A new parallel radiosity algorithm based on task mapping

The main problem for simulating large scenes based on radiosity algorithm is a highly time-consuming process. We present a task mapping model to achieve a good load balance on a distributed memory system. And we keep our focus on the solutions for datum sets with complex geometric models. In order to evaluate the proposed method, a parallel radiosity algorithm based on our models has been tested on supercomputer simulator with cellular network. An efficient performance in terms of scalability and speedup has been obtained.