At 11:44 AM 18/07/99 -0700, George Woltman wrote:
>2) Me. Some would argue that I share some of the credit for any
>Mersenne discoveries. Be aware that I will donate any share to charity.
I think that giving some to George is a great idea, with one change: Buy
him a new computer with some of the money.
I distinctly remember him telling me that he couldn't look into Pentium
III optimizations because he didn't have a Pentium III.
Not only would the most deserving person (IMHO) get some of the prize,
but it would be in such a way that it benefitted GIMPS as well.
>3) Scott Kurowski. He has real expenses in running the PrimeNet server
>that should be reimbursed and has been instrumental in GIMPS' growth.
I don't think that this would be a good idea. I can't speak for Scott (I
can barely speak for myself ;-) ), but it seems to me that entropia
benefits from GIMPS mostly due to the publicity it gets.
I'm sure the publicity would be far more favorable if he can say that
entropia is _donating_ its services, rather than having to admit that he is
getting some of the money.
>7) Anyone that makes a mathematical or algorithmic breakthrough that
>speeds up the search process. I'm talking about a doubling in search speed
>not a 1% speedup in assembly code.
I think that this would be great -- but I seriously doubt that any
improvement will be found. We can't get any better with FP FFTs, since we
don't have any zero-padding any more, and you've specifically disallowed
implementational improvements. A switch to integer FFTs might be better
for huge FFT lengths, but integer arithmetic is currently very slow
compared to FP arithmetic, so I doubt that will end up helping.
Really, the only hope I can see for a significant improvement is a really
fast implementation of Schonhage-Strasen multiplication, but
Schonhage-Strassn is almost infamously slow.
Colin Percival
PS. As for the idea of buying computers just to search for primes, GIMPS
currently has maybe $5 million of computers working on it. Whatever could
be bought would probably have less impact than a single article in a major
newspaper about GIMPS. (For reference, when the NY Times wrote about my
computation of the 5 trillionth bit of Pi, I got over 200 new volunteers).
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
