A NVidia CUDA k-mean's implementation using the Thrust library. The implementation uses scatter-gather and prefix scans. The project was developed as a part of the Parallel & Distributed Systems in Aristotle Univeristy of Thessloniki Unfortunately, it is quite old and a bit messy and needs some refactoring.
krishnaiitd/kMeansCUDA
A NVidia CUDA k-Means implementation for the Parallel & Distributed Course 2009
C++