FPGA acceleration of rigid-molecule docking codes
Computer Architecture and Automated Design Laboratory, Department of Electrical and Computer Engineering, Boston University; Boston, MA 02215, USA
IET Computers & Digital Techniques, 2009
@article{sukhwani2010fpga,
title={FPGA acceleration of rigid-molecule docking codes},
author={Sukhwani, B. and Herbordt, M.C.},
journal={Computers & Digital Techniques, IET},
volume={4},
number={3},
pages={184–195},
year={2010},
publisher={IET}
}
Modelling the interactions of biological molecules, or docking, is critical both to understanding basic life processes and to designing new drugs. The field programmable gate array (FPGA) based acceleration of a recently developed, complex, production docking code is described. The authors found that it is necessary to extend their previous three-dimensional (3D) correlation structure in several ways, most significantly to support simultaneous computation of several correlation functions. The result for small-molecule docking is a 100-fold speed-up of a section of the code that represents over 95% of the original run-time. An additional 2% is accelerated through a previously described method, yielding a total acceleration of 36x over a single core and 10x over a quad-core. This approach is found to be an ideal complement to graphics processing unit (GPU) based docking, which excels in the protein-protein domain.
August 15, 2011 by hgpu
