Research on FFT Algorithms for Exa-Scale Computing Environment

• Project/Area Number: 24500029
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Software
• Research Institution: University of Tsukuba
• Principal Investigator: TAKAHASHI Daisuke 筑波大学, システム情報系, 教授 (00292714)
• Project Period (FY): 2012-04-01 – 2016-03-31
• Project Status: Completed (Fiscal Year 2015)
• Budget Amount: ¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000); Fiscal Year 2014: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000); Fiscal Year 2013: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000); Fiscal Year 2012: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
• Keywords: 高速フーリエ変換 / エクサスケール計算環境

Outline of Final Research Achievements

In this research, we investigated an implementation of parallel fast Fourier transform (FFT) algorithm for exa-scale computing environment. Parallel FFTs require intensive all-to-all communication, which affects the performance of parallel FFTs. An automatic tuning facility for selecting the optimal parameters of the communication hiding, the radices, and the block size is implemented. The performance results demonstrate that the implementation of
parallel FFTs with automatic tuning is efficient for improving the performance.

Research Products

• [Presentation] Automatic Tuning for Parallel FFTs on Intel Xeon Phi Clusters (2016)
  • Author(s): Daisuke Takahashi
  • Organizer: 2016 SIAM Conference on Parallel Processing for Scientific Computing (PP16)
  • Place of Presentation: Pierre and Marie Curie University, Paris, France
  • Year and Date: 2016-04-12
  • Int'l Joint Research
• [Presentation] Automatic Tuning for Parallel FFTs on Intel Xeon Phi Clusters (2016)
  • Author(s): Daisuke Takahashi
  • Organizer: 2016 Conference on Advanced Topics and Auto Tuning in High-Performance and Scientific Computing (2016 ATAT in HPSC)
  • Place of Presentation: National Taiwan University, Taipei, Taiwan
  • Year and Date: 2016-02-19
  • Int'l Joint Research
• [Presentation] Xeon Phiにおける並列FFTの実現と評価 (2015)
  • Author(s): 高橋大介
  • Organizer: 日本応用数理学会2015年度年会
  • Place of Presentation: 金沢大学角間キャンパス，石川県金沢市
  • Year and Date: 2015-09-09
• [Presentation] Xeon Phiクラスタにおける並列FFTの自動チューニング (2015)
  • Author(s): 高橋大介
  • Organizer: 第20回計算工学講演会
  • Place of Presentation: つくば国際会議場，茨城県つくば市
  • Year and Date: 2015-06-08
• [Presentation] Automatic Tuning for Parallel FFTs on GPU Clusters (2015)
  • Author(s): Daisuke Takahashi
  • Organizer: 2015 SIAM Conference on Computational Science and Engineering (CSE15)
  • Place of Presentation: Salt Palace Convention Center, Salt Lake City, USA
  • Year and Date: 2015-03-14 – 2015-03-18
• [Presentation] GPUクラスタにおける並列FFTの自動チューニング (2014)
  • Author(s): 高橋大介
  • Organizer: 日本応用数理学会2014年度年会
  • Place of Presentation: 政策研究大学院大学，東京都港区
  • Year and Date: 2014-09-03 – 2014-09-05
• [Presentation] Implementation of Parallel FFTs on GPU Clusters (2014)
  • Author(s): Daisuke Takahashi
  • Organizer: 2014 Conference on Advanced Topics and Auto Tuning in High Performance Scientific Computing
  • Place of Presentation: Taipei, Taiwan
  • Invited
• [Presentation] Experience of Implementing Parallel FFTs on GPU Clusters (2013)
  • Author(s): Daisuke Takahashi
  • Organizer: Special Session: Legacy HPC Application Migration 2013 (LHAM) (held in conjunction with IEEE MCSoC-13)
  • Place of Presentation: Tokyo, Japan
• [Presentation] Implementation of Parallel 1-D FFT on GPU Clusters (2013)
  • Author(s): Daisuke Takahashi
  • Organizer: 2013 IEEE 16th International Conference on Computational Science and Engineering(CSE 2013)
  • Place of Presentation: Sydney, Australia
• [Presentation] An Implementation of Parallel 1-D FFT on the K computer (2012)
  • Author(s): Daisuke Takahashi, Atsuya Uno and Mitsuo Yokokawa
  • Organizer: 2012 IEEE 14th International Conference on High Performance Computing and Communications (HPCC-2012)
  • Place of Presentation: Liverpool, United Kingdom