Complex network analysis: The need for speed

Complex network analysis has become a critical research area for the design and evaluation of complex systems, due to the increasing success of modeled phenomena. Accordingly, there is a growing need for easy-to-use and fast network analysis software packages. In this work, we report on a comprehensive benchmark of state-of-the-art network analysis software. We perform exhaustive experiments on a set of eight competitors for 15 important network analysis tasks. All methods are evaluated on 13 publicly available datasets. We find that none of the competitors clearly dominates our competitive evaluation, while for specific tasks, the running times differ by two orders of magnitude and more. Our study helps researchers to select an appropriate tool for their network analysis tasks.