Performance Analysis of Lattice QCD on GPUs in APGAS Programming Model (UnrefereedWorkshop Manuscript)

• Koichi Shirahata

  Tokyo Institute of Technology

• Jun Doi

  IBM Research - Tokyo

• Mikio Takeuchi

  IBM Research - Tokyo

抄録

The APGAS programming model abstracts deep memory hierarchy such as distributed memory and GPU device memory by a global view of data and asynchronous operations on massively parallel computing environments. However, how much GPUs accelerate applications using the APGAS model remains unclear. In order to understand the effectiveness of using GPUs in the APGAS model, we give a comparative performance analysis of the APGAS model in X10 on GPUs with a standard massage passing model using lattice QCD. Our experimental results on TSUBAME2.5 show that our X10 CUDA implementation on 32 GPUs exhibits 19.4x speedup over X10 C++ on multi-core CPUs, and comparative performance with MPI CUDA in weak scaling. The results indicate that the APGAS programming model on GPUs scales well and accelerates the lattice QCD application significantly.

収録刊行物

• 情報処理学会研究報告. [ハイパフォーマンスコンピューティング]

  情報処理学会研究報告. [ハイパフォーマンスコンピューティング] 2015 (20), 1-8, 2015-02-23

  一般社団法人情報処理学会

キーワード

• PGAS

詳細情報

• CRID: 1571417127846453248
• NII論文ID: 110009877729
• NII書誌ID: AN10463942
• 本文言語コード: en
• データソース種別:
  • CiNii Articles