Hmm. Not sure about the point.
It doesn't search the real data, right? (I checked). It happens to have a predefined set of data. So, how do we know if the speed comes from genuine qualities of the tech, not from a simple fact that the data set is small?