GIMPS just found the new largest discovered prime number. This was all done on an i7, not a supercomputer. Imagine if we looked for it on Tainhe-2, or even an advanced beowulf cluster? Why the hell don't they, is it not worth the recognition or the prize money?
Who is going to fund that shit?
Nobody gives a shit about large primes, even the RSA numbers are not even close to being that big, why would anyone fund using supercomputers for searching for large primes, when there's other interesting computational problems to solve in other domains? Like molecular biology.
Number theory autism just isn't that interesting, anon.
GIMPS was found by a distributed computing effort. Which is sort of like a Beowulf cluster, except that you don't need fast interconnect between the nodes.
Using a supercomputer or an actual cluster would be a waste of resources when you can just use thousands of disparate PCs instead.