Research And Implementation Of The Highly Efficient Ray Tracing Based On GPU

GPU architecture through the process in-depth analysis and GPU-based streaming programming model, this paper uses uniform grid acceleration structure, and realizes ray tracing based on GPU of stream programming model. Throughout the rendering process needs to be done a lot of light intersection calculation, reducing the efficiency of the entire scene rendering. To solve this problem, under the premise of no reducing the quality of rendering, the paper introduces the hardware Z-buffered rasterization to accelerate ray to determine the main way to improve the efficiency of the primary light is generated, and its experimental verification carried out.Through the analyse of the process of the GPU architecture, this paper obtained the conclusion that GPU operating performance related not only with the structural design of the GPU,but also with its frequency and effectiveness of the implementation.The higher the frequency and effectiveness of the implementation, GPU performance is relatively the better. By comparative analysis of the pre-experimental and improved experimental data, proved that the improved strategy effectively increases the efficiency of ray tracing, and its effect is good.