FPGA Based OpenCL Acceleration of Genome Sequencing Software
Authors: Ashish Sirasao (Xilinx Inc.), Elliott Delaye (Xilinx Inc.), Ravi Sunkavalli (Xilinx Inc.), Stephen Neuendorffer (Xilinx Inc.)
Abstract: The Smith-Waterman algorithm produces the optimal pairwise alignment between two sequences of proteins or nucleotides and is frequently used as a key component of alignment and variation detection tools for next-generation sequencing data. In this paper an efficient and scalable implementation of the Smith-Waterman algorithm is written in OpenCL and implemented on a Xilinx Virtex-7 FPGA which shows >2x compute performance and 18x-20x performance per watt advantage compared with 12 core and 60 core CPUs. Against a GPU we show a 30% performance advantage and 11.6x better performance per watt. These results were achieved using an off the shelf PCIe accelerator card by optimizing kernel throughput of a systolic array architecture with compiler pipelining of the OpenCL kernels, carefully mapping memories used during computation and adjusting the number of systolic nodes per compute unit to fully utilize the FPGA resources.
Two-page extended abstract: pdf