Parallel computing in sound field analysis by bread slicing method

Yasuhiro Tokita, Yasuhiro Oikawa, Yoshio Yamasaki

    Research output: Contribution to journalArticle

    1 Citation (Scopus)

    Abstract

    The use of the bread slicing method to reduce the amount of communication among nodes of a computer cluster on parallel computing of the distributed memory system, was investigated. The computer cluster consisted of 128 calculating nodes were used for the study. The sound wave in a three dimensional sound field was used to determine nodes distance. Boundary conditions was used to calculate sound pressure at grid points. The 256 GB memory space of the cluster was used for dividing the calculating field. The calculation of the sound field was performed using a spatial grid that is effective to calculate field with less memory space. The field was divided by the bread slicing method that requires a shorter communicating time among the processes. The method has the ability to increase the speed and scale of a distributed memory system.

    Original languageEnglish
    Pages (from-to)113-116
    Number of pages4
    JournalAcoustical Science and Technology
    Volume29
    Issue number1
    DOIs
    Publication statusPublished - 2008

    Fingerprint

    slicing
    sound fields
    distributed memory
    grids
    communicating
    sound pressure
    sound waves
    communication
    boundary conditions

    Keywords

    • Communication
    • Distributed memory system
    • MPI
    • Parallel computing
    • Three dimensional sound fields

    ASJC Scopus subject areas

    • Acoustics and Ultrasonics

    Cite this

    Parallel computing in sound field analysis by bread slicing method. / Tokita, Yasuhiro; Oikawa, Yasuhiro; Yamasaki, Yoshio.

    In: Acoustical Science and Technology, Vol. 29, No. 1, 2008, p. 113-116.

    Research output: Contribution to journalArticle

    Tokita, Yasuhiro ; Oikawa, Yasuhiro ; Yamasaki, Yoshio. / Parallel computing in sound field analysis by bread slicing method. In: Acoustical Science and Technology. 2008 ; Vol. 29, No. 1. pp. 113-116.
    @article{e1bdd6611dde4cceada9d16c5c61d483,
    title = "Parallel computing in sound field analysis by bread slicing method",
    abstract = "The use of the bread slicing method to reduce the amount of communication among nodes of a computer cluster on parallel computing of the distributed memory system, was investigated. The computer cluster consisted of 128 calculating nodes were used for the study. The sound wave in a three dimensional sound field was used to determine nodes distance. Boundary conditions was used to calculate sound pressure at grid points. The 256 GB memory space of the cluster was used for dividing the calculating field. The calculation of the sound field was performed using a spatial grid that is effective to calculate field with less memory space. The field was divided by the bread slicing method that requires a shorter communicating time among the processes. The method has the ability to increase the speed and scale of a distributed memory system.",
    keywords = "Communication, Distributed memory system, MPI, Parallel computing, Three dimensional sound fields",
    author = "Yasuhiro Tokita and Yasuhiro Oikawa and Yoshio Yamasaki",
    year = "2008",
    doi = "10.1250/ast.29.113",
    language = "English",
    volume = "29",
    pages = "113--116",
    journal = "Acoustical Science and Technology",
    issn = "1346-3969",
    publisher = "Acoustical Society of Japan",
    number = "1",

    }

    TY - JOUR

    T1 - Parallel computing in sound field analysis by bread slicing method

    AU - Tokita, Yasuhiro

    AU - Oikawa, Yasuhiro

    AU - Yamasaki, Yoshio

    PY - 2008

    Y1 - 2008

    N2 - The use of the bread slicing method to reduce the amount of communication among nodes of a computer cluster on parallel computing of the distributed memory system, was investigated. The computer cluster consisted of 128 calculating nodes were used for the study. The sound wave in a three dimensional sound field was used to determine nodes distance. Boundary conditions was used to calculate sound pressure at grid points. The 256 GB memory space of the cluster was used for dividing the calculating field. The calculation of the sound field was performed using a spatial grid that is effective to calculate field with less memory space. The field was divided by the bread slicing method that requires a shorter communicating time among the processes. The method has the ability to increase the speed and scale of a distributed memory system.

    AB - The use of the bread slicing method to reduce the amount of communication among nodes of a computer cluster on parallel computing of the distributed memory system, was investigated. The computer cluster consisted of 128 calculating nodes were used for the study. The sound wave in a three dimensional sound field was used to determine nodes distance. Boundary conditions was used to calculate sound pressure at grid points. The 256 GB memory space of the cluster was used for dividing the calculating field. The calculation of the sound field was performed using a spatial grid that is effective to calculate field with less memory space. The field was divided by the bread slicing method that requires a shorter communicating time among the processes. The method has the ability to increase the speed and scale of a distributed memory system.

    KW - Communication

    KW - Distributed memory system

    KW - MPI

    KW - Parallel computing

    KW - Three dimensional sound fields

    UR - http://www.scopus.com/inward/record.url?scp=37749041682&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=37749041682&partnerID=8YFLogxK

    U2 - 10.1250/ast.29.113

    DO - 10.1250/ast.29.113

    M3 - Article

    AN - SCOPUS:37749041682

    VL - 29

    SP - 113

    EP - 116

    JO - Acoustical Science and Technology

    JF - Acoustical Science and Technology

    SN - 1346-3969

    IS - 1

    ER -