compared to that of using 6 cores. However, by adopting cache mechanism, inter-node data transfer time was shortened to 5% using the same program.

AB - Effective parallel programming for a PC cluster integrating multi-core processors requires programmers two types of skill that have different nature; that is multi-thread programming to use multiple cores, and programming for inter-node communication by using libraries such as MPI. For the ease of programming development, we propose a Thread Virtualization Environment (TVE) which virtualizes multiple cores in multiple nodes connected by a network as if they are in one node. If we use this environment, we only need the knowledge of multi-thread programming techniques to effectively utilize the computing resources in multiple nodes. As long inter-node communication delay can severely downgrade performance in some applications, we implemented a caching mechanism on each node so that the number of inter-node communication can be reduced. As the result of executing Monte-Carlo method, whose algorithm requires few data transfers, on TVE, we confirmed that performance scaled well as number of cores increased. On the other hand, Levenshtein Distance computation with frequent data transfers, performance using 30 cores was 0.029 times compared to that of using 6 cores. However, by adopting cache mechanism, inter-node data transfer time was shortened to 5% using the same program.

KW - Cell/B.E.

KW - Multicore programming

KW - Parallel programming

KW - PC cluster

KW - Playstation3

UR - http://www.scopus.com/inward/record.url?scp=84858645572&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84858645572&partnerID=8YFLogxK

U2 - 10.2316/P.2010.724-027

DO - 10.2316/P.2010.724-027

M3 - Conference contribution

SN - 9780889868786

SP - 32

EP - 39

BT - Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Systems

ER -