using NVIDIA FFT CUDA Library to solve FFT problem

GPU实现的MapReduce framework,对于学习并行编程和cuda平台的编程方面有着极好的参考价值，里面附带论文。用户要求有NViDIA显卡，并且安装cuda编程环境。-We design and implement Mars, a MapReduce framework, on graphics processors (GPUs). MapReduce is a distributed programming framework originally proposed by

在CUDA下的FFT变换程序，可以供爱好GPU编程的参考学习。-It it the FFT transform in CUDA, and very suitable for the programer who is interested in GPU.