(加急)基于GPU的并行加速渲染算法的研究

Research of the Parallel Rendering Algorithm Based on GPU

摘要:

目前,GPU渲染技术仍然存在不足之处,在分配渲染任务时,没有充分发挥各个处理器的优势,浪费了性能较好的处理器,影响了渲染速度的提高。针对以上问题,根据武进区邹区现代农业产业园人口疏散地域接收安置动画的项目,在原有GPU渲染架构的基础上提出了一种基于sort-last架构的带反馈的动态负载均衡算法。在分配渲染任务之前,对所有处理器进行性能统计,将时长较长的场景分配给性能较好的处理器,时长短的分配给剩余的处理器,待有处理器完成渲染任务时,将剩余场景时间较长的,分配给第一轮任务先完成的处理器,以此类推。实验结果表明,该方法对解决上述问题具有较好的效果,解决了目前集群渲染存在的不足之处和负载不平衡的问题,最终实现了加速渲染。

Abstract:

At present, the existing GPU rendering technology still exist deficiencies, in the allocation of rendering task, did not give full play to the advantages of each processor, wasted better processor, influence the improvement of the rendering speed. According to the receiving resettlement animation projects Wujin District Zou industrial park of modern agricultural population evacuation areas, to solve above problems, on the basis of the original GPU rendering architecture, proposed based on a sort - last architecture with feedback of dynamic load balancing algorithm. Performance statistics for all processors, long scene assigned to good processor performance, short scene assigned to the rest of the processor, before rendering assignment, when one processor finish rendering task, the remaining longer scene assigned to the first round of task to complete the processor, and so on. The experimental results show that this method has a good effect to solve the above problems, to solve the shortcomings and load imbalance cluster rendering currently exist, accelerate rendering finally.