说明：  CUDA是一个更适合于并行计算的架构，提供了硬件的直接访问接口，并率先提供了针对GPU(图形处理器)编程的C语言开发环境。 CUDA编程指南2.0
(CUDA is a more suitable for parallel computing architecture, providing direct access to the hardware interface and the first to provide for the GPU (graphics processor) programming, C language development environment. CUDA Programming Guide 2.0)