Idea for next gauntlet

[jasong]

Maybe I'm just bugging you guys, but there is a wonderful group of people(about as good as you guys) at Riesel Sieve. Basically, they're trying to prove the Riesel conjecture throught brute force. The Riesel conjecture states that for every odd k of the form k*2^n-1 below 509203 there is an n that makes it prime. Supposedly, k=509203 is the lowest k that is ALWAYS composite.

Anyway, their sieving program is going to become undependable somewhere around 150 trillion and they're trying to knock down as many k's as possible before their programmer is forced to determine what certain aspects of the new sieving program will be. From what I can gather, the fewer k the final product has to be made for, the faster the sieve.

If you guys would run a Gauntlet with their LLR-Net client for even one or two weeks, it would help out TREMENDOUSLY. Also, sieving is still helpful(I think they're around 110 trillion at the moment), so "old"(non-FX or 64 cpus are still good for LLR I think[?]) Athlons could help too.

Here's the main site!!! (http://rieselsieve.com/)

[rsbriggs]

There doesn't appear to be a Free-DC team there....

[jasong]

Originally posted by rsbriggs
There doesn't appear to be a Free-DC team there.... Yeah, that's true. The Riesel conjecture is basically the Sierpenski problem(I want to say,"Seventeen or Bust," but I've never conclusively established the link) except it's -1 instead of +1, which means totally different numbers become difficult to crack.

Seventeen or Bust is more popular, but IMO, Riesel Sieve deserves more attention than it's getting. They've got 80 k's to find, while Seventeen or Bust is down to 10. Maybe I'm imagining this, but I'm thinking it would be easier to find a prime with 80 k's available than just 10, since keeping to the pattern established by the server would keep the n lower for the same number of tests.

Another possible project, in terms of math, is the Euler "something-or-other" project.
Basically they'll have a 3 number pattern assigned to you, like "2,1,3." This would mean you would have to find 3 whole numbers(1 on one side of the equal sign and 3 on the other) which, when squared, equal the same amount. Basically the numbers are: power,number of terms on one side, number of terms on the other. This project has no use that I'm aware of, but it might be fun. I've oversimplified the assignment process(there's no overlap in the actual application), but I think you get the basic idea.

[em99010pepe]

In fact there's a Free-DC team......

Team LLR Stats (http://stats.rieselsieve.com/teamllr.php)
Team Sieve Stats (http://stats.rieselsieve.com/teamsieve.php)

Carlos

[Pconfig]

I'd say go for rieselsieve, i needs more attention and at this point (low N) when they would have way more compu's there will be a prime found every week or so :)

Don't forget to pop on IRC: freenode.net #rieselsieve