Challenge for code tuning of plasma particle-in-cell code using a manycore processor

2017 Fiscal Year Final Research Report

• Project/Area Number: 26600152
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Computational science
• Research Institution: Nagoya University
• Principal Investigator: Saito Shinji 名古屋大学, 理学研究科, 特任准教授 (60528165)
• Project Period (FY): 2014-04-01 – 2018-03-31
• Keywords: 計算物理 / 宇宙物理 / プラズマ粒子シミュレーション

Outline of Final Research Achievements

We study code tuning of plasma particle-in-cell (PIC) code for many core processor (Xeon Phi). In PIC method, calculations for charged particle motion and electric current density are the most expensive. It is the most important object for PIC method to reduce their calculation cost by using many core system. We have reduced the computational cost for particle motion, while there are some difficulty to reduce the cost for electric current density, because of random memory access of many charged particles for current density grids. It is the most important task for future study to redistribute particle data to realize sequential memory access that enable compilers to vectorize the calculation.

Free Research Field

プラズマ物理