Co-design of hardware and software to accelerate a numerical integration with high precision.

2017 Fiscal Year Final Research Report

• Project/Area Number: 15H03602
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Computational science
• Research Institution: Hitotsubashi University
• Principal Investigator: DAISAKA HIROSHI 一橋大学, 大学院商学研究科, 教授 (80399295)
• Co-Investigator(Kenkyū-buntansha): 湯浅 富久子 大学共同利用機関法人高エネルギー加速器研究機構, 計算科学センター, 准教授 (00203943) ; 中里 直人 会津大学, コンピュータ理工学部, 上級准教授 (00392051) ; 似鳥 啓吾 国立研究開発法人理化学研究所, 計算科学研究機構, 研究員 (80600824) ; 石川 正 大学共同利用機関法人高エネルギー加速器研究機構, 計算科学センター, 准教授 (90184481)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 多倍長精度演算 / アクセラレータ / FPGA / ファインマンループ図形演算 / エルミート積分

Outline of Final Research Achievements

We developed a dedicated system to accelerate numerical integrations with multi precision operations. We designed a processor enabling 4/6/8 precision operations and implemented it on a Field-programmable gate array. We also constructed a software library with OpenCL which supports a calculation with high precision. This library enables us to perform a calculation with high precision at high speed on GPUs. Both can be utilized uniformly by our developed compiler system. Applying to our applications, we demonstrated a validity of our system.

Free Research Field

惑星科学