Back to Search Start Over

基于容器技术的水文水动力模型软硬件适配方法.

Authors :
张海嘉
刘家宏
梅超
王佳
高希超
Source :
Application Research of Computers / Jisuanji Yingyong Yanjiu. Jan2024, Vol. 41 Issue 1, p188-205. 6p.
Publication Year :
2024

Abstract

With the development and popularity of advanced RISC machines(ARM), adapting the X86-based scientific computing software to ARM-based computing platforms is one of the key issues for scientific research and operational applications. It is urgent to explore the software-hardware-adaptation methods of advanced RISC machines for professional computational models. This paper systematically analysed the key points for software hardware adaptation of ARM. By comparing the existing software adaptation technologies, it selected container technology to enable the hydrodynamic model to run on ARM-based computing platforms. Container technology could package the application software and its dependencies into a portable container, which made the suggested software run freely without any dependency on the underlying architecture and operating system. Taking the hydrological hydrodynamic model, TELEMAC, as an example, it explained the principle of Docker image composition in detail, and built the TELEMAC image environment through Dockerfile. The method was verified by carrying out the computational cases. The results show that the TELEMAC image can run safely on the Kunpeng 920 processor-based openEuler and Kirin V10. The case calculation results are consistent with the standard results, and the model calculation efficiency is high. The proposed method realizes the adaptation from an X86-based platform to ARM-based computing platforms for professional scientific computing models. This study can also provide a reference for the localization adaptation of other software. [ABSTRACT FROM AUTHOR]

Details

Language :
Chinese
ISSN :
10013695
Volume :
41
Issue :
1
Database :
Academic Search Index
Journal :
Application Research of Computers / Jisuanji Yingyong Yanjiu
Publication Type :
Academic Journal
Accession number :
175061736
Full Text :
https://doi.org/10.19734/j.issn.1001-3695.2023.05.0200