It is a sign that there is something wrong with the methodology. Well crafted C should hit the performance limits of the machine. If anything rates faster than C, then there is something going on that the author does not understand. This point of comparison is like a sanity check on the whole methodology, and it fails.