Performance evaluation of macrodataflow computation on shared memory multiprocessors

Kento Aida, Kiyoshi Iwasaki, Hironori Kasahara, Seinosuke Narita

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

    3 Citations (Scopus)

    Abstract

    The coarse grain parallel processing on shared memory multiprocessor systems has been implemented using multi-tasking. However, this scheme has drawbacks such as difficulty in the extraction of parallelism among coarse grain tasks by ordinary users and large dynamic scheduling overhead caused by operating system calls or run-time library calls. On the other hand, in the proposed Fortran macro-dataflow computation scheme, the compiler automatically generates coarse grain tasks called macrotasks, exploits parallelism among macrotasks, and generates a dynamic scheduling routine that schedules macrotasks to processors at run-time with small overhead. This paper presents performance evaluation of macrodataflow computation on shared memory multiprocessor systems. The results on four processors of a KSR1 show that macrodataflow computation reduces execution time to 1/2.81 of sequential execution time while an ordinary multi-threading reduces execution time to 1/2.19 of sequential execution time.

    Original languageEnglish
    Title of host publicationIEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings
    Place of PublicationPiscataway, NJ, United States
    PublisherIEEE
    Pages50-54
    Number of pages5
    Publication statusPublished - 1995
    EventProceedings of the 1995 IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Victoria, BC, Can
    Duration: 1995 May 171995 May 19

    Other

    OtherProceedings of the 1995 IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing
    CityVictoria, BC, Can
    Period95/5/1795/5/19

    Fingerprint

    Data storage equipment
    Scheduling
    Multitasking
    Macros
    Processing

    ASJC Scopus subject areas

    • Signal Processing

    Cite this

    Aida, K., Iwasaki, K., Kasahara, H., & Narita, S. (1995). Performance evaluation of macrodataflow computation on shared memory multiprocessors. In IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings (pp. 50-54). Piscataway, NJ, United States: IEEE.

    Performance evaluation of macrodataflow computation on shared memory multiprocessors. / Aida, Kento; Iwasaki, Kiyoshi; Kasahara, Hironori; Narita, Seinosuke.

    IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings. Piscataway, NJ, United States : IEEE, 1995. p. 50-54.

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

    Aida, K, Iwasaki, K, Kasahara, H & Narita, S 1995, Performance evaluation of macrodataflow computation on shared memory multiprocessors. in IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings. IEEE, Piscataway, NJ, United States, pp. 50-54, Proceedings of the 1995 IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing, Victoria, BC, Can, 95/5/17.
    Aida K, Iwasaki K, Kasahara H, Narita S. Performance evaluation of macrodataflow computation on shared memory multiprocessors. In IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings. Piscataway, NJ, United States: IEEE. 1995. p. 50-54
    Aida, Kento ; Iwasaki, Kiyoshi ; Kasahara, Hironori ; Narita, Seinosuke. / Performance evaluation of macrodataflow computation on shared memory multiprocessors. IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings. Piscataway, NJ, United States : IEEE, 1995. pp. 50-54
    @inproceedings{d133b1d5f2da4f7591cc45705f4a2317,
    title = "Performance evaluation of macrodataflow computation on shared memory multiprocessors",
    abstract = "The coarse grain parallel processing on shared memory multiprocessor systems has been implemented using multi-tasking. However, this scheme has drawbacks such as difficulty in the extraction of parallelism among coarse grain tasks by ordinary users and large dynamic scheduling overhead caused by operating system calls or run-time library calls. On the other hand, in the proposed Fortran macro-dataflow computation scheme, the compiler automatically generates coarse grain tasks called macrotasks, exploits parallelism among macrotasks, and generates a dynamic scheduling routine that schedules macrotasks to processors at run-time with small overhead. This paper presents performance evaluation of macrodataflow computation on shared memory multiprocessor systems. The results on four processors of a KSR1 show that macrodataflow computation reduces execution time to 1/2.81 of sequential execution time while an ordinary multi-threading reduces execution time to 1/2.19 of sequential execution time.",
    author = "Kento Aida and Kiyoshi Iwasaki and Hironori Kasahara and Seinosuke Narita",
    year = "1995",
    language = "English",
    pages = "50--54",
    booktitle = "IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings",
    publisher = "IEEE",

    }

    TY - GEN

    T1 - Performance evaluation of macrodataflow computation on shared memory multiprocessors

    AU - Aida, Kento

    AU - Iwasaki, Kiyoshi

    AU - Kasahara, Hironori

    AU - Narita, Seinosuke

    PY - 1995

    Y1 - 1995

    N2 - The coarse grain parallel processing on shared memory multiprocessor systems has been implemented using multi-tasking. However, this scheme has drawbacks such as difficulty in the extraction of parallelism among coarse grain tasks by ordinary users and large dynamic scheduling overhead caused by operating system calls or run-time library calls. On the other hand, in the proposed Fortran macro-dataflow computation scheme, the compiler automatically generates coarse grain tasks called macrotasks, exploits parallelism among macrotasks, and generates a dynamic scheduling routine that schedules macrotasks to processors at run-time with small overhead. This paper presents performance evaluation of macrodataflow computation on shared memory multiprocessor systems. The results on four processors of a KSR1 show that macrodataflow computation reduces execution time to 1/2.81 of sequential execution time while an ordinary multi-threading reduces execution time to 1/2.19 of sequential execution time.

    AB - The coarse grain parallel processing on shared memory multiprocessor systems has been implemented using multi-tasking. However, this scheme has drawbacks such as difficulty in the extraction of parallelism among coarse grain tasks by ordinary users and large dynamic scheduling overhead caused by operating system calls or run-time library calls. On the other hand, in the proposed Fortran macro-dataflow computation scheme, the compiler automatically generates coarse grain tasks called macrotasks, exploits parallelism among macrotasks, and generates a dynamic scheduling routine that schedules macrotasks to processors at run-time with small overhead. This paper presents performance evaluation of macrodataflow computation on shared memory multiprocessor systems. The results on four processors of a KSR1 show that macrodataflow computation reduces execution time to 1/2.81 of sequential execution time while an ordinary multi-threading reduces execution time to 1/2.19 of sequential execution time.

    UR - http://www.scopus.com/inward/record.url?scp=0029223868&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=0029223868&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:0029223868

    SP - 50

    EP - 54

    BT - IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings

    PB - IEEE

    CY - Piscataway, NJ, United States

    ER -