The code is prepared to execute on the H100. Just executing runAll script should be enough. It is important to use CUDA 12.3 or greater.
This repository includes the DPX microbenchmark code and the NW and DTW implementations for CPU and GPU