Merry Christmas and a prime! (M50 related)

[chalsall]

Originally Posted by DanielBamberger

That's certainly what it feels like.

Just to share the joke, this is a standard question to young Buddhists. The answer is "silence".

I always answered with "autogratification".

For some reason I was never invited back.

[GP2]

Originally Posted by DanielBamberger

I wondered how strange that really is.

There are much more improbable things, for instance Six nines in pi

[airsquirrels]

Good news! It is prime!

That is to say, the 19760000th 64 bit residue is prime. As are 43/1527 of the first 31400000 20,000 interval residues.

[chalsall]

Originally Posted by airsquirrels

Good news! It is prime!

Extreme coolness!

[GP2]

Originally Posted by airsquirrels

Good news! It is prime!

That is to say, the 19760000th 64 bit residue is prime. As are 43/1527 of the first 31400000 20,000 interval residues.

With any luck, the final 64 bit residue will be prime too!!!

[R. Gerbicz]

Originally Posted by Madpoo

My theory is that we've missed a prime that will later be found by double-checking.

If we reckon that the error rate (so far) on first-time checks is around 3-4%, and we now have 50 Mersenne primes of ever increasing size, I think given that data it's safe to say that eventually a prime will most definitely be missed, sooner or later. My theory is that it's already happened.

And we won't see this in reality, because you set (manually?) these as Mersenne prime and delete/hide all LL/PRP results done, like for https://www.mersenne.org/report_expo...exp_hi=&full=1 , here we even don't see the first LL test on this exponent from Martin Nowak. I would list the first few (correct and errorneous) LL/PRP results.

[pacionet]

How long will the final checks take? When the exponent will be announced?

[retina]

Originally Posted by DanielBamberger

The question should be: Assuming you tossed 100 "heads" in a row, what's the probability ...

Already discussed before. It like de-ja-vu

http://mersenneforum.org/showthread.php?t=13314

[ATH]

Originally Posted by pacionet

How long will the final checks take? When the exponent will be announced?

All current tests will probably be done within 48h, but it will not be announced before Jan 3rd at the earliest depending on when it is best to do a press release.

[Madpoo]

Originally Posted by R. Gerbicz

And we won't see this in reality, because you set (manually?) these as Mersenne prime and delete/hide all LL/PRP results done, like for https://www.mersenne.org/report_expo...exp_hi=&full=1 , here we even don't see the first LL test on this exponent from Martin Nowak. I would list the first few (correct and errorneous) LL/PRP results.

I don't think there's anything in the report page that masks or otherwise hides other results for Mersenne prime exponents, like the one you used as an example.

That said, there is a system in place that will block additional "is prime" results from being accepted for any known primes, but that's just because people like to put in the known primes and run a test just to see what the output looks like. Accepting those would mean a lot of extraneous results for those.

I did just look in the database and saw that there's no actual entry for M25964951 in the LL results table. That probably just means that, at the time (this was on the old Primenet iteration, v4) it may not have automatically accepted "is prime" results and instead triggered the notification email.

Even with the current setup, an automatic result will be entered into the database, but someone trying to submit a new prime via the manual results page will trigger an exception that emails us the information but prevents it from being logged into the database. That's because manual results can come from GPU programs that lack certain validation bits and pieces, so someone could try spamming us with false positives.

In such a scenario (which has happened with the cudalucas false positive issue), we could see a new prime that never got an actual entry into the LL results table.

Yeah, I chalk the weirdness of those "missing" LL results for other known primes to a Primenet version 4 issue. The last 3 to be found all came in after 2008 when Primenet version 5 launched, and they all have LL result entries.

Maybe somewhere George has the raw results for the older prime discoveries that were manually sent or emailed to him and we can add them in just for completion.

Anyway, rest assured that if a prime was discovered by way of a double-check, any older "false negatives" that exist will in fact appear. It will work that way now, and I'll make sure that's the case if we ever do have that situation (since it'll prove my theory)

[ewmayer]

Originally Posted by Mark Rose

Well we already knew that from earlier posted information. Ernst posted it would take at 72 hours to complete a run at 4 ms/iter, meaning it would be at least around 64 million.

Yes, I inadvertently leaked a lower bound by giving those data without mentioning how long the DC in question had been going. If you were paying close attention, you'd have noticed that a later poster inadvertently leaked an upper bound, even before the 'starts with a 7' leak. That one was more subtle, but hey, we like to keep the local Sherlock-Holmesian sleuths on their toes, it adds to the fun.