When that was written in 2010 it probably was the fastest grep. It's not like GNU grep is completely devoid of any performance optimisations.