Multigrain automatic parallelization in Japanese Millennium Project IT21 Advanced Parallelizing Compiler

Hironori Kasahara, M. Obata, K. Ishizaka, Keiji Kimura, H. Kaminaga, H. Nakano, K. Nagasawa, A. Murai, H. Itagaki, J. Shirako

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

    7 Citations (Scopus)

    Abstract

    This paper describes OSCAR multigrain parallelizing compiler which has been developed in Japanese Millennium Project IT21 "Advanced Parallelizing Compiler" project and its performance on SMP machines. The compiler realizes multigrain parallelization for chip-multiprocessors to high-end servers. It hierarchically exploits coarse grain task parallelism among loops, subroutines and basic blocks and near fine grain parallelism among statements inside a basic block in addition to loop parallelism. Also, it globally optimizes cache use over different loops, or coarse grain tasks, based on data localization technique to reduce memory access overhead. Current performance of OSCAR compiler for SPEC95fp is evaluated on different SMPs. For example, it gives us 3.7 times speedup for HYDRO2D, 1.8 times for SWIM, 1.7 times for SU2COR, 2.0 times for MGRID, 3.3 times for TURB3D on 8 processor IBM RS6000, against XL Fortran compiler ver.7.1 and 4.2 times speedup for SWIM and 2.2 times speedup for TURB3D on 4 processor Sun Ultra80 workstation against Forte6 update 2.

    Original languageEnglish
    Title of host publicationProceedings - International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002
    PublisherInstitute of Electrical and Electronics Engineers Inc.
    Pages105-111
    Number of pages7
    ISBN (Print)0769517307, 9780769517308
    DOIs
    Publication statusPublished - 2002
    EventInternational Conference on Parallel Computing in Electrical Engineering, PARELEC 2002 - Warsaw, Poland
    Duration: 2002 Sep 222002 Sep 25

    Other

    OtherInternational Conference on Parallel Computing in Electrical Engineering, PARELEC 2002
    CountryPoland
    CityWarsaw
    Period02/9/2202/9/25

    Fingerprint

    Subroutines
    Sun
    Servers
    Data storage equipment

    Keywords

    • Algorithms
    • Data analysis
    • Government
    • Memory architecture
    • Multiprocessing systems
    • Parallel processing
    • Program processors
    • Testing
    • Usability
    • Workstations

    ASJC Scopus subject areas

    • Computer Science Applications
    • Electrical and Electronic Engineering

    Cite this

    Kasahara, H., Obata, M., Ishizaka, K., Kimura, K., Kaminaga, H., Nakano, H., ... Shirako, J. (2002). Multigrain automatic parallelization in Japanese Millennium Project IT21 Advanced Parallelizing Compiler. In Proceedings - International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002 (pp. 105-111). [1115213] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/PCEE.2002.1115213

    Multigrain automatic parallelization in Japanese Millennium Project IT21 Advanced Parallelizing Compiler. / Kasahara, Hironori; Obata, M.; Ishizaka, K.; Kimura, Keiji; Kaminaga, H.; Nakano, H.; Nagasawa, K.; Murai, A.; Itagaki, H.; Shirako, J.

    Proceedings - International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002. Institute of Electrical and Electronics Engineers Inc., 2002. p. 105-111 1115213.

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

    Kasahara, H, Obata, M, Ishizaka, K, Kimura, K, Kaminaga, H, Nakano, H, Nagasawa, K, Murai, A, Itagaki, H & Shirako, J 2002, Multigrain automatic parallelization in Japanese Millennium Project IT21 Advanced Parallelizing Compiler. in Proceedings - International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002., 1115213, Institute of Electrical and Electronics Engineers Inc., pp. 105-111, International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002, Warsaw, Poland, 02/9/22. https://doi.org/10.1109/PCEE.2002.1115213
    Kasahara H, Obata M, Ishizaka K, Kimura K, Kaminaga H, Nakano H et al. Multigrain automatic parallelization in Japanese Millennium Project IT21 Advanced Parallelizing Compiler. In Proceedings - International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002. Institute of Electrical and Electronics Engineers Inc. 2002. p. 105-111. 1115213 https://doi.org/10.1109/PCEE.2002.1115213
    Kasahara, Hironori ; Obata, M. ; Ishizaka, K. ; Kimura, Keiji ; Kaminaga, H. ; Nakano, H. ; Nagasawa, K. ; Murai, A. ; Itagaki, H. ; Shirako, J. / Multigrain automatic parallelization in Japanese Millennium Project IT21 Advanced Parallelizing Compiler. Proceedings - International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002. Institute of Electrical and Electronics Engineers Inc., 2002. pp. 105-111
    @inproceedings{9fa236d31c2a4e7aaeff526e19d43701,
    title = "Multigrain automatic parallelization in Japanese Millennium Project IT21 Advanced Parallelizing Compiler",
    abstract = "This paper describes OSCAR multigrain parallelizing compiler which has been developed in Japanese Millennium Project IT21 {"}Advanced Parallelizing Compiler{"} project and its performance on SMP machines. The compiler realizes multigrain parallelization for chip-multiprocessors to high-end servers. It hierarchically exploits coarse grain task parallelism among loops, subroutines and basic blocks and near fine grain parallelism among statements inside a basic block in addition to loop parallelism. Also, it globally optimizes cache use over different loops, or coarse grain tasks, based on data localization technique to reduce memory access overhead. Current performance of OSCAR compiler for SPEC95fp is evaluated on different SMPs. For example, it gives us 3.7 times speedup for HYDRO2D, 1.8 times for SWIM, 1.7 times for SU2COR, 2.0 times for MGRID, 3.3 times for TURB3D on 8 processor IBM RS6000, against XL Fortran compiler ver.7.1 and 4.2 times speedup for SWIM and 2.2 times speedup for TURB3D on 4 processor Sun Ultra80 workstation against Forte6 update 2.",
    keywords = "Algorithms, Data analysis, Government, Memory architecture, Multiprocessing systems, Parallel processing, Program processors, Testing, Usability, Workstations",
    author = "Hironori Kasahara and M. Obata and K. Ishizaka and Keiji Kimura and H. Kaminaga and H. Nakano and K. Nagasawa and A. Murai and H. Itagaki and J. Shirako",
    year = "2002",
    doi = "10.1109/PCEE.2002.1115213",
    language = "English",
    isbn = "0769517307",
    pages = "105--111",
    booktitle = "Proceedings - International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002",
    publisher = "Institute of Electrical and Electronics Engineers Inc.",

    }

    TY - GEN

    T1 - Multigrain automatic parallelization in Japanese Millennium Project IT21 Advanced Parallelizing Compiler

    AU - Kasahara, Hironori

    AU - Obata, M.

    AU - Ishizaka, K.

    AU - Kimura, Keiji

    AU - Kaminaga, H.

    AU - Nakano, H.

    AU - Nagasawa, K.

    AU - Murai, A.

    AU - Itagaki, H.

    AU - Shirako, J.

    PY - 2002

    Y1 - 2002

    N2 - This paper describes OSCAR multigrain parallelizing compiler which has been developed in Japanese Millennium Project IT21 "Advanced Parallelizing Compiler" project and its performance on SMP machines. The compiler realizes multigrain parallelization for chip-multiprocessors to high-end servers. It hierarchically exploits coarse grain task parallelism among loops, subroutines and basic blocks and near fine grain parallelism among statements inside a basic block in addition to loop parallelism. Also, it globally optimizes cache use over different loops, or coarse grain tasks, based on data localization technique to reduce memory access overhead. Current performance of OSCAR compiler for SPEC95fp is evaluated on different SMPs. For example, it gives us 3.7 times speedup for HYDRO2D, 1.8 times for SWIM, 1.7 times for SU2COR, 2.0 times for MGRID, 3.3 times for TURB3D on 8 processor IBM RS6000, against XL Fortran compiler ver.7.1 and 4.2 times speedup for SWIM and 2.2 times speedup for TURB3D on 4 processor Sun Ultra80 workstation against Forte6 update 2.

    AB - This paper describes OSCAR multigrain parallelizing compiler which has been developed in Japanese Millennium Project IT21 "Advanced Parallelizing Compiler" project and its performance on SMP machines. The compiler realizes multigrain parallelization for chip-multiprocessors to high-end servers. It hierarchically exploits coarse grain task parallelism among loops, subroutines and basic blocks and near fine grain parallelism among statements inside a basic block in addition to loop parallelism. Also, it globally optimizes cache use over different loops, or coarse grain tasks, based on data localization technique to reduce memory access overhead. Current performance of OSCAR compiler for SPEC95fp is evaluated on different SMPs. For example, it gives us 3.7 times speedup for HYDRO2D, 1.8 times for SWIM, 1.7 times for SU2COR, 2.0 times for MGRID, 3.3 times for TURB3D on 8 processor IBM RS6000, against XL Fortran compiler ver.7.1 and 4.2 times speedup for SWIM and 2.2 times speedup for TURB3D on 4 processor Sun Ultra80 workstation against Forte6 update 2.

    KW - Algorithms

    KW - Data analysis

    KW - Government

    KW - Memory architecture

    KW - Multiprocessing systems

    KW - Parallel processing

    KW - Program processors

    KW - Testing

    KW - Usability

    KW - Workstations

    UR - http://www.scopus.com/inward/record.url?scp=44649120286&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=44649120286&partnerID=8YFLogxK

    U2 - 10.1109/PCEE.2002.1115213

    DO - 10.1109/PCEE.2002.1115213

    M3 - Conference contribution

    AN - SCOPUS:44649120286

    SN - 0769517307

    SN - 9780769517308

    SP - 105

    EP - 111

    BT - Proceedings - International Conference on Parallel Computing in Electrical Engineering, PARELEC 2002

    PB - Institute of Electrical and Electronics Engineers Inc.

    ER -