基于四阶龙格-库塔法的多弹道并行计算研究
Research on Parallel Calculation of Multi-ballstic Based on Fourth-order Runge-Kutta Method
范磊 1周琳2
作者信息
- 1. 西安电子科技大学,陕西 西安 710071;航空工业洛阳电光设备研究所 光电控制技术重点实验室,河南 洛阳 471023
- 2. 航空工业洛阳电光设备研究所 光电控制技术重点实验室,河南 洛阳 471023
- 折叠
摘要
弹道计算是一类典型的计算密集型问题,在飞行器设计及航迹规划中起着至关重要的作用,对该问题的求解通常具有计算量大、耗时长等问题.为实现弹道解算加速,本文利用多线程并行计算技术设计了基于四阶龙格-库塔法的多弹道加速计算架构,通过任务划分将多弹道解算映射到不同的线程并调度至处理器内核.通过试验分析了多弹道计算关键因素的影响,试验结果验证了利用多核多线程并行加速架构可以在保证高精度要求下有效地提高多弹道计算效率,并且可应用于提高微分方程类问题的解算效率.
Abstract
Ballistic calculation is a type of typical computing intensive problem,which plays a vital role in aircraft design and trajectory planning.The solution to this problem usually involves a large amount of calculation and time-consuming problems.In order to realize the acceleration of ballistic calculation,this paper uses multi-threaded parallel computing technology to design a multi-ballistic acceleration calculation architecture based on the fourth-order Runge-Kutta method,and maps the multi-ballistic calculation to different threads through task division.The threads are then scheduled to the processor cores for load balancing.Experiments are designed to analyze the influences of key factors in multi-ballistic calculation,and the simulation results show using multi-ballistic parallel acceleration architecture can effectively improve the time efficiency of ballistic calculation and the computing efficiency for the differential equations with high precision and high computational complexity.The developed architecture can also be used to accelerate the solving procedure for the differential equation problems.
关键词
弹道计算/微分方程/并行运算/龙格-库塔法/多核多线程技术Key words
ballistic calculation/differential equations/parallel computing/Runge-Kutta method/multi-threaded technology引用本文复制引用
基金项目
国家自然科学基金(61502371)
航空科学基金(201951081003)
陕西省自然科学基础研究计划项目(2018JQ6047)
出版年
2024
国家科技期刊平台
NSTL
万方数据