中国机械工程 ›› 2025, Vol. 36 ›› Issue (9): 1942-1950.DOI: 10.3969/j.issn.1004-132X.2025.09.005

• 机械基础工程 • 上一篇    

基于CPU-GPU的超音速流场N-S方程数值模拟

卢志伟(), 张皓茹, 刘锡尧, 王亚东, 张卓凯, 张君安   

  1. 西安工业大学机电工程学院, 西安, 710021
  • 收稿日期:2024-07-15 出版日期:2025-09-25 发布日期:2025-10-15
  • 通讯作者: 卢志伟
  • 作者简介:卢志伟*(通信作者),男,1979年生,副教授、博士研究生导师。研究方向为流体润滑与并行计算。E-mail:luzhiwei@xatu.edu.cn
  • 基金资助:
    国家自然科学基金(52301101);陕西省科技厅项目(2025ZG-JBGS-007);陕西省教育厅项目(23JC040)

Numerical Simulation of N-S Equations for Supersonic Flow Fields Based on CPU-GPU

Zhiwei LU(), Haoru ZHANG, Xiyao LIU, Yadong WANG, Zhuokai ZHANG, Jun'an ZHANG   

  1. School of Mechatronic Engineering,Xi'an Technological University,Xi'an,710021
  • Received:2024-07-15 Online:2025-09-25 Published:2025-10-15
  • Contact: Zhiwei LU

摘要:

为深入分析超音速流场的特性并提高数值计算效率,设计了一种高效的加速算法。该算法充分利用中央处理器-图形处理器(CPU-GPU)异构并行模式,通过异步流方式实现数据传输及处理,显著加速了超音速流场数值模拟的计算过程。结果表明:GPU并行计算速度明显高于CPU串行计算速度,其加速比随流场网格规模的增大而明显提高。GPU并行计算可以有效提高超音速流场的计算速度,为超音速飞行器的设计、优化、性能评估及其研发提供一种强有力的并行计算方法。

关键词: 超音速流场, 中央处理器-图形处理器, 异构计算, 有限差分

Abstract:

To thoroughly analyze the characteristics of supersonic flow fields and enhance the efficiency of numerical computations, an efficient acceleration algorithm was designed. The algorithm herein fully leveraged the CPU-GPU heterogeneous parallel architecture and achieved data transmission and processing through asynchronous streaming, significantly accelerating the computational processes of supersonic flow field numerical simulations. The results demonstrate that the computational speed of GPU parallel processing is markedly faster than that of CPU serial processing, and the speedup ratio exhibits a pronounced increasing trend as the scale of the flow field grid expands. GPU parallel computing may effectively improve the computational speed of supersonic flow field simulations, providing a robust parallel computing method for the design, optimization, performance evaluation, and development of supersonic aircrafts.

Key words: supersonic flow field, central processing unit - graphics processing unit(CPU-GPU), heterogeneous computing, finite difference

中图分类号: