high performance computing on graphics processing units: hgpu.org

BACKGROUND: With the exploding volume of data generated by continuously evolving high-throughput technologies, biological network analysis problems are growing larger in scale and craving for more computational power. General Purpose computation on Graphics Processing Units (GPGPU) provides a cost-effective technology for the study of large-scale biological networks. Designing algorithms that maximize data parallelism is the key in leveraging the power of GPUs. RESULTS: We proposed an efficient data parallel formulation of the All-Pairs Shortest Path problem, which is the key component for shortest path-based centrality computation. A betweenness centrality algorithm built upon this formulation was developed and benchmarked against the most recent GPU-based algorithm. Speedup between 11 to 19% was observed in various simulated scale-free networks. We further designed three algorithms based on this core component to compute closeness centrality, eccentricity centrality and stress centrality. To make all these algorithms available to the research community, we developed a software package gpu-fan (GPU-based Fast Analysis of Networks) for CUDA enabled GPUs. Speedup of 10-50x compared with CPU implementations was observed for simulated scale-free networks and real world biological networks. CONCLUSIONS: gpu-fan provides a significant performance improvement for centrality computation in large-scale networks. Source code is available under the GNU Public License (GPL) at http://bioinfo.vanderbilt.edu/gpu-fan/.