Newer milestone thread - Page 58

davieddy · 2012-08-10, 05:38

Quote:

Originally Posted by bcp19

Just because an exponent is 'the lowest unverified' so far, does that statistically make it more likely to be a prime? Why is it considered a 'plum'?

Yes. Probability of being prime is inversely proportional to the exponent.
Time taken to LL is proportional to exponent².
So probability/year of finding a prime is proportional to 1/exponent³.

Humble toilers such as myself like these plums, and resent them being reserved for Chalsall's Chums.

Look after the waveFRONT speed, and let the tail look after itself.

If Primenet can't exploit GPUs properly yet, it should retire all the hoarded TF assignments >53M.

D

davieddy · 2012-08-10, 10:48

Quote:

Originally Posted by axn

Nothing much. With that, I'll shut up.

Please don't.
I (for one) respect your contributions immensely.
You need a thicker skin to resist the morons around here.

David

Prime95 · 2012-08-10, 14:01

Quote:

Originally Posted by axn

Nothing much. With that, I'll shut up.

Some of your erato.cu optimizations are also in my GPU sieving code. I guess that also makes you one of the two dozen or so that have contributed actual working code - thanks! PM me, if you'd like a copy of the new sieve code to see if you can find some more micro-optimizations.

Uncwilly · 2012-08-12, 06:42

Quote:

Originally Posted by Uncwilly

All exponents below 24,212,801 have been tested and double-checked.
All exponents below 38,195,891 have been tested at least once.

Countdown to testing all exponents below M(42643801) once: 28
Countdown to testing all exponents below M(43112609) once: 35

Countdown to proving M(25964951) is the 42nd Mersenne Prime: 185
Countdown to proving M(30402457) is the 43rd Mersenne Prime: 35,330

All exponents below 24,212,801 have been tested and double-checked.
All exponents below 41,060,057 have been tested at least once.

Countdown to testing all exponents below M(42643801) once: 13
Countdown to testing all exponents below M(43112609) once: 18

Countdown to proving M(25964951) is the 42nd Mersenne Prime: 70
Countdown to proving M(30402457) is the 43rd Mersenne Prime: 28,626

It looks like 3 of these milestones will get retired by the end of 2012.

davar55 · 2012-08-12, 13:34

GIMPS is close to testing all Mersenne numbers up to 2^43112609-1,
which will eventually prove that M47* = M47.

As long as we have no proof that the Mersenne Primes are infinite in
number, continuing the search beyond any given limit such as M47 is
of course interesting but unbounded.

Hence my question: is there any way to measure the probability that
M47 is the last Mersenne Prime?

garo · 2012-08-12, 13:34

A few observations:

1. The pfrakes exponent that chalsall linked to above is most likely the same user testing it twice. I suspect this was because the first-time test was done in v4 and the second time test done in v5 and v5 does not remember the v4 IDs for checking if the same user returned two LLs for the same test. Also, George has a second check - the shift count - and if both results are submitted using Prime95 it becomes very hard - if not impossible - to fake these.

2. George's reply on what GPU72 has done for us is pretty much to the point. I would just like to point out that I am one of those who runs his own script periodically to get plum DCs and I found that I was able to get what I wanted even when GPU72 was grabbing exponents. I do think George's idea of leaving 1 mod 4 or even just 1 mod 4n where n>1 exponents for non-GPU72 users is a good idea.

3. My count of exponents between 25-26M is now 23/72. So I am keeping with the pace.

bcp19 · 2012-08-12, 16:20

Quote:

Originally Posted by sonjohan

I get it that poaching is not supposed to happen, but how can we advance the milestones faster?
One of the 2 DC left in the 24M range is over 1 year old.
How can you see whether progress is being made on these or not?

I've been tracking 24-26M assignments for some time, and the two you mention have the following stats:
xx,xxx,801 - 10.90% on 5/9, 11.00% on 8/12
xx,xxx,607 - 61.20% on 4/8, 61.20% on 5/7, reassigned on 5/9, no progress to date.

ATH · 2012-08-12, 21:14

Quote:

Originally Posted by Prime95

.

Since you are following this thead, I want to ask: Do you plan to implement any assignment recycling rules like these or similar?: http://www.mersenneforum.org/showpos...&postcount=443

I think it would be nice with something preventing assignments for getting stuck and running 2+ years with little progress.

Prime95 · 2012-08-13, 01:06

Quote:

Originally Posted by ATH

Since you are following this thead, I want to ask: Do you plan to implement any assignment recycling rules like these or similar?

At the time I proposed that, there was a fair amount of opposition so I elected to not implement any changes.

Dubslow · 2012-08-13, 01:38

Quote:

Originally Posted by Prime95

At the time I proposed that, there was a fair amount of opposition so I elected to not implement any changes.

Please consider this post the second vote for implement those rules.

chalsall · 2012-08-13, 02:37

Quote:

Originally Posted by Dubslow

Please consider this post the second vote for implement those rules.

Third.

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

2012-08-12, 13:34	#632
davar55 May 2004 New York City 5·7·11² Posts	GIMPS is close to testing all Mersenne numbers up to 2^43112609-1, which will eventually prove that M47* = M47. As long as we have no proof that the Mersenne Primes are infinite in number, continuing the search beyond any given limit such as M47 is of course interesting but unbounded. Hence my question: is there any way to measure the probability that M47 is the last Mersenne Prime?

2012-08-12, 13:34	#633
garo Aug 2002 Termonfeckin, IE 2768₁₀ Posts	A few observations: 1. The pfrakes exponent that chalsall linked to above is most likely the same user testing it twice. I suspect this was because the first-time test was done in v4 and the second time test done in v5 and v5 does not remember the v4 IDs for checking if the same user returned two LLs for the same test. Also, George has a second check - the shift count - and if both results are submitted using Prime95 it becomes very hard - if not impossible - to fake these. 2. George's reply on what GPU72 has done for us is pretty much to the point. I would just like to point out that I am one of those who runs his own script periodically to get plum DCs and I found that I was able to get what I wanted even when GPU72 was grabbing exponents. I do think George's idea of leaving 1 mod 4 or even just 1 mod 4n where n>1 exponents for non-GPU72 users is a good idea. 3. My count of exponents between 25-26M is now 23/72. So I am keeping with the pace.