Static task cluster size determination in homogeneous distributed systems

Hidehiro Kanemitsu, Gilhyon Lee, Hidenori Nakazato, Takashige Hoshiai, Yoshiyori Urano

    Research output: Chapter in Book/Report/Conference proceedingChapter

    2 Citations (Scopus)

    Abstract

    In a distributed system, which consists of an unknown number of processors, it is important to derive an appropriate number of processors by which the good schedule length is obtained by a task scheduling. Many task clustering heuristics have been proposed to determine the number of processors and to minimize the schedule length for scheduling a directed acyclic graph (DAG) application. However, those heuristics are not aware of the actual number of existing processors. As a result, the number of processors determined by an existing task clustering may exceed that of actually existing processors. Therefore, conventional approaches adopt merging of each cluster for reducing the number of clusters at the expense of decreasing degree of task parallelism. In this paper, we present a static cluster size determination method, which derives the lower bound of the cluster size with considering the DAG structure and the task size to data size ratio to suppress the schedule length with the small number of processors. Our experimental evaluations by simulations show that the lower bound of each cluster size determined by the proposed method has a good impact on both the schedule length and the processor utilization.

    Original languageEnglish
    Title of host publicationSoftware Automatic Tuning: From Concepts to State-of-the-Art Results
    PublisherSpringer New York
    Pages229-252
    Number of pages24
    ISBN (Print)9781441969347
    DOIs
    Publication statusPublished - 2010

    Fingerprint

    Size determination
    Scheduling
    Merging

    Keywords

    • Cluster size
    • DAG
    • Task clustering
    • Task scheduling

    ASJC Scopus subject areas

    • Engineering(all)

    Cite this

    Kanemitsu, H., Lee, G., Nakazato, H., Hoshiai, T., & Urano, Y. (2010). Static task cluster size determination in homogeneous distributed systems. In Software Automatic Tuning: From Concepts to State-of-the-Art Results (pp. 229-252). Springer New York. https://doi.org/10.1007/978-1-4419-6935-4_14

    Static task cluster size determination in homogeneous distributed systems. / Kanemitsu, Hidehiro; Lee, Gilhyon; Nakazato, Hidenori; Hoshiai, Takashige; Urano, Yoshiyori.

    Software Automatic Tuning: From Concepts to State-of-the-Art Results. Springer New York, 2010. p. 229-252.

    Research output: Chapter in Book/Report/Conference proceedingChapter

    Kanemitsu, H, Lee, G, Nakazato, H, Hoshiai, T & Urano, Y 2010, Static task cluster size determination in homogeneous distributed systems. in Software Automatic Tuning: From Concepts to State-of-the-Art Results. Springer New York, pp. 229-252. https://doi.org/10.1007/978-1-4419-6935-4_14
    Kanemitsu H, Lee G, Nakazato H, Hoshiai T, Urano Y. Static task cluster size determination in homogeneous distributed systems. In Software Automatic Tuning: From Concepts to State-of-the-Art Results. Springer New York. 2010. p. 229-252 https://doi.org/10.1007/978-1-4419-6935-4_14
    Kanemitsu, Hidehiro ; Lee, Gilhyon ; Nakazato, Hidenori ; Hoshiai, Takashige ; Urano, Yoshiyori. / Static task cluster size determination in homogeneous distributed systems. Software Automatic Tuning: From Concepts to State-of-the-Art Results. Springer New York, 2010. pp. 229-252
    @inbook{72ae846d5187492ab45ef9d6d59967b3,
    title = "Static task cluster size determination in homogeneous distributed systems",
    abstract = "In a distributed system, which consists of an unknown number of processors, it is important to derive an appropriate number of processors by which the good schedule length is obtained by a task scheduling. Many task clustering heuristics have been proposed to determine the number of processors and to minimize the schedule length for scheduling a directed acyclic graph (DAG) application. However, those heuristics are not aware of the actual number of existing processors. As a result, the number of processors determined by an existing task clustering may exceed that of actually existing processors. Therefore, conventional approaches adopt merging of each cluster for reducing the number of clusters at the expense of decreasing degree of task parallelism. In this paper, we present a static cluster size determination method, which derives the lower bound of the cluster size with considering the DAG structure and the task size to data size ratio to suppress the schedule length with the small number of processors. Our experimental evaluations by simulations show that the lower bound of each cluster size determined by the proposed method has a good impact on both the schedule length and the processor utilization.",
    keywords = "Cluster size, DAG, Task clustering, Task scheduling",
    author = "Hidehiro Kanemitsu and Gilhyon Lee and Hidenori Nakazato and Takashige Hoshiai and Yoshiyori Urano",
    year = "2010",
    doi = "10.1007/978-1-4419-6935-4_14",
    language = "English",
    isbn = "9781441969347",
    pages = "229--252",
    booktitle = "Software Automatic Tuning: From Concepts to State-of-the-Art Results",
    publisher = "Springer New York",

    }

    TY - CHAP

    T1 - Static task cluster size determination in homogeneous distributed systems

    AU - Kanemitsu, Hidehiro

    AU - Lee, Gilhyon

    AU - Nakazato, Hidenori

    AU - Hoshiai, Takashige

    AU - Urano, Yoshiyori

    PY - 2010

    Y1 - 2010

    N2 - In a distributed system, which consists of an unknown number of processors, it is important to derive an appropriate number of processors by which the good schedule length is obtained by a task scheduling. Many task clustering heuristics have been proposed to determine the number of processors and to minimize the schedule length for scheduling a directed acyclic graph (DAG) application. However, those heuristics are not aware of the actual number of existing processors. As a result, the number of processors determined by an existing task clustering may exceed that of actually existing processors. Therefore, conventional approaches adopt merging of each cluster for reducing the number of clusters at the expense of decreasing degree of task parallelism. In this paper, we present a static cluster size determination method, which derives the lower bound of the cluster size with considering the DAG structure and the task size to data size ratio to suppress the schedule length with the small number of processors. Our experimental evaluations by simulations show that the lower bound of each cluster size determined by the proposed method has a good impact on both the schedule length and the processor utilization.

    AB - In a distributed system, which consists of an unknown number of processors, it is important to derive an appropriate number of processors by which the good schedule length is obtained by a task scheduling. Many task clustering heuristics have been proposed to determine the number of processors and to minimize the schedule length for scheduling a directed acyclic graph (DAG) application. However, those heuristics are not aware of the actual number of existing processors. As a result, the number of processors determined by an existing task clustering may exceed that of actually existing processors. Therefore, conventional approaches adopt merging of each cluster for reducing the number of clusters at the expense of decreasing degree of task parallelism. In this paper, we present a static cluster size determination method, which derives the lower bound of the cluster size with considering the DAG structure and the task size to data size ratio to suppress the schedule length with the small number of processors. Our experimental evaluations by simulations show that the lower bound of each cluster size determined by the proposed method has a good impact on both the schedule length and the processor utilization.

    KW - Cluster size

    KW - DAG

    KW - Task clustering

    KW - Task scheduling

    UR - http://www.scopus.com/inward/record.url?scp=84860545183&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84860545183&partnerID=8YFLogxK

    U2 - 10.1007/978-1-4419-6935-4_14

    DO - 10.1007/978-1-4419-6935-4_14

    M3 - Chapter

    AN - SCOPUS:84860545183

    SN - 9781441969347

    SP - 229

    EP - 252

    BT - Software Automatic Tuning: From Concepts to State-of-the-Art Results

    PB - Springer New York

    ER -