ABSTRACT
The gene is an ordered pair of bases in DNA, so the permutation of these bases may explain some biological reactions scientifically. The sequence of genes plays an important role in human development. Given the importance of gene sequencing, we need to increase the speed of this gene sequencing and matching process. Gene sequencing is a large-scale project with huge data, and we have got a good treatment to deal with the huge data of this gene sequencing. Big data has four typical characteristics: volume, velocity, variety, and value, which bring difficulties to the structural design of the computer system. The traditional computer systems are ill-equipped to meet the typical demands of big data applications, so a heterogeneous computing system based on graphics processing unit (GPU) and Field Programmable Gate Array (FPGA) has emerged as an effective framework for processing big data applications. Although the utilization of cloud computing platforms and GPU platforms can accelerate the execution of gene sequencing algorithms, both approaches still possess their respective limitations. In the case of cloud computing platforms, maintaining cluster operation requires a lot, and from the perspective of efficiency, the calculation efficiency of an individual node is not very high. Each GPU chip consumes a significant amount of power and thus consumes a lot of power to process the task. In this chapter, we, according to the knuth morris pratt (KMP) and burrows wheeler aligner (BWA) algorithms, design an accelerator to accelerate the gene sequencing. The accelerator is intended to have broad applicability and low power costs. Experimental results demonstrate that compared to central processing units (CPUs) the accelerator achieves a speedup ratio of 5x while consuming only 0.10 W of power. Compared with other platforms, we make it a balance between speedup rate and power cost. In general, the implementation of this study is important to improve the acceleration effect and reduce power consumption.
