Parallelized Force-Directed Edge Bundling on the GPU

How to draw large scale spatial interaction data clearly and quickly is a challenge in high performance data visualization research and application field. Force-Directed Edge Bundling (FDEB) helps display graph clearly with significant clutter reduction, but with high time complexity. This paper presents a parallelized FDEB on the GPU (GPU-FDEB), which reforms FDEB and achieves a balanced partitioning of data and calculation to suit computation on the GPU. GPU-FDEB addresses the problem of high time complexity and accelerates FDEB by an order of magnitude.