NEW PRIME DISCOVERY!!!!

[jjjjL]

k=5359 is DONE! :)

Check out our announcement at http://www.seventeenorbust.com/

5359*2^5054502 + 1 is prime.

The prime was reported several days ago and we have just completed a full proth test to prove the number's primality.

We have not been assigning new tests for k=5359 from the site since the day after we discovered the new prime while we waited for the verification. I kind of expected to see someone on the forum notice this and start rumors but if anyone noticed, you didn't say anything publically.

The problem with next.txt was related to this. Mike - It has been fixed for awhile if you want to use it again otherwise you can just use the extended range stats.

Speaking of which...

Sievers: Now you only have to sieve 11 k's instead of 12! :) I haven't made a new .dat file yet. If MikeH can put a new one on his site, that would be good. Also, when you make a new DAT file, please make only one that does BOTH 300k - 3M and 3M-20M (dual). Then if the people who are updating the sieve threads could lock the old reservation pages, and make one, unified page that has a link to the single DAT file, and tells people to update for the switchover to 11k, that would be great. Awhile ago, there was talk of updating to a DAT file with 1M - 20M. That might be wise considering that all the numbers n < 1M have been double checked with SB. There should probably be some discussion on the sieving forum before you proceed so that everyone can decide what's best. I'm just advocating that we make a single sieve to avoid having to post two messages to reserve ranges going forward.

Also, I am not going to stop people from submitting factors for k=5359 (at least for awhile) but I think the sieve stats should stop giving credit to those factors in a few days.

Factorers: Those of you running the P-1 factorer can do whatever you want to modify the dat files. Just make sure to remove k=5359 so you don't spend time checking those values anymore.

Anyway, good job everyone! All our participants hard work continues to make SB a success. 6 down, 11 to go! If you haven't yet, go checkout the main announcement for details about the discovery at www.seventeenorbust.com

Cheers,
Louie

[Alien88]

That's also why I didn't respond to some of the posts because I didnt want to give it away that we had found a prime until it was verified!

Good work guys!

[Keroberts1]

well december definatly seems to b e a good month for us

[Frodo42]

Finally :D ... congratulations everyone


Factorers: Those of you running the P-1 factorer can do whatever you want to modify the dat files. Just make sure to remove k=5359 so you don't spend time checking those values anymore.

I just removed the two lines saying k=5359 and then some number in the next line, I guess this gives 4847 a lot more weight, but only on the other side of 20G or?
P-1 does not make any tests for 5359 anymore, but I can't tell if the tests for 4847 are the right ones ... I'm not quite sure I figures out the SoB.dat file the right way.

btw. I had a suspicion that maybe a prime was underway with all the weird stuff happening but did not want to cry wolf and I guess I had more or less given up hope that it would show up anytime soon.

[Keroberts1]

i believe proth sieve is programmed to only run with a 12 K dat file. if that is not true then how do you modify the dat file to work for 11 K's I tried deleting everything from k=5359 to k=10223. That just gave an error.

[Nuri]

I was wondering why there were no new secret tests assigned at 5359. :D

That's great news. Well done everyone. :cheers:

BTW, as far as the sieve and P-1 are concerned, I think it would be the best course of action to wait for the official dat file release. That would prevent any problems that might occur.

[ceselb]

Great news. :cheers:

It is true that proth_sieve is hard coded only to work with 12 k's, but I'm sure mklasson will release a new version shortly.

Good work everyone.

[b2uc]

After Mikael changed a few things around in Proth_sieve for us at Riesel Sieve I believe the hardcoded k's were taken out..since we have 99. I believe the new versions will look for a sob.dat or riesel.dat and choose which formula to run. So I believe the newest version may work without the 12k's

If someone would delete the k and change the first line to 11 and test it out..it may just work. Of course this throws MikeH's factor density checking into the tweak bin.

Congrats to all...it's one hell of an accomplishment.

Lee Stephens
B2
www.rieselsieve.com

[Nuri]

I guess, still it has to be done by Mikael. Riesel numbers have the form k*2^n-1 whereas Sierpinski numbers have k*2^n+1. I guess this needs some changes (although probably minor) be done in coding as well.

[Nuri]

The remaining test stats needs to be updated. ;)

Remaining tests n < 6000000 n/a 28704 tests
Remaining tests n > 6000000 n/a 470356 tests

[ceselb]

Nuri: Just checked, the new version of proth_sieve can do both. We're all set. :D

[b2uc]

Ceselb..Is there any speed differences?...try running it with 12 k's still in. Didn't know if he kept or dropped the SoB tweaks in the newest version that can do either kinds of numbers.

[ceselb]

Tried to run the new version with the old dat. Same speed as far as I can see, in any case not more than 0.5% lower.

[b2uc]

Ok..now what are the speeds with 11 k's:)

Boy I'm happy for you guys...however...Mikael already took care of his end..now it's MikeH who has to work overtime for that gap analysis to work out..hehe Might be the perfect time to also take the sieved factors out.

[Samidoost]

Long have we waited. And now, at last, the waiting has paid off.
5359*2^5054502+1 is prime!!!
Congratulations to all the participants of the SB.
Specially to Randy Sundquist and his Team ExtremeDC, Louie and Dave.

Now the SB can be really proud of discovering finally a really big and really important prime. Now a Keller Proth number is the greatest non-Mersenne prime.

:|party|: :drums: :elephant: :|party|:

Waiting for k=28433 and k=4847:thumbs:
Payam:D

[ceselb]

Just tried with 11k's. Went from 215 to 222kp/sec, about 3%.

And as I've mentioned in another post, removing the factors so far made a 2.5% increase. Rasing the lower bound to 1M may also give a slightly faster rate.

[Keroberts1]

mine crashes as soon as it finishes initializing. I don't know what i did wrong exactly waht needs to be done to alter the dat file?

[b2uc]

Originally posted by Keroberts1
mine crashes as soon as it finishes initializing. I don't know what i did wrong exactly waht needs to be done to alter the dat file?

1. Run the new client
2. Change the 12 to 11 in the first line
3. Delete the k=5359 and the n's that belong to it.

[ceselb]

You get a crash (divide by zero) if you don't change 12 to 11 in the first line.

[Keroberts1]

i did all of those are you all using version 0.37? has there been anything new released recently?

[ceselb]

Yes, 0.39 (http://n137.ryd.student.liu.se/proth_sieve.php) is available. 0.37 can't handle anything else than 12 k's.

[Keroberts1]

got it i missed that one

[Death]

If I can only connect through port 3128.

[Mystwalker]

Very good! :thumbs:

The dry spell is finally over. :|party|:

Hope this is a good omen for ppl with less patience. ;)

[Death]

can server tell to client when receiving intermediate blocks to drop current test?
there's ~100 or so pendind tests at 5359...

[mklasson]

Congratulations!

Here's to hoping this will give the project a little boost,
Mikael

[jjjjL]

The server won't actively try to kill pending tests for k=5359. No new ones have been assigned in over a week so there aren't that many. If you have one, feel free to drop it. The server is already configured to not reassign it.

In the end, canceling a few dozen half-done tests won't save too much project power. Preventing them from being assigned that last week was enough. If you want to increase the project throughput, run the sieve (http://www.free-dc.org/forum/forumdisplay.php?s=&forumid=51) a little so we don't have to test as many numbers in the future.

Cheers,
Louie

[Firebirth]

Couldn't help noticing that this was the 2nd most time consuming k for primality-testing! - Couldn't be much better!

[Keroberts1]

look at all of the new users today... This is great now all we need is the P4 fix which i believe is on the way. We're definatly on the high road again.

[expinete]

I think an email should be sent to all the members. This new prime could bring back some old members.

[Alien88]

Originally posted by Death
If I can only connect through port 3128.

Would you like me to setup sbp.pns.net to listen on port 3128?

[Slatz]

Originally posted by Keroberts1
look at all of the new users today... This is great now all we need is the P4 fix which i believe is on the way. We're definatly on the high road again.

congrats all...yeah..nice boost in the new members
we've got it posted on AnandTech's homepage hoping to encourage some more members for our TeAm!!!
linky (http://www.anandtech.com/index.html)

Slatz

[Alien88]

Also, This prime is also the largest non-mersenne prime on record.

[Death]

Originally posted by Alien88
Would you like me to setup sbp.pns.net to listen on port 3128?

OMG! You can do this? :notworthy

I'm behind our corporate firewall. And no connections through the 80 or 23 ports =(.

[Alien88]

Originally posted by Death
OMG! You can do this? :notworthy

I'm behind our corporate firewall. And no connections through the 80 or 23 ports =(.


Give it a try.. sbp.pns.net is listening on ports 80, 8080, 22, and 3128.

[Keroberts1]

Has MikeH been by since the prime was found? Just wondering if he's going ot be releasing a new dat file soon. I managed ot figure out how ot get mine running wiht 11 k values but i'd still like to get the official SoB.dat if yo uare going to be removing sieved values and low p values. Just trying to keep up with whats going on

[Death]

Originally posted by Alien88
Give it a try.. sbp.pns.net is listening on ports 80, 8080, 22, and 3128.

thats not a port. =((



C:\Documents and Settings\dead>tracert 216.163.34.106

Tracing route to sbp.pns.net [216.163.34.106]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms 10.44.54.5
2 10.7.0.1 reports: Destination net unreachable.

Trace complete.


at the same time



C:\Documents and Settings\dead>ping sbp.pns.net -a

Pinging sb-proxy.pns.net [216.163.34.106] with 32 bytes of data:

Reply from 10.7.0.1: Destination net unreachable.
Reply from 10.7.0.1: Destination net unreachable.
Reply from 10.7.0.1: Destination net unreachable.
Reply from 10.7.0.1: Destination net unreachable.

Ping statistics for 216.163.34.106:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 0ms, Average = 0ms


I give up. :cry:

[MikeH]

I'm out of town for two days and I miss all the fun :rolleyes:

Congratulations to everyone that has helped contribute to this magnificent discovery. Great Christmas present to everyone. :|party|: :cheers: :thumbs:

I'll take a look now at putting an 11k sob.dat file together, but this has been a seriously long day, so more likely I'll sort it tomorrow. I'll also check with ceselb's suggestion of shrinking the .dat file, and of lifting the base from n=300K to 400K, 500K, 1M (I'm open to suggestions, I'll try a few tests).

Right now I'm still scoring for 5359 factors. Hopefully that should be sorted tomorrow (if not then weekend), but either way I'll back date and freeze all 5359 scores from the day that Louie announced the discovery (that seems fair to me).

[Death]

Originally posted by Alien88
Give it a try.. sbp.pns.net is listening on ports 80, 8080, 22, and 3128.

It seems that's there's something strange happens.

Client can't connect to server. No ping or tracert to sbp.pns.net.
But IE, that users proxy receive this -



  


I believe that this is a block for client.

How can it be?

[MathGuy]

ping (and for the most part tracert) uses ICMP and can be blocked by a firewall (or router or...) completely independently of any particular ports. In your case, 10.7.0.1 (which is probably some local corporate asset - router, firewall, etc.) is refusing to pass the ICMP packet on. This doesn't mean, however, that you can't connect on any correctly forwarded port. In short, "just because you can't ping it doesn't mean you can't connect to it."

What you have to do to connect through a firewall is make sure that the local (client) traffic is wrapped in such a way that the firewall is willing and able to re-wrap it and forward it to one of the ports that the server is listening on.

If you want some more help dealing with this, e-mail me at kerryj@math.bsu.edu and I'll do what I can to help. Several of my clients routinely connect through various firewalls and proxies so I have some experience along these lines...

[Ken_g6[TA]]

Originally posted by Death
OMG! You can do this? :notworthy

I'm behind our corporate firewall. And no connections through the 80 or 23 ports =(. I believe he's behind an HTTP proxy, not just a normal firewall.

Edit: MystWalker has a suggestion for HTTP proxies here (http://www.free-dc.org/forum/showthread.php?s=&threadid=4832).

[biwema]

By the way


The server won't actively try to kill pending tests for k=5359. No new ones have been assigned in over a week so there aren't that many. If you have one, feel free to drop it. The server is already configured to not reassign it.

Are the numbers with k=5359 and an exponent smaller than the new prime going to be tested?
I think it makes sense in order to prove that the new prime is the smallest one with k=5359. That could be useful for statistical purposes.

[wblipp]

Originally posted by biwema
Are the numbers with k=5359 and an exponent smaller than the new prime going to be tested?
I think it makes sense in order to prove that the new prime is the smallest one with k=5359. That could be useful for statistical purposes.
But to really be sure, we would have to double check every exponent below the new prime, just like GIMPS must double check each exponent to be sure they haven't missed any primes. The smallest primes are called Keller Primes, so I suppose that finding them would be the Keller Problem. It would be about twice as much work to solve the Keller Problem - I don't think we can spare the resouces.

[jjjjL]

Originally posted by wblipp
But to really be sure, we would have to double check every exponent below the new prime, just like GIMPS must double check each exponent to be sure they haven't missed any primes. The smallest primes are called Keller Primes, so I suppose that finding them would be the Keller Problem. It would be about twice as much work to solve the Keller Problem - I don't think we can spare the resouces.

I agree. Although this new prime is in all likelihood a Keller prime for the series 5359*2^n+1, it is not in our interest to make sure that it is.

I think double checking resources should first be put to work checking numbers for which no prime is known.

Louie