Design and Implementation of A High-Performance Rendering Algorithm for Large Terrain Systems based on GPU Parallel Computing

Although traditional 3D terrain algorithms can improve the rendering efficiency of the terrain, they often ignore the performance of the terrain itself. The use of four textures is not sufficient to deal with complex terrain scenes. Therefore, a terrain virtual texture mapping algorithm based on GPU instancing was proposed, which can map multiple surface textures of more terrains to one virtual texture before rendering, preventing the operation of GPU pipelines from being interrupted due to the change of surface textures. In this case, the parallel computing power of GPU can be better utilized, completing the efficient rendering of diverse terrains with fewer rendering times. The experimental results have shown that this algorithm can increase the diversity of terrain changes so as to applicable to more complex large terrain systems.