This library provides the GPU-SW algorithm, a GPGPU-enabled classical pairwise alignment one.
The dependency in this project is ArrayFire (for the GPGPU). It could be thought as a wrapper of CUDA, OpenCL, and CPU multithreading, providing a unified programming interface and releasing us from optimizing for each GPU. Although it doesn't need to write kernel code, the GPGPU speed achieves to/is faster rather than one obtained by writing the native code.
Copyright (c) 2017 Heartsh
Licensed under the MIT license.