組合せ最適化による並列数値シミュレーションの静的負荷分散  [in Japanese] Static Load Balancing for Parallel Numerical Simulation by Combinatorial Optimization (Special Issue on Parallel Processings)  [in Japanese]

Access this Article

Search this Article

Author(s)

Abstract

並列数値シミュレーション用高水準言語NSLのための静的負荷分散手法について検討する.計算領域を複数のブロックに分割し,計算量と通信量を考慮して各ブロックに適切なプロセッサを割り当てることにより,シミュレーションの実行時間を最小化する.この負荷分散法を組合せ最適化問題として定式化し,分枝限定法を用いて最適解を求めた.規模の大きな問題では計算量の問題から最適解を求めることができないが,本論文で提案する近似アルゴリズムにより短時間で精度の良い近似解が求まることを数値実験で示す.近似による誤差は妥当な条件下では最適値から15%以下であった.また,数値実験の実行時間から最適解の求解時間を見積もる近似式を求めた.本手法は評価関数の変更によって広範囲の並列処理応用に適用可能である.A static load balancing scheme is discussed for parallel numerical simulation language NSL.NSL partitions computational domain into multiple blocks,and allocates processors optimally for each block in accordance with computation and communication cost.This allocation problem is formulated as a combinatorial optimization problem,and solved by branch-and-bound method.Though large problems cannot be solved by this method because of combinatorial explosion,an effective approximation algorithm is presented and evaluated by numerical simulation.The error of this approximation algorithm is less than 15% under reasonable condition.The execution time for this optimization is also measured in numerical simulation to induce the estimation equatin.The allocation method presented here is widely applicable by adapting evaluation function for each purpose.

A static load balancing scheme is discussed for parallel numerical simulation language NSL. NSL partitions computational domain into multiple blocks, and allocates processors optimally for each block in accordance with computation and communication cost. This allocation problem is formulated as a combinatorial optimization problem, and solved by branch-and-bound method. Though large problems cannot be solved by this method because of combinatorial explosion, an effective approximation algorithm is presented and evaluated by numerical simulation. The error of this approximation algorithm is less than 15% under reasonable condition. The execution time for this optimization is also measured in numerical simulation to induce the estimation equation. The allocation method presented here is widely applicable by adapting evaluation function for each purpose.

Journal

  • Transactions of Information Processing Society of Japan

    Transactions of Information Processing Society of Japan 39(6), 1746-1756, 1998-06-15

    Information Processing Society of Japan (IPSJ)

References:  13

Cited by:  3

Codes

  • NII Article ID (NAID)
    110002722188
  • NII NACSIS-CAT ID (NCID)
    AN00116647
  • Text Lang
    JPN
  • Article Type
    Journal Article
  • ISSN
    1882-7764
  • NDL Article ID
    4491430
  • NDL Source Classification
    ZM13(科学技術--科学技術一般--データ処理・計算機)
  • NDL Call No.
    Z14-741
  • Data Source
    CJP  CJPref  NDL  NII-ELS  IR  IPSJ 
Page Top