I have a large set (million) elements. Can node-kmeans process this data ? And how long to resolve this ?