A multi-grain parallelizing compilation scheme for OSCAR (Optimally scheduled advanced multiprocessor)

H. Kasahara, H. Honda, A. Mogi, A. Ogura, K. Fujiwara, S. Narita

研究成果: Conference contribution

9 被引用数 (Scopus)

抄録

This paper proposes a multi-grain parallelizing compilation scheme for Fortran programs. The scheme hierarchically exploits parallelism among coarse grain tasks, such as, loops, subroutines or basic blocks, among medium grain tasks like loop iterations and among near fine grain tasks like statements. Parallelism among the coarse grain tasks called the macrotasks is exploited by carefully analyzing control dependences and data dependences. The macrotasks are dynamically assigned to processor clusters to cope with run-time uncertainties, such as, conditional branches among the macrotasks and variation of execution time of each macrotask. The parallel processing of macrotasks is called the macro-dataflow computation. A macrotask composed of a Do-all loop, which is assigned onto a processor cluster, is processed in the medium grain in parallel by processors inside the processor cluster. A macrotask composed of a sequential loop or a basic block is processed on a processor cluster in the near fine grain by using static scheduling. A macrotask composed of subroutine or a large sequential loop is processed by hierarchically applying macro-dataflow computation inside a processor cluster. Performance of the proposed scheme is evaluated on a multiprocessor system named OSCAR. The evaluation shows that the multi-grain parallel processing effectively exploits parallelism from Fortran programs.

本文言語English
ホスト出版物のタイトルLanguages and Compilers for Parallel Computing - 4th International Workshop, Proceedings
編集者Utpal Banerjee, David Gelernter, Alex Nicolau, David Padua
出版社Springer Verlag
ページ283-297
ページ数15
ISBN(印刷版)9783540554226
DOI
出版ステータスPublished - 1992
イベント4th Workshop on Languages and Compilers for Parallel Computing, 1991 - Santa Clara, United States
継続期間: 1991 8 71991 8 9

出版物シリーズ

名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
589 LNCS
ISSN(印刷版)0302-9743
ISSN(電子版)1611-3349

Other

Other4th Workshop on Languages and Compilers for Parallel Computing, 1991
CountryUnited States
CitySanta Clara
Period91/8/791/8/9

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

フィンガープリント 「A multi-grain parallelizing compilation scheme for OSCAR (Optimally scheduled advanced multiprocessor)」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル