二维传热数值计算程序在图形卡上的实现  

IMPLEMENTION OF TWO DIMENSIONAL NUMERICAL CONDUCTIVE HEAT TRANSFER ON GPU

作　　者：杨昆仑[1] 宋耀祖[1] 任建勋[1] 

机构地区：[1]清华大学航天航空学院,北京100084

出　　处：《工程热物理学报》2008年第1期151-153,共3页Journal of Engineering Thermophysics

摘　　要：本文针对通用图形处理器(GPGPU)的特点,对二维传热数值求解算法及三类不同边界条件问题在GPU上的求解进行了分析。图形处理器是一种专用并行处理器,主要用于三维场景渲染加速,其结构和程序设计方式与CPU完全不同,但却比CPU具有更强的浮点运算能力和内存带宽。随着图形处理器可编程特性的发展,它已越来越超出原有的应用范围,向通用计算领域发展,成为一个新的研究领域。采用Jacobi迭代,能够在GPU上实现二维导热方程的求解。本文在已有研究的基础上,实现了二维导热计算的混合边界条件问题在GPU上的求解方法,通过计算时间的比较显示GPU在计算传热学中有良好应用前景。The numerical simulation of two dimensional conductive heat transfers with combined boundary condition on Graphic Processing Unit （GPU） is investigated in this paper. GPU is dedicated graphics rendering device with highly parallel structure, and it has more processing power and memory bandwidth than CPU. As the developing of its programming ability, GPU is exceeding the limitation of its traditional applications. General purpose GPU（GPGPU） has become to a new research area. By using Jacobi arithmetic, two dimensional heat transfers with combined boundary condition is implemented, and the result shows that GPU has very good potential application in numerical heat transfer.

关 键 词：导热 数值计算 图形处理器 GPGPU 

分 类 号：TK124[动力工程及工程热物理—工程热物理]