[QUOTE=tha;367075]The original owner had stopped working on it about 9 days ago.[/QUOTE]How do you know that? Perhaps it was already finished and simply waiting to update the server the next time it got an Internet connection.

[QUOTE=retina;367082]How do you know that? Perhaps it was already finished and simply waiting to update the server the next time it got an Internet connection.[/QUOTE]

Indeed, correctly would have been 'stopped reporting progress'. A few times I started work on exponents but waited to turn in the result (halting work a day before finishing) until work by the owner had either resumed or become more likely to have been abandonded.

Oh so because you're giving the rightful owner "just a bit more time" (whatever the ---- that means) to finish in the case where they are actually trying to finish the assignment, then that makes your poaching completely fine?

And in the entirely possible case where they DO report their work, unknowing of the fact that had they taken "an amount of time" longer to finish it, they would have lost all of it, you've wasted almost enough time to have just done some legitimate assignment instead.


This assignment rule re-adjustment actually started from a bit of bumbling about organized poaching to clear out the milestones. The idea was to use our combined judgement to determine which assignments were really being held up and which ones were just slow.

Instead, you're using just your own personal judgement to decide [B]for me[/B] if [B]I[/B] am doing [B]my own[/B] assignment properly or not.

As it happens, I'm fortunate enough to have joined this project only recently and with a brand-new computer so I can complete more than on time so I don't have to worry about getting poached. But I can say that if I was in the shoes of someone labouring for months on a single assignment and waiting for next time I go into town to bring my results.txt to the library on a USB stick, only to see someone with a fancy new computer steal the work, I would be very unhappy.


In the past, people [B]have implicitly agreed to[/B] a one-year reservation limit on assignments that someone else may want. GIMPS has never really needed to enforce that time limit because the idea was "it'll get done eventually."

Now that people's impatience with the milestones has gotten higher, we're planning to grandfather that agreement to give people the time [B]they agreed on[/B].

The statement is "You get one year," not " You get as long as it takes before some guy takes your assignment."


Really, the bottom line is you are making GIMPS infringe on its end of the contract.

I agree with the principles of what you are writing, TheMawn, and as someone who likes doing DC's, about six per year, on my 7-year-old part-time-operating machine, I have felt threatened by poaching myself. But tha has indicated an intention to stop this activity now that the new (re)assignment rules are in place, and has in fact [URL="http://www.mersenneforum.org/showpost.php?p=367074"]just tried[/URL] to quell an over-enthusiastic attempt by someone to put forward a list of DC assignments in the 32M range for poaching.

The real problem is combatting poachers who do it surreptitiously and without any stated justification. I fear there is little to be done about them. But we can at least make it clear, as you have just done, that the behaviour is not condoned.

[QUOTE]Countdown to proving M(30402457) is the 43rd Mersenne Prime: 3[/QUOTE]

This is all due to poaching?

[QUOTE=retina;367021][QUOTE=retina;367014]30399143 was by Fredrik also.[/QUOTE]Add 30286357 to the list.[/QUOTE]And 30207883.

Anyone know where telix5000 went? So close and yet so far.

[QUOTE=petrw1;367459]Anyone know where telix5000 went? So close and yet so far.[/QUOTE]After receiving so many "result not needed" responses he/she probably turned off P95 and took up some other hobby.

[QUOTE=retina;367482]After receiving so many "result not needed" responses he/she probably turned off P95 and took up some other hobby.[/QUOTE]

I, personally, recommend autogratification. Or miniature trains. Or both (at the same time is extra fun!)....

[QUOTE=chalsall;367485]I, personally, recommend autogratification. Or miniature trains. Or both (at the same time is extra fun!)....[/QUOTE]Yes indeed ! Let GIMPS be reserved to professionals, or at least semi-professionals. All those amateurs should be whipped away.

Jacob

30170867
 

So this time it was Mr. ANONYMOUS doing the poaching.

So who is the one that was too shy to admit it?

Man, this is a real pain. After the implementation of the new recycling criteria, poaching is becoming even more unacceptable, due to lack of excuses.
I had hoped that the (excellent) measures recently taken would deter people from poaching, but there are at least a few impatient and unethical ones that keep on doing it. Unfortunately, there isn´t much more we can do about it.

I expect that the poaching will drop off once we reach the M43 milestone. Then the new assignment rules will be able to do their job for future milestones. Once M43 is placed in its proper slot, expect to see the extinction of poaching. What would be the reason for doing it, after all, if the assignment and recycling rules are functioning as intended? Certainly it becomes harder to justify on the part of the poacher.

[QUOTE]Countdown to proving M(30402457) is the 43rd Mersenne Prime: 0[/QUOTE]

:party:

Yep, our old friend Fredrik finished the last one: 30184789

Congratulations! Thank you to everyone who did their best to make this possible without poaching.

Congrats to GIMPS.

We've all been watching that milestone approach.
Now it's time to start watching the next one.
But that may be the only significant one reached this year.

[QUOTE=davar55;367664]Congrats to GIMPS.

We've all been watching that milestone approach.
Now it's time to start watching the next one.
But that may be the only significant one reached this year.[/QUOTE]

No, I think M44 will be confirmed this year. There are only 8,000 exponents left to double-check, and with the new assignment rules, they should be checked in an efficient manner.

[QUOTE=NBtarheel_33;367666]There are only 8,000 exponents left to double-check...[/QUOTE]
Ok, there are 60 days gone already, 300 days left, with 2 per day, I can do 600. As I look to [URL="http://www.gpu72.com/reports/workers/dc/graph/1-10/"]these graphs[/URL], chalsall can do about 3000 alone, monst another 500, and kracker another 500, other will do the rest, and we may still have time for some beer... :razz:

And we didn't count curtisc, neither chuck norris...

[SIZE=1]Disclaimer: this was a joke, the hottest period of the year just started in Thai, so I may need to stop our partnership soon, but seriously speaking, 8000 DC in 10 months, that is piece of cake for the first 20 guys in the top only! (and yes, I know, that curtisc, same as chuck norris, doesn't do DCs...)[/SIZE]

[QUOTE=LaurV;367670]And we didn't count curtisc, neither chuck norris...[/QUOTE]

Chuck Norris once found a Mersenne Prime with a composite exponent.

[QUOTE=NBtarheel_33;367666]M44[/QUOTE]Are we there yet?

:bear:

[QUOTE=NBtarheel_33;367666]No, I think M44 will be confirmed this year. There are only 8,000 exponents left to double-check, and with the new assignment rules, they should be checked in an efficient manner.[/QUOTE]

Indeed.

Countdown to proving M(32582657) is the 44th Mersenne Prime: 7,959

This seems to be falling at a pretty substantial rate.

Also, the remaining candidates below M48 seem to be falling at a very fast rate.

[QUOTE=Primeinator;367684]Indeed.

Countdown to proving M(32582657) is the 44th Mersenne Prime: 7,959

This seems to be falling at a pretty substantial rate.

Also, the remaining candidates below M48 seem to be falling at a very fast rate.[/QUOTE]

I think we could see both drop this year....mostly due the the new rules.

Countdown to proving only 44 Mersenne Primes less than 10M digits: ?
 

Countdown to proving only 44 Mersenne Primes less than 10M digits: ?

The query below is limited to 1000 results:
[url]http://mersenne.org/assignments/default.php?exp_lo=1&exp_hi=33219278&execm=1&exfirst=1&exp1=1&extf=1[/url]

Is there a way to get just the count of exponents within a range without returning each individual status?

[QUOTE=retina;367705]Countdown to proving only 44 Mersenne Primes less than 10M digits: ?

The query below is limited to 1000 results:
[url]http://mersenne.org/assignments/default.php?exp_lo=1&exp_hi=33219278&execm=1&exfirst=1&exp1=1&extf=1[/url]

Is there a way to get just the count of exponents within a range without returning each individual status?[/QUOTE]

Not exactly, AFAIK, but you can get the counts for 30M-33M (<30M is fully proven) from [URL="http://www.mersenne.org/primenet/"]the summary page[/URL]:
[CODE]
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
Exponent Range | Composite | Status Unproven | Assigned | Available |
Start Count P | F LL-D | LL LLERR NO-LL | TF P-1 LL LL-D | TF P-1 LL LL-D |
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
30000000 58120 1 | 36902 21098 119 | 119 | |
31000000 57836 | 36761 18613 2462 | 2029 | 433 |
32000000 57852 1 | 36913 9203 11735 | 8567 | 3167 |
[/CODE]
Then checking 1000 results (starting with your link) at a time for 33M to 33.2M, I find 3590 in that range, for a total of 119+2462+11735+3590=17,906 DCs remaining.

[QUOTE=Mini-Geek;367707]Not exactly, AFAIK, but you can get the counts for 30M-33M (<30M is fully proven) from [URL="http://www.mersenne.org/primenet/"]the summary page[/URL]:
[CODE]
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
Exponent Range | Composite | Status Unproven | Assigned | Available |
Start Count P | F LL-D | LL LLERR NO-LL | TF P-1 LL LL-D | TF P-1 LL LL-D |
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
30000000 58120 1 | 36902 21098 119 | 119 | |
31000000 57836 | 36761 18613 2462 | 2029 | 433 |
32000000 57852 1 | 36913 9203 11735 | 8567 | 3167 |
[/CODE]
Then checking 1000 results (starting with your link) at a time for 33M to 33.2M, I find 3590 in that range, for a total of 119+2462+11735+3590=17,906 DCs remaining.[/QUOTE]Okay, thanks.

This query returns all results but contains duplicates:
[url]http://www.mersenne.org/report_LL/default.php?exp_lo=28873849&exp_hi=33219278&exp_date=&user_only=0&user_id=&exdchk=1&exbad=1&exfactor=1&txt=1&dispdate=1[/url]

After removing duplicates I got 17999. Where are the extra 93?

Using the same report_LL page and duplicate removal for each individual range I get these counts:
30M=118 - discrepancy of -1
31M=2459 - discrepancy of -3

There appears to be no obvious way to get reliable numbers. And each individual query suffers from the problem of not being atomic with figures potentially changing between each access.

I guess it will have to do until the number drops below 1000 and a single query can return the entire range.

[QUOTE=retina;367710]Okay, thanks.

This query returns all results but contains duplicates:
[url]http://www.mersenne.org/report_LL/default.php?exp_lo=28873849&exp_hi=33219278&exp_date=&user_only=0&user_id=&exdchk=1&exbad=1&exfactor=1&txt=1&dispdate=1[/url]

After removing duplicates I got 17999. Where are the extra 93?

Using the same report_LL page and duplicate removal for each individual range I get these counts:
30M=118 - discrepancy of -1
31M=2459 - discrepancy of -3

There appears to be no obvious way to get reliable numbers. And each individual query suffers from the problem of not being atomic with figures potentially changing between each access.

I guess it will have to do until the number drops below 1000 and a single query can return the entire range.[/QUOTE]

The discrepancies of 1 and 3 per million can probably be explained by recent completions (the status I quoted is updated every hour, I think). Stats are now (2014-02-24 16:01 UTC):
[CODE]
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
Exponent Range | Composite | Status Unproven | Assigned | Available |
Start Count P | F LL-D | LL LLERR NO-LL | TF P-1 LL LL-D | TF P-1 LL LL-D |
----------=-----=-- | -----=-----=-----=-----=----- | -----=-----=-----=----- | -----=-----=-----=----- |
30000000 58120 1 | 36902 21099 118 | 118 | |
31000000 57836 | 36761 18616 2459 | 2027 | 432 |
32000000 57852 1 | 36913 9211 11727 | 8563 | 3163 |[/CODE]
I think that looking at assignments (as I was, based on your first post) isn't the right way to go, because of the fact that not all 'status unproven' are assigned, starting in the 31M block (as the stats above show). So, based on using your report_LL query for 33M-33.2M and finding 3693 distinct in that range, I think the actual figure as of ~16:01 UTC is 118+2459+11727+3693=17,997

[QUOTE=retina;367705]Countdown to proving only 44 Mersenne Primes less than 10M digits: ?[/QUOTE]

[CODE]mysql> select count(*) from Exponent where Exponent<33219278 and (Status=4 or Status=5);
+----------+
| count(*) |
+----------+
| 17994 |
+----------+[/CODE]

[QUOTE=LaurV;367670]Ok, there are 60 days gone already, 300 days left, with 2 per day, I can do 600. As I look to [URL="http://www.gpu72.com/reports/workers/dc/graph/1-10/"]these graphs[/URL], chalsall can do about 3000 alone, monst another 500, and kracker another 500, other will do the rest, and we may still have time for some beer...[/QUOTE]Don't forget that not all GIMPS work is performed by GPU's : the top 30 on that graph is not the top thirty on GIMPS.
Jacob

[QUOTE=S485122;367726]Don't forget that not all GIMPS work is performed by GPU's : the top 30 on that graph is not the top thirty on GIMPS.[/QUOTE]

Of course not. GPUs should only be doing TF'ing at the moment, and for the foreseeable future.

LaurV et al's point, however, is valid. Under the new assignment rules proving M(32582657) is the 44th Mersenne Prime will very likely be accomplished this year (mostly using CPUs).

[QUOTE=LaurV;367670]Ok, there are 60 days gone already, 300 days left, with 2 per day, I can do 600. As I look to [URL="http://www.gpu72.com/reports/workers/dc/graph/1-10/"]these graphs[/URL], chalsall can do about 3000 alone, monst another 500, and kracker another 500, other will do the rest, and we may still have time for some beer... :razz:[/QUOTE]

I can do [I]much[/I] better than 2 per per day if I really wanted to/stopped messing around.

:razz:

My poor minions.

[QUOTE=chalsall;367728] Under the new assignment rules proving M(32582657) is the 44th Mersenne Prime will very likely be accomplished this year (mostly using CPUs).[/QUOTE]

Nope. All assignments prior to March 1, 2014 are grandfathered -- given a full year to complete. I'm guessing we won't prove M#44 until June 2015.

[QUOTE=Prime95;367734]Nope. All assignments prior to March 1, 2014 are grandfathered -- given a full year to complete. I'm guessing we won't prove M#44 until June 2015.[/QUOTE]

What about poaching? :wink:

[QUOTE=chalsall;367738]What about poaching? :wink:[/QUOTE]

The recent wave of poaching was after a long time of 'residue' building up. I think we will see a low level again for lack of residue. By the time new residue will be build up the new rules will be in effect.

[QUOTE=chalsall;367719][CODE]mysql> select count(*) from Exponent where Exponent<33219278 and (Status=4 or Status=5);
+----------+
| count(*) |
+----------+
| 17994 |
+----------+[/CODE][/QUOTE]Is this some private DB you have, or access to the primenet server?

It seems the report_LL is currently the closest approximation to get the value.

[QUOTE=retina;367744]Is this some private DB you have, or access to the primenet server?[/QUOTE]

I've said it before, I'll say it yet again: I have no privileged access to Primenet.

[QUOTE=retina;367744]It seems the report_LL is currently the closest approximation to get the value.[/QUOTE]

[CODE]mysql> select UTC_TIMESTAMP(),count(*) from Exponent where Exponent<33219278 and (Status=4 or Status=5);
+---------------------+----------+
| UTC_TIMESTAMP() | count(*) |
+---------------------+----------+
| 2014-02-24 23:07:11 | 17957 |
+---------------------+----------+[/CODE]

[QUOTE=chalsall;367745]I've said it before, I'll say it yet again: I have no privileged access to Primenet.[/QUOTE]

I take it this is a query of GPUto72's DB (populated by pulling data from PrimeNet), then?

[QUOTE=Mini-Geek;367746]I take it this is a query of GPUto72's DB (populated by pulling data from PrimeNet), then?[/QUOTE]

Yes.

Anyone could do this.

[QUOTE=petrw1;364159]1 down...1753[/QUOTE]

1277 to 61 bits. 1 to go. Early April or so.

All exponents below [B][COLOR="DarkRed"]30,412,999[/COLOR][/B] have been tested and double-checked.
All exponents below [B][COLOR="SandyBrown"]45,807,607[/COLOR][/B] have been tested at least once.

Countdown to testing all exponents below M([B][COLOR="Blue"]57885161[/COLOR][/B]) once: 11,331

Countdown to proving M([COLOR="MediumTurquoise"]32582657[/COLOR]) is the [COLOR="MediumTurquoise"]44[/COLOR]rd Mersenne Prime: 7,564
Countdown to proving M([COLOR="Green"]37156667[/COLOR]) is the [COLOR="green"]45[/COLOR]th Mersenne Prime: 93,410

Cracked top-100 DC for 365 days :smile: All that's left is LL and I don't see that happening without more hardware :unsure:

[QUOTE=TheMawn;368372]Cracked top-100 DC for 365 days :smile: All that's left is LL and I don't see that happening without more hardware :unsure:[/QUOTE]

Just a couple GPUs from top-100 in all categories Life Time and YTD.

[QUOTE=TheMawn;368372]Cracked top-100 DC for 365 days :smile: All that's left is LL and I don't see that happening without more hardware :unsure:[/QUOTE]

Always the limiting factor. :down:

March 13, 2014. All exponents below 46 million have been tested at least once.

Today, for the first time (in the modern era), the projected end of the 79.3M range is in 2017 or earlier.
The current projection is 12/30/17.
This is based upon the changes in the P90 years remaining.
It will move downward as we move along It may actually come before the end of 2015.

[QUOTE=Uncwilly;369171]
It will move downward as we move along It may actually come before the end of 2015.[/QUOTE]

I wonder what the effect of tablets replacing pc's will be. For sure pc's won't be upgraded as often. I have no legitimate reason to upgrade my 2007 era pc. Plenty of illegitimate reasons though.

[QUOTE=tha;369182]I wonder what the effect of tablets replacing pc's will be. For sure pc's won't be upgraded as often. I have no legitimate reason to upgrade my 2007 era pc. Plenty of illegitimate reasons though.[/QUOTE]
There are plenty of office desk PC's that will remain as such. They will also get 'life-cycle' replacements. A tablet can't touch a wide screen monitor with a real keyboard attached.

[QUOTE=Uncwilly;369171]Today, for the first time (in the modern era), the projected end of the 79.3M range is in 2017 or earlier.
The current projection is 12/30/17.
This is based upon the changes in the P90 years remaining.
It will move downward as we move along It may actually come before the end of 2015.[/QUOTE]
When is the projected end of the 100M range?

[QUOTE=c10ck3r;369215]When is the projected end of the 100M range?[/QUOTE]I do not have data to do that projection. I am using the change in P90 years as posted in the "classic" report.

Just in time to celebrate with a green beer:

[QUOTE]March 17, 2014: All exponents below 47,000,000 tested at least once. [/QUOTE]

[QUOTE=philmoore;369227]Just in time to celebrate with a green beer:[/QUOTE]

Your cat prevents me.

[QUOTE=tha;369182] I have no legitimate reason to upgrade my 2007 era pc. Plenty of illegitimate reasons though.[/QUOTE]

What? Is GIMPS not a legitimate enough reason? :shock:
:lol:

[QUOTE=kracker;369230]Your cat prevents me.[/QUOTE]

Fortunately, you are right. My cat insisted I celebrate with a traditional stout. And a few fiddle tunes, but the cat was not too happy about that.

I don't remember the thread, but there was recent discussion of finding multiple factors. Over about nine hours today, while I was at work, my GTX 570 found three. This was in addition to 21 No Factor results for the combined output of the 570 and 580.

And then there were none.....
 

[QUOTE=petrw1;368008]1277 to 61 bits. 1 to go. Early April or so.[/QUOTE]

Tomorrow there should be no 60 column on GPU72

[QUOTE=petrw1;370285]Tomorrow there should be no 60 column on GPU72[/QUOTE]

Column or row? And, are you planning to wipe out 2048 73-74 bit exponents overnight?

Just got done. No more exponents trial factored to less than 61 bits.
Good job, petrw1, the "tail" in the table got shorter.
Tackling 61 bits will be way harder a job, though...

Check what bounds p-1 has been run to against the numbers that have not been TFed to 61 bits. There will certainly be no factors smaller than B1. That might save some time.

Chris

[QUOTE=chris2be8;370316]Check what bounds p-1 has been run to against the numbers that have not been TFed to 61 bits. There will certainly be no factors smaller than B1. That might save some time.

Chris[/QUOTE]

There are no exponents that fit this criteria. See above.

[QUOTE=lycorn;370296]Tackling 61 bits will be way harder a job, though...[/QUOTE]

Can GPU's not do that? I could clear 15,000 67->68 overnight with my hardware alone...

[QUOTE=TheMawn;370331]Can GPU's not do that? I could clear 15,000 67->68 overnight with my hardware alone...[/QUOTE]

Try 61 and see what happens :wink:

[QUOTE=TheMawn;370331]Can GPU's not do that? [/QUOTE]

Yes they can, and many exponents have already been cleared in that range by GPUs. The thing is, efficiency at these low bit levels is way lower than above 64 bits. First, because GPU sieving is only enabled above 64 bits, and second, the kernels used, particularly for exponents < 1M, have not been extensively optimized.
That said, GPUs are in any case much faster than CPUs, but we can by no means expect them to perform these tasks at the speeds they use to.
As an example, my GTX560Ti TFs a 950,000 exponent from 61 to 62 bits in ~14 minutes.
Also note that even if GPU sieving was enabled and the kernels optimized, the tests would take longer anyway, because these exponents are much smaller than the "mainstream" ones. A 1M exponent, tested to the same levels and in the same conditions than a 60M, would take roughly 60 times more, The fact that we are testing to lower levels partially makes up for this, but still...

I have to admit the Primenet Summary page looks [B]really[/B] nice. I hadn't checked it in a while, but the tails have sure been trimmed down! Still waiting for the 20M range to join 0M and 10M with their separate headings for ECM factoring :smile:

Also, looking at the milestones, you can see how much of a difference the new rules have already made:
[QUOTE]March 26, 2014: All exponents below 48,000,000 tested at least once.
March 17, 2014: All exponents below 47,000,000 tested at least once.
March 13, 2014: All exponents below 46,000,000 tested at least once.
February 14, 2014: All exponents below 30,000,000 double-checked.
February 4, 2014: All exponents below 29,000,000 double-checked.
January 12, 2014: All exponents below 28,000,000 double-checked.
January 9, 2014: All exponents below 27,000,000 double-checked. [/QUOTE]

As of now, all exponents below 50,000,000 have been tested at least once but that hasn't been noted on the site, yet.

That's nine millions in the last four months. The previous nine before that:

[QUOTE]December 11, 2013: All exponents below 45,000,000 tested at least once.
November 9, 2012: All exponents below 44,000,000 tested at least once.
August 27, 2012: All exponents below 25,000,000 double-checked.
April 23, 2011: All exponents below 22,000,000 double-checked.
January 7, 2010: All exponents below 20,000,000 double-checked.[/QUOTE]

The[SIZE=2] "Countdown to testing all exponents below M(57885161) once" fell below 10000 (at 9,957).
I hope that the milestone will be reached soon enough :)
(It doesn't seem too far away, although there'll probably be some exponents lagging behind.)

edit: I just refreshed the page and now the countdown is at 10,150. :confused:
[/SIZE]

I just changed the web page to only consider error-free LL tests in "all exponents below x are tested once" calculations. This was done so that it matches the rules when getting first-time assignments.

Wow some progress
 

Jan4: Countdown to proving M(32582657) is the 44rd Mersenne Prime: 12,721
Apr9: Countdown to proving M(32582657) is the 44th Mersenne Prime: 5,509

3 Months and about 65% done.
Dare I say by Fireworks day it could easily be done.!!!

[QUOTE=petrw1;370739]3 Months and about 65% done.
Dare I say by Fireworks day it could easily be done.!!![/QUOTE]Not a chance without some serious amount of "poaching" of the stragglers.

Even with the new assignment rules?

(I did not check anything, I am just honestly asking, you know, we put some hopes in these new rules...)

[QUOTE=LaurV;370750]Even with the new assignment rules?

(I did not check anything, I am just honestly asking, you know, we put some hopes in these new rules...)[/QUOTE]Here is one to start you off: 31455961

And there are hundreds more just like it. Are you prepared to "poach" all of those to meet the 2014-07-04 deadline?

I personally have no remorse to poach few of those :razz:
Fortunately for the assignees, I have plenty of other things to do... Chris is torturing me with those ranges of his, to factor... So don't count on me this time. I was only the curious monkey.

Disclaimer: this is more of a (serious) joke, and not an invitation to poaching! I don't approve indiscriminate poaching.

[QUOTE=lycorn;370296]Just got done. No more exponents trial factored to less than 61 bits.
Good job, petrw1, the &quot;tail&quot; in the table got shorter.
Tackling 61 bits will be way harder a job, though...[/QUOTE]

With exponents this small would the P-1 trick in [url]http://mersenneforum.org/showthread.php?t=14265&page=8[/url] post 83 be a quicker way to prove there are no factors smaller than your desired bound? Or arbooker's Pollard-Strassen app. mentioned in the same thread?

Chris

[QUOTE=LaurV;370758]Fortunately for the assignees, I have plenty of other things to do... Chris is torturing me with those ranges of his, to factor... So don't count on me this time.[/QUOTE]

Hey, you've been very well paid in kind! :razz: :wink: :smile:

[QUOTE=chris2be8;370791]With exponents this small would the P-1 trick in [URL]http://mersenneforum.org/showthread.php?t=14265&page=8[/URL] post 83 be a quicker way to prove there are no factors smaller than your desired bound? Or arbooker's Pollard-Strassen app. mentioned in the same thread?

Chris[/QUOTE]
B2 is limited to around 1e17 which limits us somewhat. There might be ways around this.

You can get round the limit on B2 by running just stage 1 and saving the result, then resuming several times with a range of B2 values. This can be used to spread stage 2 over several machines (or several cores on one machine if you have enough RAM). See the ecm man page for details.

The ultimate limit is that B1 can't be more than 9007199254740996 so B2 can't be more that that squared. But I don't think anyone will reach that limit in this century.

Chris

Can you get a higher value of P? Do you need to?

our first 10M range under 200,000 remaining
 

Thanks in a big part to GPU72 advancing the bit levels the 60-69M range now has less than 200,000 unfactored exps.

Shortening the tail...
 

May 1, 2014. All exponents below 51 million have been tested at least once.

[QUOTE=NBtarheel_33;372438]May 1, 2014. All exponents below 51 million have been tested at least once.[/QUOTE]
I noticed that this morning. The attached graph, that looks like a city skyline,
shows the gap between the tail end of OneLL and TwoLL.
It is (OneLL - TwoLL) / OneLL expressed as a %age.

I don't like that graph. I'd say we should zoom in a lot more but I'm guessing the resolution doesn't exist.

In the 90's the overall gap would have been massive on account of there being so few exponents to begin with. The surge in 2007 is interesting, though.

[QUOTE=TheMawn;372454]In the 90's the overall gap would have been massive on account of there being so few exponents to begin with. The surge in 2007 is interesting, though.[/QUOTE]

I don't think 2007 is a result of a surge in DC, as much as a lag in LL. Probably a lot of focus was on finding the 10-mill digit prime, so the lowest of LL was lagging.

[QUOTE=TheMawn;372454]I don't like that graph. I'd say we should zoom in a lot more but I'm guessing the resolution doesn't exist.[/QUOTE]
Well, I have fairly good data from ~Oct 2010. Is this good enough for the current period?

BTW, I have all the data that I could get my hands on.

Much better. I still don't know what it says, if anything.

Just the milestones
 

:coffee:

With a straight scale
 

Emphasises the tick up at the end.

Straight scale, with primes, and rich data.

[QUOTE=Uncwilly;372515]Straight scale, with primes, and rich data.[/QUOTE]
... data, yes, more data is good.

[QUOTE=markr;372505]Emphasises the tick up at the end.[/QUOTE]

I like that one :-)

I did not make that big rock rolling, others took their big canons and made that happen. But my now not so new anymore pc removed the little stones that prevented the big rock from rolling down the slope. It worked a lot better than I had hoped for. Kudos to all who banded together to make this happen.

[QUOTE=petrw1;370739]Jan4: Countdown to proving M(32582657) is the 44rd Mersenne Prime: 12,721
Apr9: Countdown to proving M(32582657) is the 44th Mersenne Prime: 5,509

3 Months and about 65% done.
Dare I say by Fireworks day it could easily be done.!!![/QUOTE]
May 16: under 3,995 or over 68% since Jan 5.

Appears I was dyslexic last time. Should have read 56% not 65%

[QUOTE=petrw1;370739]Jan4: Countdown to proving M(32582657) is the 44rd Mersenne Prime: 12,721
Apr9: Countdown to proving M(32582657) is the 44th Mersenne Prime: 5,509

3 Months and about 65% done.
Dare I say by Fireworks day it could easily be done.!!![/QUOTE]
Is today this fireworks day you've mentioned?

July 1: Countdown to proving M(32582657) is the 44th Mersenne Prime: 2,234
(BTW, over what were you taking the percentages?)

What's "44rd"? :razz:

[QUOTE]What's "44rd"? :razz:[/QUOTE]That comes after "43th".

[QUOTE=Qubit;377152]Is today this fireworks day you've mentioned?

July 1: Countdown to proving M(32582657) is the 44th Mersenne Prime: 2,234
(BTW, over what were you taking the percentages?)[/QUOTE]

I fail.....:davieddy:

Percentages relative to Jan 4th remaining.

82% now....only 18% short of my prediction.

How about by Ghost and Goblin day?

Maybe by Firework Night.

[QUOTE=Qubit;377152]Is today this fireworks day you've mentioned?

July 1: Countdown to proving M(32582657) is the 44th Mersenne Prime: 2,234
(BTW, over what were you taking the percentages?)[/QUOTE]

2,234 more in the next two days... should be a piece of cake.

[QUOTE=Primeinator;377219]2,234 more in the next two days... should be a piece of cake.[/QUOTE]

Not all countries celebrate their "birthday" on July 4. Petrw1 and I, among a number of others I would imagine, are from Canada, one of those little countries that are sort of hard to find. "Fireworks Day" would likely have referred to yesterday.

TBH when IT first posted I meany July 4th knowing there are a lot more USA than Canada members.

[QUOTE=petrw1;377234]TBH when IT first posted I meany July 4th knowing there are a lot more USA than Canada members.[/QUOTE]Oh, so there are only two countries in the world?

[size=1][color=grey]And you didn't even mention the two countries with the highest population.[/color][/size]