针对瞬态分析耗时长的问题,研究了基于CUDA的并行计算方法。CUDA是NVIDIA公司推出的基于GPU的并行编程模型,采用C语言编程。它利用GPU多处理器的特点来实现大规模数据的并行计算,从而显著提高问题的求解速度。本文将CUDA技术应用到瞬态分析问题的显式求解上,较大的提高了求解效率。 Aiming at the problem of time-consuming transient analysis, a parallel computation method based on CUDA is studied. CUDA is NVIDIA's GPU-based parallel programming model introduced in C programming language. It takes advantage of GPU multiprocessor features to achieve large-scale data parallel computing, thus significantly improving the speed of problem solving. In this paper, CUDA technology is applied to the explicit solution of transient analysis problems, which greatly improves the efficiency of the solution.