图形处理器并行计算用于离子发动机粒子模拟

 为了研究离子发动机羽流对航天器的影响，采用质点网格-蒙特卡罗碰撞方法对离子发动机羽流中的交换电荷离子进行了模拟。利用计算设备统一架构技术，开发出一套基于图形处理器的并行粒子模拟程序。随机数生成采用并行MT19937伪随机数生成器算法，电场方程使用完全近似存储格式的代数多重网格法求解。r-z轴对称坐标系中，在z=0 m处获得的电流密度均值为4.5×10^-5 A/m²，图形处理器所得结果与中央处理器模拟结果吻合。在16核心的NVIDIA GeForce 9400 GT图形显示卡上，取得相对于Intel Core 2 E6300中央处理器4.5~10.0倍的加速比。

Particle simulation of ion thruster using parallel computation on graphic processor unit

A numerical simulation has been conducted using the particle-in-cell with Monte Carlo collision (PIC-MCC) method to model charge-exchange (CEX) xenon ions in the plume of ion thruster. A parallel particle simulation code has been developed by using the Compute Unified Device Architecture (CUDA) for graphic processor unit (GPU). The random number is generated by the parallel Mersenne Twister pseudorandom number generator algorithm. The full approximation storage-algebraic multigrid method (FAS-AMG) is used to update the electric field. In an axis-symmetric(r-z) coordinate, the average ion current flux is 4.5×10^-5 A/m² at z=0 m, and GPU’s result agrees well with CPU’s. Compared with a single CPU Intel Core 2 E6300, 16-processor GPU NVIDIA GeForce 9400 GT represents 4