Virtual molecular docking is a computational method used in computer-aided drug discovery that calculates the binding affinity of a small molecule drug candidate to a target protein. High-throughput virtual screenings calculate the binding affinities for a large number of molecules at once and ranks potential drug candidates to greatly reduces the time and cost of suggesting new potential pharmaceuticals. This high-throughput screening is a task parallel process and therefore well-suited for distributed computing. In this study, we use the open source Hadoop framework implementing the MapReduce paradigm for distributed computing on a cloud platform and the widely used molecular docking program, AutoDock. The initial implementation of AutoDockCloud showed a speed-up of 450 on Kandinsky, a cloud computer located at Oak Ridge National Laboratory. Further modifications show promise for a greater speed-up of large chemical library screenings and also incorporates and distributes the pre-docking procedures. Copyright © 2012 John Wiley & Sons, Ltd.