high performance computing on graphics processing units: hgpu.org

In this paper we investigate a solution to the problem of simulating the spread of epidemics in real-world trading networks. We developed an application that uses parallel computing devices (e.g. GPUs – Graphical Processing Units) with OpenCL (Open Computing Language). Furthermore, we use the epidemiological SIRmodel to represent the nodes of the trading network. Initially, the epidemic grows locally in every node. At certain points of time a transaction happens between severals nodes to spread the epidemic spatially. Our results show that a computational speedup of at least 8 times can be achieved using modern GPUs. Additional research is required to further accelerate the computation.