(M48) NEW MERSENNE PRIME! LARGEST PRIME NUMBER DISCOVERED!

[chalsall]

Congratulations to everyone involved!!!

If anyone wants to help get the announcement on Slasdot (and have a Slashdot account) please go to http://slashdot.org/recent and "+" the "New largest known prime number found: 2^57,885,161-1" story.

I submitted the first paragraph of George's press release anonymously.

[ATH]

Originally Posted by Jeff Gilchrist

I just finished my Prime95 AVX double check and it also reports PRIME! Hopefully one of the non-Prime95 runs will finish soon for extra-extra confirmation.

You completed this just 112.5 hours after it was found which is 7.0ms/iteration at the most. Which processor do you have? How many threads it use?

How long after 23:30 UTC when it was found did you start your run?

[aketilander]

Which computer and processer was used for the discovery?

[Xyzzy]

If our (not-so-elegant) programming is correct, every six digit string is present in the decimal expansion.

#!/usr/bin/perl -w
use strict;
my $line;
open M, "48.txt";
        $line = <M>;
        chomp $line;
close M;
for my $a ( 0 .. 9 ) {
        for my $b ( 0 .. 9 ) {
                for my $c ( 0 .. 9 ) {
                        for my $d ( 0 .. 9 ) {
                                for my $e ( 0 .. 9 ) {
                                        for my $f ( 0 .. 9 ) {
                                                        my $str = join ( '', $a, $b, $c, $d, $e, $f );
                                                        print "$str\n" if $line !~ m/$str/;
                                        }
                                }
                        }
                }
        }
}

You can generate 48.txt with:

echo '2^57885161-1' | bc | tr -d '\\\n' | sed 's/$/\n/' > 48.txt

[kladner]

Add my congratulations to everyone!

Salud, dinero, amor, y tiempo para gustar lo!
(Here's to) Health, money, love, and time to enjoy it!

[NBtarheel_33]

Googling "48th Mersenne prime", I found that we are in something called "Wopular": http://www.wopular.com/48th-mersenne-prime-found

"Hacker News" links to the press release: http://news.ycombinator.com/

Interestingly, there's also a crank out there that swears up and down that he has found M48, M49, and M50 on an iPhone...laissez le bon temps rouleau!

[Batalov]

Originally Posted by ATH

Factors below 10¹¹:

(2^57885161+1)/3: 7061989643

2^57885161-3: 29, 167, 269

2^57885160*(2^57885161 - 1) + 1: 7, 22127627

Some more factors:
* (Mp+2)/3 has prime factors 7061989643 * 100271784664656667 (no other factors < 2e18)
* Mp-2 has prime factors 29 * 167 * 269 * 453527886599 (no other factors < 1e11)
* 2*Mp-1 has prime factors 193 * 1549 * 1707450949 (no other factors < 1e9)

[petrw1]

Originally Posted by NBtarheel_33

Interestingly, there's also a crank out there that swears up and down that he has found M48, M49, and M50 on an iPhone...laissez le bon temps rouleau!

What was the date? I need to know if it was before or after I found them on my Abacus and Double-checked them on my Slide Rule.

[Prime95]

Originally Posted by aketilander

Which computer and processer was used for the discovery?

CPU Model : Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz
CPU Speed, MHz : 2992
Software Name : Prime95
Software Version : WindowsService,Prime95,v27.7,build 2

[ixfd64]

Originally Posted by NBtarheel_33

Seconded. We have the following evidence:

1. 57881749 is not assigned to anyone... and there's no reason that it shouldn't be. 57881573, for instance, was handed out to ANONYMOUS yesterday. Other neighboring exponents have been handed out in recent days. If 57881749 were truly free, it wouldn't be for long.

2. 57881749 indeed has a curtisc pedigree; it was P-1'ed (the usual curtisc S1-only job) back in September. It is an almost certainty that this exponent was therefore assigned to curtisc for LL testing, as curtisc does not normally (ever?) perform P-1 for its own sake.

3. The link in (2) shows no known LL result for 57881749. So, based on the logic in (2), it should show up on a list of team assignments for curtisc, right? Well, I have an empty test account that belongs to curtisc's team, and I just looked at their (thousands of) team assignments. You guessed it...57881749 is nowhere to be found!

4. The milestone estimates/calculations detailed in this thread (and those that I have secondarily performed) indicate an exponent in the neighborhood of 57.8-57.9 million.

Four hours exactly to go...but I'm going to agree with ixfd64 and make my final guess that M48 = 2^57881749-1.

It turns out we were both wrong. But it's indeed strange for an exponent to have such a property. Methinks George did this just to throw us off. *cue the X-Files theme*

[chalsall]

Amusingly, this exponent passed through two ANONYMOUS assignments before finally being picked up by Curtis on 2012-12-12 -- one assignment on 2012-08-14 which then expired 60 days later, then another on 2012-10-13 which expired 60 days later.

Neither had done a single cycle on it. Or, at least, not reported.