An Architecture for Distributed Behavioral Models with GPUs

We describe an architecture for massive simulation of a distributed behavioral model using graphics hardware. By leveraging on the recent programmable capabilities of GPUs we implemented the model capable of managing a large aggregate motion of birds in a virtual environment that can avoid both static and dynamic obstacles. We demonstrate the effectiveness of our GPU implementation by comparing the results to a CPU implementation and, emphasize the modularity of the proposed architecture that favors reusability in several contexts. Categories and Subject Descriptors (according to ACM CCS): I.6.8 [Simulation and modelling]: Animation

• Type: Conference
• Year: 2006
• Where: EGITALY
• Authors: Rosario De Chiara, Ugo Erra, Vittorio Scarano
• Keywords: Behavioral Model, Computer Graphics, EGITALY 2006, Large Aggregate Motion, Massive Simulation