Publication

Exact sequence matching lets researchers find recurring patterns and mutations in DNA, information that can support faster diagnoses and personalized treatments. Because it is computationally heavy, it often slows down the whole analysis.
This paper presents a multi-core FPGA implementation of the Knuth-Morris-Pratt (KMP) algorithm that processes sequences in parallel while remaining flexible. On a Xilinx Alveo U280 it is up to 2.68x faster and 7.46x more energy efficient than Bowtie2, a state-of-the-art aligner, and it surpasses previous hardware implementations of KMP by up to 19.38x in throughput.
Authors: Beatrice Branchini, Sofia Breschi, Alberto Zeni, Marco Domenico Santambrogio
Published in: IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), 2022