Super fast implementation of ICP in CUDA for compute capable devices 3.5 or higher - View it on GitHub