Performance Evaluation of 3D Game Rendering with and without Frustum Culling and PVS Culling
In this paper, the performance of 3D game rendering with and without application of frustum culling and PVS (Potentially Visible Set) culling is evaluated in FPS(frame per second). Both frustum culling and PVS culling are two representative methods to efficiently remove invisible contents before a rendering stage in a large 3D game scene. As a result of the performance evaluation, frustum culling and PVS culling have a large effect on rendering performance. In case of not applying any culling, the average performance is decreased by 15.8% compared to the case where all culling is applied. Also, in this paper, the number of portals for indoor static contents in the 3D game contents used in the experiment is larger, so that the application of PVS culling has more influence on the performance than frustum culling.