At 04:38 15-10-98 PDT, you wrote:
>
>hi,
>
>this is going to sound like an urban myth, but at least from my
>understanding, it is not.
>
>A friend of mine is a computational number theorist and she knows some
>NSA-types quite well. She told me that they (I'm not sure who "they" is)
>have been doing some work on Mersenne primes. She is often telling me
>that they are so much more advanced than the "outside world" in
>basically every field of computational number theory (among other
>subjects).
Hum, this looks suspicious. As I once said to a friend, there is much more
mathematicial working in the open field than for NSA. In the last Scientific
American, there was a comment about a encryption scheme wich was developed by
the private sector which was better than some propose by the NSA.
IMHO, those people cannot be related to the NSA which is probably the
most secret institution in the world!! A NSA worker will certainly
never speak about his job.
>She told me that they have made some vague statements to her about how
>they (this "they" again) have found (at least) two new Mersenne primes.
>The exponent of one of them is supposed to be around 5,200,000 while the
>other one is around 18,000,000.
It can be done the hard way if you have a computer powerful enough. Basicly,
if you have a teraflops computer available, scanning up to 5 200 000 would
take
about a year. However, going to 18 000 000 would take about ten years even
with such a computer.
I am not convince than NSA would be involve in such effort. Or, maybe they are
simpy running their own version of GIMPS on theirs computers!
Yvan Dutil
