https://hgpu.org/?p=2084
Scalable and highly parallel implementation of Smith-Waterman on graphics processing unit using CUDA