high performance computing on graphics processing units: hgpu.org

Computing driving directions interactively on continental road networks requires preprocessing. This step can be costly, limiting our ability to incorporate new optimization functions, including traffic information or personal preferences. We show how the performance of the state-of-the-art customizable route planning (CRP) framework is boosted by GPUs, even though it has highly irregular structure. Our experimental study reveals that our method is an order of magnitude faster than a highly-optimized parallel CPU implementation, enabling interactive personalized driving directions on continental scale.