基于Brook在GPU的应用

Application of GPU Based on Brook

依据现代GPU在通用计算方面的功能,提出了基于Brook在GPU上执行通用计算的实现方法,运用图像分割、快速傅立叶算法和光线跟踪3个应用对该方法进行了评估。评估结果表明GPU在Brook环境下通用计算能力优于CPU。

According to the capability of modern GPU （Graphics Process Unit） in general computing, this paper brings forward a Brook-based implementation method about processing general computing in GPU. Then, it evaluates the method through three applications：picture partition, FFT algorithm implementation and ray tracking. The result shows that under the Brook environment GPU owns a higher general computing capability than CPU.