2026-03-07T09:59:00Z https://ipsj.ixsq.nii.ac.jp/oai

oai:ipsj.ixsq.nii.ac.jp:00239733 2025-01-19T08:11:24Z 1164:2240:11467:11720

Development of a Python Interface for Code-To-Code Adapter (CoToCoA) Library Development of a Python Interface for Code-To-Code Adapter (CoToCoA) Library Haichao, Zhao Keiichiro, Fukazawa Takeshi, Iwashita Takeshi, Nanri Yohei, Miyake Yuto, Katoh Haichao, Zhao Keiichiro, Fukazawa Takeshi, Iwashita Takeshi, Nanri Yohei, Miyake Yuto, Katoh スーパーコンピューティング環境 This paper presents the development of a Python interface for the Code-To-Code Adapter (CoToCoA) library, a framework designed for coupled computations that enables direct program-to-program communication without intermediate files. Considering CoToCoA's current support for C and Fortran, as well as its reliance on the MPI library for communication, this study aims to enhance its usability by integrating it with Python. To effectively integrate CoToCoA with Python, we explored two development approaches: direct development in Python and the use of ctypes to call CoToCoA's C functions from Python. Furthermore, we conducted performance evaluations focusing on CoToCoA's two types of data transfer methods to evaluate the efficiency of the developed Python interface. The newly developed Python interface enables the efficient transfer of data from C or Fortran-based simulation programs to Python, where it can be effectively analyzed or processed. This capability significantly enhances the practical utility of CoToCoA in computational science by enabling efficient data handling across programming languages. This paper presents the development of a Python interface for the Code-To-Code Adapter (CoToCoA) library, a framework designed for coupled computations that enables direct program-to-program communication without intermediate files. Considering CoToCoA's current support for C and Fortran, as well as its reliance on the MPI library for communication, this study aims to enhance its usability by integrating it with Python. To effectively integrate CoToCoA with Python, we explored two development approaches: direct development in Python and the use of ctypes to call CoToCoA's C functions from Python. Furthermore, we conducted performance evaluations focusing on CoToCoA's two types of data transfer methods to evaluate the efficiency of the developed Python interface. The newly developed Python interface enables the efficient transfer of data from C or Fortran-based simulation programs to Python, where it can be effectively analyzed or processed. This capability significantly enhances the practical utility of CoToCoA in computational science by enabling efficient data handling across programming languages. technical report 情報処理学会 2024-09-23 application/pdf 研究報告ハイパフォーマンスコンピューティング（HPC） 4 2024-HPC-196 1 6 2188-8841 AN10463942 https://ipsj.ixsq.nii.ac.jp/record/239733/files/IPSJ-HPC24196004.pdf eng