(M48) NEW MERSENNE PRIME! LARGEST PRIME NUMBER DISCOVERED! - Page 13

Batalov · 2013-01-28, 22:27

Quote:

Originally Posted by Tony Reix

Wonderful ! A new candidate Mersenne prime ! Nice work, guys !

I do not have access to any big Bull computer for checking this new candidate. But others should have access to big machines, for sure. How many days does it take today for checking such a big candidate ? maybe I have to read all the previous comments in this thread... I'll try to read them.

Hey, hey, Tony!
Nice to hear from you again.

Yes, sometime by Friday we will be done. (Even Wednesday with CudaLucas, fingers crossed.)

frmky · 2013-01-28, 22:37

Still need a CUDALucas run? I can run it here on a GTX 480.
Edit: guess not.

LaurV · 2013-01-29, 02:08

Quote:

Originally Posted by Prime95

I don't think GIMPS' reputation was not tarnished...

Told ya... double negation is tricky in English..

Rodrigo · 2013-01-29, 03:20

Quote:

Originally Posted by LaurV

Told ya... double negation is tricky in English..

That isn't not the truth.

philmoore · 2013-01-29, 07:46

From another thread:

Quote:

Originally Posted by Prime95

Yes, this will be the 8th heads.

For you sleuthers out there: Although the average ratio between successive Mersenne prime exponents is predicted to be 2^(e^(-gamma)), about 1.47576, the expected ratio to which you would expect to search to have a 50% chance to find the next Mersenne prime is about 1.30964, which translates, when multiplied by 43112609, to the next Mersenne prime having an exponent below 56,462,000. I could be wrong about George's interpretation, but if I am right, this certainly narrows the range.

Phil

(Too lazy to sleuth on his own, or else too many papers to grade, but if you figure it out, I would appreciate a PM. I promise to be discrete and not contact SlashDot.)

flashjh · 2013-01-29, 19:31

<24 Hours for the CUDALucas run

ewmayer · 2013-01-29, 19:42

Quote:

Originally Posted by philmoore

For you sleuthers out there: Although the average ratio between successive Mersenne prime exponents is predicted to be 2^(e^(-gamma)), about 1.47576, the expected ratio to which you would expect to search to have a 50% chance to find the next Mersenne prime is about 1.30964, which translates, when multiplied by 43112609, to the next Mersenne prime having an exponent below 56,462,000. I could be wrong about George's interpretation, but if I am right, this certainly narrows the range.

But that is at best an average property having zilch (or very nearly zilch) to do with the magnitude of any given new prime found.

I will say only that enough data have been leaked in this thread to narrow the exponent to within 1%, without any knowledge of GIMPS server traffic and such. But you have to learn how filter out the obvious noise, and thread sub-discussions which actually take things farther away from the correct direction.

I will explain after the official announcement.

dabaichi · 2013-01-29, 19:45

Quote:

Originally Posted by flashjh

<24 Hours for the CUDALucas run

Great! Will we know the exponent tomorrow?

chalsall · 2013-01-29, 19:46

Quote:

Originally Posted by dabaichi

Great! Will we know the exponent tomorrow?

Only when George decides. Probably not tomorrow.

flashjh · 2013-01-29, 19:47

George will have my results tomorrow; I know Jeff will be done with the AVX run @~ the same time (I think he's ahead a little). As for the official release...

NBtarheel_33 · 2013-01-29, 22:33

Quote:

Originally Posted by Tony Reix

Wonderful ! A new candidate Mersenne prime ! Nice work, guys !

I do not have access to any big Bull computer for checking this new candidate. But others should have access to big machines, for sure. How many days does it take today for checking such a big candidate ? maybe I have to read all the previous comments in this thread... I'll try to read them.

For now, I'm home, waiting for last medecine exams, since I had a (TRANSIENT !!) cerebral vascular accident 3 weeks ago. For now, they found no explanation... since my 54 years body is in perfect shape, the body of a sportman they said.

While waiting for news about checking this candidate, you could have a look at my last hobby : photography.
Look at: http://www.skeptic.com/ .
You may appreciate different kinds of photographs, like Paris or ... nice ladies.

Tony

Long time, no see! Good to see you here again, Tony! It has certainly been a very long wait for a new prime, and if you have been around lurking on the forum, you know about all of the false positives we were getting for a while...very frustrating!!

I'm guessing you're still working @ Bull. I guess the hardware is really getting more and more impressive with each passing year. Those 16-core Itanium systems are probably way out-of-date by now. GIMPS itself recently went above 120 TFLOPS sustained.

And how about Curtis Cooper? Three Mersenne primes! He has certainly put himself and his university in the annals of history.

2013-01-28, 22:37	#134
frmky Jul 2003 So Cal 4104₈ Posts	Still need a CUDALucas run? I can run it here on a GTX 480. Edit: guess not. Last fiddled with by frmky on 2013-01-28 at 22:37

2013-01-29, 19:31	#138
flashjh "Jerry" Nov 2011 Vancouver, WA 1,123 Posts	Almost done <24 Hours for the CUDALucas run

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
I think I discovered new largest prime	harrik	Miscellaneous Math	8	2020-07-14 20:06
Number of zero's in largest prime...	Heather	Math	90	2006-04-01 22:06
Largest known prime not Mersenne?	Orgasmic Troll	Lounge	5	2005-12-23 22:00
Sixth bigest prime (no Mersenne's) discovered!	gbvalor	Math	4	2004-01-08 20:52
New largest prime number???	McBryce	Lounge	39	2003-08-12 19:35

2013-01-29, 19:47	#142
flashjh "Jerry" Nov 2011 Vancouver, WA 1,123 Posts	George will have my results tomorrow; I know Jeff will be done with the AVX run @~ the same time (I think he's ahead a little). As for the official release...