when will the main prp reach 31337

[Keroberts1]

Subject is pretty self explanatory. Just looking for opinions. I figure with the discovery of one prime in the next 3-6 months we could be testing 13,400,000 N tests by this time next year.

[Joe O]

It will definitely take something. Discovery of a prime, or primes. Many more contributors. More, newer, faster computers. An algorithmic breakthrough.
At our current rate of progress, we will be testing 13,400,000 N tests in June of 2007.

[Keroberts1]

12,000 increase a day. 365 days in a year. 4380000 a year! We'll be there long before that. Als owe may not have see nthe full speed up from the new client yet and with the new client we have been seeing alot more new participants.

[Joe O]

Granted, my numbers were computed Dec 1st, and we have had a speedup since then. But, even the most optimistic date would be April of 2006 and September of 2006 much more likely. You also have to take into account that the computations get harder/slower as n increases. Double the n and the test takes four times as long if the complexity is O(n^2). Eight times as long if it is O(n^3). I don't remember what Dave concluded, I will try to find his post, but let's take O(n^2) as a rough guideline.
We are nearly at 8M for n now. At 12M for n, the tests will take at least 2.25 times longer.
Your other post, about PRP poses a very good question that applies here. "How can we speed things up?" P-1 would do that. Hopefully, Louie will have time to incorporate the recent speedups into the P-1 code. Yes, we could start using Prime95 for P-1, but I think that we would have to be careful. I don't think that it chooses optimal bounds for B1 and B2 for S|B. Louie redid the computatation for the P-1 program he released to better reflect S|B. Perhaps, Louie is incorporating a P-1 test in V3.
You were looking for opinions, this is mine.

[vjs]

I'm thinking testing n=13m numbers with the 31337 account at this point is pointless. When that particular account was setup, a prime at that level would have yeilded the worlds largest prime.

Now although a prime at n=13m would be great it wouldn't be as benifital (from a PR stand point) as it was before. I believe we should really totally eliminate that account altogether...

If a new high-n account were created I would make it a n=33.4m account. A prime at this level would be a world record prime and actually worth some money to someone.
Also we might get a couple non-sob people coming to test our very large prime candidates.


IMHO

[royanee]

It's actually not a good idea to try the larger numbers though, because it's very possible that it will be sieved out, and the larger ones take a whole lot longer. I do think that we should get rid of the 31337 account though.

[Keroberts1]

if the new sieve client works out that will also speed the project up considerably. Does anyone know if there's been anyword on that?

[Mystwalker]

b2 just wrote a status update here: http://www.rieselsieve.com/phpBB2/viewtopic.php?p=4435#4435

[vjs]

For those of you who don't have access to the site, username etc....

<copy/paste>

From b2
Site Admin

I had to make a few decisions last night on exactly what the sieve can assume and not assume. The memory bitmap is where most of the speed problems lie. When we make certain assumptions about our numbers...we can tweak out some speed improvements and cut the page faults completely out. Memory allocation is being worked on now.

Expect the executeable in the next week or so...but it won't be the super duper sieve that we will eventually release...more of a step between what we have now..and what we will have in the future.

But FASTER :)

<end>

[royanee]

Yay! :)

By the way, thanks for the copy/paste vjs.

[vjs]

I've created a new thread in the sieve section regarding this topic....

[Keroberts1]

could the main PRP reach the 31337 I know many of you were skeptical about that happening be the end of 2005 but i believe this revives hope that it could. Not that it mens anything it'll just be a nice thing ot count down towards. Plus we get a nice 7-10 % sieve boost at the same time.