Preceding activation scheme with graph unfolding for the parallel processing system 'Harray'

Hayato Yamana, Takashi Hagiwara, Jun Kohdate, Yoichi Muraoka

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

    1 Citation (Scopus)

    Abstract

    The purpose of this work is to propose and evaluate the preceding activation scheme with graph unfolding, which translate a Fortran program into a dataflow graph and executes it efficiently. The problems in restructuring a Fortran program into a dataflow graph are that a Fortran program is not written in a single assignment rule and it has an explicit control flow. These problems result in little parallelism because many gate operations, such as T/F gates, are introduced in the dataflow graph to synchronize the data movement. Therefore, discarding these gate operations is the key to exposing parallelism in a Fortran program. The preceding activation scheme with graph unfolding is proposed to discard these gate operations. The result of the performance evaluation by the 'Harray' software simulator is presented. It is shown that the execution speed with the proposed scheme for flow graphs without backward branches is about 1.5 times as fast as that with the extended activation scheme which initiates the execution only after it is confirmed that a basic block will be selected at a conditional branch. Moreover, the execution speed is 2.7 times as fast as that with the extended activation scheme if a flow graph including backward branches is unfolded by the proposed scheme.

    Original languageEnglish
    Title of host publicationProc Supercomput 89
    Editors Anon
    Place of PublicationPiscataway, NJ, United States
    PublisherPubl by IEEE
    Pages675-684
    Number of pages10
    Publication statusPublished - 1989
    EventProceedings: Supercomputing '89 - Reno, NV, USA
    Duration: 1989 Nov 131989 Nov 17

    Other

    OtherProceedings: Supercomputing '89
    CityReno, NV, USA
    Period89/11/1389/11/17

    Fingerprint

    Parallel processing systems
    Chemical activation
    Flow graphs
    Flow control
    Simulators

    ASJC Scopus subject areas

    • Engineering(all)

    Cite this

    Yamana, H., Hagiwara, T., Kohdate, J., & Muraoka, Y. (1989). Preceding activation scheme with graph unfolding for the parallel processing system 'Harray'. In Anon (Ed.), Proc Supercomput 89 (pp. 675-684). Piscataway, NJ, United States: Publ by IEEE.

    Preceding activation scheme with graph unfolding for the parallel processing system 'Harray'. / Yamana, Hayato; Hagiwara, Takashi; Kohdate, Jun; Muraoka, Yoichi.

    Proc Supercomput 89. ed. / Anon. Piscataway, NJ, United States : Publ by IEEE, 1989. p. 675-684.

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

    Yamana, H, Hagiwara, T, Kohdate, J & Muraoka, Y 1989, Preceding activation scheme with graph unfolding for the parallel processing system 'Harray'. in Anon (ed.), Proc Supercomput 89. Publ by IEEE, Piscataway, NJ, United States, pp. 675-684, Proceedings: Supercomputing '89, Reno, NV, USA, 89/11/13.
    Yamana H, Hagiwara T, Kohdate J, Muraoka Y. Preceding activation scheme with graph unfolding for the parallel processing system 'Harray'. In Anon, editor, Proc Supercomput 89. Piscataway, NJ, United States: Publ by IEEE. 1989. p. 675-684
    Yamana, Hayato ; Hagiwara, Takashi ; Kohdate, Jun ; Muraoka, Yoichi. / Preceding activation scheme with graph unfolding for the parallel processing system 'Harray'. Proc Supercomput 89. editor / Anon. Piscataway, NJ, United States : Publ by IEEE, 1989. pp. 675-684
    @inproceedings{d28a8605bc3a4a9bb196d313e4537258,
    title = "Preceding activation scheme with graph unfolding for the parallel processing system 'Harray'",
    abstract = "The purpose of this work is to propose and evaluate the preceding activation scheme with graph unfolding, which translate a Fortran program into a dataflow graph and executes it efficiently. The problems in restructuring a Fortran program into a dataflow graph are that a Fortran program is not written in a single assignment rule and it has an explicit control flow. These problems result in little parallelism because many gate operations, such as T/F gates, are introduced in the dataflow graph to synchronize the data movement. Therefore, discarding these gate operations is the key to exposing parallelism in a Fortran program. The preceding activation scheme with graph unfolding is proposed to discard these gate operations. The result of the performance evaluation by the 'Harray' software simulator is presented. It is shown that the execution speed with the proposed scheme for flow graphs without backward branches is about 1.5 times as fast as that with the extended activation scheme which initiates the execution only after it is confirmed that a basic block will be selected at a conditional branch. Moreover, the execution speed is 2.7 times as fast as that with the extended activation scheme if a flow graph including backward branches is unfolded by the proposed scheme.",
    author = "Hayato Yamana and Takashi Hagiwara and Jun Kohdate and Yoichi Muraoka",
    year = "1989",
    language = "English",
    pages = "675--684",
    editor = "Anon",
    booktitle = "Proc Supercomput 89",
    publisher = "Publ by IEEE",

    }

    TY - GEN

    T1 - Preceding activation scheme with graph unfolding for the parallel processing system 'Harray'

    AU - Yamana, Hayato

    AU - Hagiwara, Takashi

    AU - Kohdate, Jun

    AU - Muraoka, Yoichi

    PY - 1989

    Y1 - 1989

    N2 - The purpose of this work is to propose and evaluate the preceding activation scheme with graph unfolding, which translate a Fortran program into a dataflow graph and executes it efficiently. The problems in restructuring a Fortran program into a dataflow graph are that a Fortran program is not written in a single assignment rule and it has an explicit control flow. These problems result in little parallelism because many gate operations, such as T/F gates, are introduced in the dataflow graph to synchronize the data movement. Therefore, discarding these gate operations is the key to exposing parallelism in a Fortran program. The preceding activation scheme with graph unfolding is proposed to discard these gate operations. The result of the performance evaluation by the 'Harray' software simulator is presented. It is shown that the execution speed with the proposed scheme for flow graphs without backward branches is about 1.5 times as fast as that with the extended activation scheme which initiates the execution only after it is confirmed that a basic block will be selected at a conditional branch. Moreover, the execution speed is 2.7 times as fast as that with the extended activation scheme if a flow graph including backward branches is unfolded by the proposed scheme.

    AB - The purpose of this work is to propose and evaluate the preceding activation scheme with graph unfolding, which translate a Fortran program into a dataflow graph and executes it efficiently. The problems in restructuring a Fortran program into a dataflow graph are that a Fortran program is not written in a single assignment rule and it has an explicit control flow. These problems result in little parallelism because many gate operations, such as T/F gates, are introduced in the dataflow graph to synchronize the data movement. Therefore, discarding these gate operations is the key to exposing parallelism in a Fortran program. The preceding activation scheme with graph unfolding is proposed to discard these gate operations. The result of the performance evaluation by the 'Harray' software simulator is presented. It is shown that the execution speed with the proposed scheme for flow graphs without backward branches is about 1.5 times as fast as that with the extended activation scheme which initiates the execution only after it is confirmed that a basic block will be selected at a conditional branch. Moreover, the execution speed is 2.7 times as fast as that with the extended activation scheme if a flow graph including backward branches is unfolded by the proposed scheme.

    UR - http://www.scopus.com/inward/record.url?scp=0024935631&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=0024935631&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:0024935631

    SP - 675

    EP - 684

    BT - Proc Supercomput 89

    A2 - Anon, null

    PB - Publ by IEEE

    CY - Piscataway, NJ, United States

    ER -