mersenneforum.org - Newer milestone thread

mersenneforum.org (https://www.mersenneforum.org/index.php)

- Data (https://www.mersenneforum.org/forumdisplay.php?f=21)

- - Newer milestone thread (https://www.mersenneforum.org/showthread.php?t=13871)

[QUOTE=NBtarheel_33;435561]Taking a look at the old [URL=http://www.mersenne.org/report_classic/]"colorful stats report"[/URL]: Only [SIZE="3"][B]19[/B][/SIZE] Mersenne numbers need factored to reach a total of 3,000,000 factored Mersenne numbers in the "classical" exponent range of 0-79.3M.

*Your* factor could be number 3,000,000![/QUOTE]

And in the full report to 1 billion if you sum the factored column there are a total of 28,859,765 exponents factored (June 4th 11pm UTC) out of the 50,847,534 prime exponents below 10[sup]9[/sup] (56.76%):

[url]http://www.mersenne.org/primenet/[/url]

[CODE]
2016-06-04 11pm UTC:

Primes: 49
Factored: 28,859,765
LL-D: 857,782
LL: 640,360
LLERR: 369
NO-LL: 20,489,209
--------------------
Total: 50,847,534
[/CODE]

62,000 in the first 1,000,000 range factored.

[url]http://www.mersenne.org/primenet/[/url]

As we get closer to finishing the < 68M range, I noticed something weird. Out of the 16 remaining, about 12 of them are probably going to expire before finishing.

I ran my little query to analyze their daily rate of progress and you can disregard the ETAs being reported for most of them... I have *real* ETA's projected at 100+ days (or even 221 days in the case of M67820569 based on an observed 0.28% per day rate).

I mean, yeah, like that one will expire in 41 more days, but geez.

On the bright side, it was assigned before the new rules went into effect. However, about half of the ones I'm looking at were assigned *after* the new rules, but the machines in question are poking along like M67739429 (0.63% daily) or simply stopped checking in like M67673351 (which would have finished 23 days ago if they hadn't given up or whatever...it was doing 8.84% daily up 'til then).

Anyway, just thought I'd throw that info out. It doesn't really mean a whole lot in regards to the new assignment rules since we have some in there that predated those, but it was still weird that so many out of the remaining were in the "likely to expire" category.

[QUOTE=Madpoo;439859]... it was still weird that so many out of the remaining were in the "likely to expire" category.[/QUOTE]I thought it was expected behaviour. The last few in any category will always be the ones that are dragging their heels.

[QUOTE=retina;439863]I thought it was expected behaviour. The last few in any category will always be the ones that are dragging their heels.[/QUOTE]

Indeed. There will always be those who abandon for whatever reason. The new rules mean that when it happens, the LL test will almost certainly be completed by the next person who gets it, not that LL tests will always finish.

[QUOTE=retina;439863]I thought it was expected behaviour. The last few in any category will always be the ones that are dragging their heels.[/QUOTE]

Yeah, it was just odd to see that half of those I think will expire were assigned [B]after[/B] the new rules were in place.

Oh well... at least they won't be getting any more cat 0/1 stuff for a while if that happens.

This day in GIMPS.....

I looked back in my records of the summary page: [url]http://www.mersenne.org/primenet/[/url] from back to 2008 almost back to when the v5 server started. I looked specifically at today Aug 23rd at 18:00 UTC and summed up the numbers with a program.

[CODE] 2016-08-23 2015-08-23 2014-08-23 2013-08-23 2012-08-23 2011-08-23 2010-08-23 2009-08-23 [COLOR="Red"][B]2008-11-20[/B][/COLOR]
18:00 UTC 18:00 UTC 18:00 UTC 18:00 UTC 18:00 UTC 18:00 UTC 18:00 UTC 18:00 UTC [COLOR="Red"][B]10:00 UTC[/B][/COLOR]
---------------------------------------------------------------------------------------------------------------------------------------------------
Primes: 49 48 48 48 47 47 47 47 46
Factored: 28923338 28576259 28291984 28021026 27871933 27686192 27462158 26245385 25682322
LL-D: 879314 791072 742738 700969 657589 611738 557954 518800 492556
LL: 638188 642571 610467 557144 506861 488761 480513 452520 425328
LLERR: 327 406 568 784 1035 1353 2099 3025 5530
NO-LL: 20406318 20837179 21201730 21567564 21810069 22059443 22344763 23627757 24241752

Total: 50847534 50847535 50847535 50847535 50847534 50847534 50847534 50847534 50847534
[/CODE]

Note that for 2008 the date is Nov 20th at 10am UTC, that is the earliest full summary page I have.

Notice that for 2013, 2014 and 2015 the total 50847535 is 1 too high compared to the 50,847,534 primes under 1 billion. I looked for errors in my summing program but I could not find any, and I did not bother adding all the numbers by hand, so I assume there was 1 too many in one of the columns on the summary page back then.

1944-08-23 [URL="http://ireport.cnn.com/docs/DOC-484553"]Romanian Insurrection[/URL] :razz:
(sorry, could not stop myself when I saw those dates!)
/offtopic

There could be a milestone coming up in the "weird and wacky" category. Maybe an award for "longest running and useless test result"

[URL="http://www.mersenne.org/M26553437"]M26553437[/URL]

There's an expired assignment that was assigned that exponent back in May of 2011, it was picked up (poached?) by someone and finished DC in Nov 2012, the assignment itself technically should have expired once the DC was done, but it looks like it was actually expired when the grandfather rules were implemented back in Jan/Feb 2014.

Despite all of that, the user has continued plugging away at it and is now 95.8 % done.

His machine is predicting an ETA of Aug 26 but it's probably one of those cases where it keeps saying "I'll be done soon" even though soon never seems to come.

My actual analysis of the progress can't project an ETA because it actually hasn't gone up even a single 1/10 of a percent for the past several months... but hey, it keeps checking in so at least it's running, even if it might only do 0.1% every 6 months?

The same user is also doing 27384901 and 27385009 which actually have logged enough progress to say that it's doing 0.001% daily, so they're estimated to finish in 2041 to 2049. Set your alarm clocks.

There's actually an older machine that's still checking in for these exponents: 25413671 and 25413779

But unlike that other one, it hasn't actually started yet (or it's still under 0.1% done?) so it's not as much fun. The curious thing about that one is the CPU running them: "Intel(R) Core(TM) i7-4702MQ CPU @ 2.20GHz"

You'd think it would be a little further along by now. :smile:

The number of LL decreased this year for the first time. This is due to an increase in LL-D.