Newer milestone thread - Page 3

ckdo · 2010-12-09, 19:00

Quote:

Originally Posted by ckdo

November 12, 2010: The trial factoring wavefront hits 79M.

That's 26 days for 1M, again.

December 9, 2010: The trial factoring wavefront hits 80M.

That's 27 days for 1M this time.

Primeinator · 2010-12-11, 08:51

Quote:

# Countdown to testing all exponents below M(32582657) once: 1
# Countdown to testing all exponents below M(37156667) once: 1

NBtarheel_33 · 2010-12-11, 08:59

Only problem is that this single exponent (which is now holding up *two* milestones) is being run on a P3 by a user who updates only every 28 days, and is not due to finish the test until May of next year...

...yeah, yeah, I know, 100.00000001 is *still* faster than 100, and all that jazz.

Uncwilly · 2010-12-11, 09:04

I suspect that there will be a real early double check of that number completed soon enough.

davieddy · 2010-12-11, 09:29

Even allowing for the 10M digit prize distorting steady
progress somewhat, you might hope that the person entrusted with this
(now rather special) LL test would be sufficiently conscious of
its significance to (at least) keep his machine running 24/7.

David

NBtarheel_33 · 2010-12-11, 09:42

Quote:

Originally Posted by davieddy

Even allowing for the 10M digit prize distorting steady
progress somewhat, you might hope that the person entrusted with this
(now rather special) LL test would be sufficiently conscious of
its significance to (at least) keep his machine running 24/7.

David

It looks like this user is a very casual GIMPSter (not that there is the least bit wrong with that!). The computer on which "the" LL test is running is called P3-900 (likely a 900 MHz Pentium III). It has one other completed result to its credit, namely 25126039, ostensibly a double-check. The user *does*, however, own or have access to what is likely a Quad, which turned in 4 exponents in the 45M range between April 18 and May 8 of this year.

I guess it might be considered prodding if George were to e-mail him/her and let them know that they are 93% of the way to their place in mathematical history (proving the order of not one, but two Mersenne primes!)...

ckdo · 2010-12-11, 11:17

Quote:

Originally Posted by NBtarheel_33

I guess it might be considered prodding if George were to e-mail him/her and let them know that they are 93% of the way to their place in mathematical history (proving the order of not one, but two Mersenne primes!)...

There's no place in history for completing the first first-time test below X.
There's not even a place in history for completing the last double-check test below X.
And neither of which proves the order of any Mersenne Primes, for common definitions of "order".

Commaster · 2010-12-11, 11:51

Maybe we will get another (read "Don't care about 100.00000001 > 100) opinion as soon as this assignment holds 3-4 Milestones ^.^

TimSorbet · 2010-12-11, 13:28

Quote:

Originally Posted by NBtarheel_33

I guess it might be considered prodding if George were to e-mail him/her and let them know that they are 93% of the way to their place in mathematical history (proving the order of not one, but two Mersenne primes!)...

I think it's a good idea for George to (politely) 'prod' the user in this way. Perhaps it will encourage them to move the test to their quad, (if possible) or at least keep the computer on 24/7 (if it's not already), and keep unnecessary CPU-intensive tasks (e.g. other DC participations) off it (if they're not already). Or it could change nothing. Or, as I'm sure Cheesehead et al will tell me, it could also anger them so they quit GIMPS, but I don't think there's a great risk of that happening just by making sure they know the significance of that test (no, it's not extremely important in the long run, but it is holding up a couple of milestones, and some people are more bugged by that than others). Besides, such a communication could very well avert it being poached, which would be far more likely to disappoint/anger the user and make them quit.

Quote:

Originally Posted by ckdo

There's no place in history for completing the first first-time test below X.
There's not even a place in history for completing the last double-check test below X.
And neither of which proves the order of any Mersenne Primes, for common definitions of "order".

I'm not going to argue about what constitutes a "place in history", but completing the last double-check test below X does prove the order of a Mersenne prime when X is a Mersenne prime, i.e. it proves that since no unknown primes exist below X, that we know that X is the Zth Mersenne number, instead of being uncertain which Mersenne number it is. (in this situation, this is the last first-time test below X, so it's not 100% certain yet) If you have a looser definition of "proving" an "order", this might instead be the last first-time test below X, or when only a certain relatively low number of first-time or double-checks remain below X.

Brian-E · 2010-12-11, 14:20

The completed first-time tests below M(37156667) must include thousands of incorrect results. So why the fuss about one incompleted test?

davieddy · 2010-12-11, 14:52

His/her Mnumber has a significantly better chance of being prime
than any other, and fewer than 10M digits would
be a turn up for the history books.

David

2010-12-11, 11:51	#30
Commaster Jun 2010 Kiev, Ukraine 3·19 Posts	Maybe we will get another (read "Don't care about 100.00000001 > 100) opinion as soon as this assignment holds 3-4 Milestones ^.^ Last fiddled with by Commaster on 2010-12-11 at 11:51

2010-12-11, 14:52	#33
davieddy "Lucan" Dec 2006 England 14512₈ Posts	His/her Mnumber has a significantly better chance of being prime than any other, and fewer than 10M digits would be a turn up for the history books. David Last fiddled with by davieddy on 2010-12-11 at 14:58

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Newer X64 build needed	Googulator	Msieve	75	2022-06-13 14:22
Performance of cuda-ecm on newer hardware?	fivemack	GMP-ECM	14	2015-02-12 20:10
Cause this don't belong in the milestone thread	bcp19	Data	30	2012-09-08 15:09
Newer msieves are slow on Core i7	mklasson	Msieve	9	2009-02-18 12:58
Use of large memory pages possible with newer linux kernels	Dresdenboy	Software	3	2003-12-08 14:47

2010-12-11, 08:59	#25
NBtarheel_33 "Nathan" Jul 2008 Maryland, USA 5×223 Posts	Only problem is that this single exponent (which is now holding up two milestones) is being run on a P3 by a user who updates only every 28 days, and is not due to finish the test until May of next year... ...yeah, yeah, I know, 100.00000001 is still faster than 100, and all that jazz.

2010-12-11, 09:04	#26
Uncwilly 6809 > 6502 """"""""""""""""""" Aug 2003 101×103 Posts 10,891 Posts	I suspect that there will be a real early double check of that number completed soon enough.

2010-12-11, 09:29	#27
davieddy "Lucan" Dec 2006 England 2×3×13×83 Posts	Even allowing for the 10M digit prize distorting steady progress somewhat, you might hope that the person entrusted with this (now rather special) LL test would be sufficiently conscious of its significance to (at least) keep his machine running 24/7. David

2010-12-11, 14:20	#32
Brian-E "Brian" Jul 2007 The Netherlands 2×11×149 Posts	The completed first-time tests below M(37156667) must include thousands of incorrect results. So why the fuss about one incompleted test?