Parallel Rendering for Large-Scale Crowd Based on Dynamic Feedback

Rendering large-scale virtual crowd with high speed is still a challenge for the computer graphics community. In this paper, we introduce a parallel rendering method for scenes with large-scale virtual crowd, which has been successfully applied to the rendering of the scene of 140 thousand square meters with several hundred thousand people on a PC cluster. A "dynamic feedback" mechanism is introduced to resolve the problem of loading imbalance in parallel rendering. It makes a dynamical plan for task assignment before each frame based on the implementation results of the former frame. Thus our rendering method could dynamically react to not only the variation of rendering task but also the variation of cluster environments. Experiments show that, our method is capable of rendering several hundred thousand individuals efficiently and has flexibility to deal with the variation of the cluster environment.