摘要
提出一种基于Open CL的CPU-GPU协同计算的大规模地形实时渲染算法,该算法侧重于把批LOD模型的构建从CPU移植到GPU.CPU主要负责把外存中的数据块实时调度到内存中,并把相应的地形块载入GPU中的显存.GPU负责利用Open CL平台并行构建LOD模型.为了避免相邻LOD模型出现裂缝,利用地形块的裂缝检测和删除顶点的方法消除裂缝;为了解决两个LOD层次的转换过程中出现地表突跳现象,采用morphing方法实现地形的平滑过渡.实验结果表明,该算法将大量的几何计算移植到GPU上并行计算,降低了CPU的计算负载,提高了LOD模型的构建效率,加快了场景的渲染速度.
The paper proposes a large-scale terrain real-time rendering algorithm using CPU-GPU cooperative computation based on OpenCL,which focuses on moving the construction of the batch LOD model from the CPU to the GPU.CPU is generally responsible for scheduling terrain blocks from the external storage to memory and uploading the corresponding terrain blocks into memory in the GPU.The GPU is responsible for building the LOD model in parallel using the OpenCL platform.To avoid the cracks between the adjacent LOD model,the crack is eliminated by the method of crack detection and deleting the vertices of the terrain block.In order to solve the terrain popping artifacts during the transition of two LOD models,the morphing method is used to realize the smooth transition of the terrain.The experimental results show that a large number of geometric calculation are done in parallel in the CPU,reducing the computation overhead of CPU,improving the efficiency of building the LOD models,and speeding up 3D scene rendering.
作者
郭向坤
林浒
刘继申
王鸿亮
GUO Xiang-kun1,2 LIN Hu2 , LIU Ji-shen1,2 , WANG Hong-liang2(1 University of Chinese Academy of Sciences, Beijing 100049, China ; 2 Shenyang Institute of Computing Technology, Chinese Academy of Sciences, Shenyang 110168, China)
出处
《小型微型计算机系统》
CSCD
北大核心
2018年第4期825-829,共5页
Journal of Chinese Computer Systems
基金
辽宁省博士启动基金项目(201501040)资助
