摘要： 通过有效利用图形硬件的图形处理单元(GPU)的运算能力和可编程性，将大量计算从CPU 分离出来。在GPU 上采用顶点和片元程序进行阴影计算，从而加速复杂场景阴影绘制。选择图像空间阴影算法进行GPU 加速绘制。用Cg 图形编程语言和OpenGL 实现了算法的绘制过程，能够满足通用的复杂3D 场景应用的需要，达到满意的实时绘制效果。

关键词: 图形处理单元；阴影算法；顶点程序；片元程序

Abstract: With the efficient use of computation capacity and programmability provided by graphics processing unit(GPU)of underlying graphics hardware, most of computation from CPU are isolated. Vertex program and fragment program are used for shadow computation, and then the paper accelerates shadow rendering on complex geometry scene. It chooses an image-space shadow algorithm, shadow mapping, for GPU accelerated rendering. The algorithm in Cg shading language and OpenGL is implemented, it can satisfy the requirement of complex 3D application, and gives convincible real-time rendering results

Key words: GPU; Shadow algorithms; Vertex program(VP); Fragment program(FP)