mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Marin's Mersenne-aries (https://www.mersenneforum.org/forumdisplay.php?f=30)
-   -   [Mission Accomplished] Let's finish primality verification through Mp#48*, M(57 885 161) (https://www.mersenneforum.org/showthread.php?t=26591)

Prime95 2021-10-01 19:55

[QUOTE=kriesel;589154]If not cleared soon...[/QUOTE]

... the world will not end.

chalsall 2021-10-01 20:29

[QUOTE=Prime95;589157]... the world will not end.[/QUOTE]

+1.

greenskull 2021-10-01 22:22

[QUOTE=Prime95;589157]... the world will not end.[/QUOTE]

Gold words! :)

Dr Sardonicus 2021-10-02 02:21

[QUOTE=Prime95;589157][QUOTE=kriesel;589154]If not cleared soon...[/QUOTE]... the world will not end.[/QUOTE]:missingteeth: :missingteeth: :missingteeth: :missingteeth:

:tu:

Besides - even if the world [i]does[/i] end, nobody will be complaining about it...

greenskull 2021-10-03 01:14

[QUOTE=Dr Sardonicus;589188]Besides - even if the world [i]does[/i] end, nobody will be complaining about it...[/QUOTE]

The transition of the world from the state of its existence to its end is a continuous process lasting non-zero time.
And in the process of this transition, as its state changes, there will be complains from some witnesses of this process.

And the reason for that is a banal limitation of the signal transmission rate.
It seems to me that a binary transition is possible only from a singularity state.

I think November 16 is the right date to end the uncertainty of the M57885161 status.

tuckerkao 2021-10-03 03:28

It seemed like that the LL tester of M[M]57773081[/M] paused at 26% and already 3 days passed the expected finishing date while still has 20 days left before the assignment expires of now.

greenskull 2021-10-03 10:15

The LL tester of M57773081 is in deep meditation and comprehends the higher meanings of being.
Let's not rush and fuss. We're here not for time, ... but for lessons ...

It is not me. I don't know him.
But it increases the validity of my forecast for November 16th.
And interference with the natural course of events makes the game unfair.

Dr Sardonicus 2021-10-03 13:16

[QUOTE=greenskull;589262]<snip>
And interference with the natural course of events makes the game unfair.[/QUOTE]This course of events is in no way "natural." It is part of a man-made project devised and implemented in the pursuit of an objective that exists entirely in the human mind.

As I understand it, once the user starts a computation, then by default it continues to completion, updating automatically every day.

The only reasons I know for a computation to "pause" are (1) some kind of system failure, and (2) human interference. [If there are other reasons I don't know about, please enlighten me.]

Case (1) would include hardware failures in the machine doing the computation, or in the system that connects it with the network; or something (e.g. a power outage) causing the computation or its connection to the network to halt or abort. Any such occurrence would IMO warrant "interference."

In case (2), the "interference" has already occurred, and warrants corrective action.

I note that once the computation has begun, the user could meditate, hang upside-down, or do any number of other things without affecting - at least directly - the course of the computation.

Perhaps the user has gone on vacation or is absent for another reason; and would, if present, have done something to resume the task. In that case, having someone else do that in the user's stead - or if the computation is FUBAR reassigning the task - would IMO be entirely appropriate.

greenskull 2021-10-03 13:46

[QUOTE=Dr Sardonicus;589275]In that case, having someone else do that in the user's stead - or if the computation is FUBAR reassigning the task - would IMO be entirely appropriate.[/QUOTE]
It seems to me that this is a mortal sin.

This approach can eventually become systemic and discredit some project participants whose computers are not fast enough or the participant does not devote 100% of his time to computing.
This is a club of like-minded people. It seems to me that army discipline is not appropriate here.
Or the project needs to tighten the general rules for all, and not solve the issue of a particular case.

I have no claim to be right. I have been watching what is happening for a while and I’m just presenting an idea that makes sense.

kriesel 2021-10-03 14:38

There are conceivable scenarios where the human interference is from other than the assigned user, and unintentional as far as GIMPS goes. Hypothetical: AndersM as a college student home for the summer leaves his Cat0-qualified computer running 24/7 doing prime95 work. A few weeks after he returns for the fall semester, one of his parents finds a computer in his bedroom still running, and turns it off to save electricity. No more progress, no more updates, until he returns for winter break and turns it on again. All well intentioned, yet the progress is delayed, and the computer loses Cat0 qualification for several months.
Or he got stuck in a quarantine or Kabul, hasn't been able to pay his rent or communicate with a landlord, it was interpreted as a breach of lease, and all his stuff got packed up into a storage locker so the apartment could be re-rented. (A coworker of mine returned from a long trip to China to discover his apartment had been re-leased and his contents had been [B]sold[/B].)
Or IT did system updates which required a reboot, and he's on vacation or working from home so won't be in the office for days or weeks to notice prime95 is no longer running on his office PC.

greenskull 2021-10-03 15:13

I consider it unacceptable to break or change the rules during the game, especially when the bets are made.

Dr Sardonicus 2021-10-03 15:39

[QUOTE=greenskull;589276][QUOTE=Dr Sardonicus;589275]In that case, having someone else do that in the user's stead - or if the computation is FUBAR reassigning the task - would IMO be entirely appropriate.[/QUOTE]It seems to me that this is a mortal sin.[/QUOTE]I don't know about "mortal sin." If something were done to resume the calculation without trying to find out why it's paused, I'd say that wouldn't be right.

But I didn't advocate taking such action without first trying to find out what's going on. Since you apparently missed it the first time, [QUOTE=Dr Sardonicus;589135]Perhaps somebody could get in touch with AndersM for a sitrep...[/QUOTE]If the user investigates and decides what should be done, and delegates the action to others, I don't see anything wrong with that. If anything like the scenarios in [url=https://www.mersenneforum.org/showpost.php?p=589285&postcount=107]this post[/url] has happened, and the user says, "Sorry, I can't finish, please reassign" I'd say that's the right thing to do. Or if the user is absent and tells someone to check, and, if it's something like the network connection being interrupted to go ahead and fix it, I don't see anything wrong with that, either.

retina 2021-10-03 16:05

I think the exponent has turned out to make a Mersenne prime, and there is a huge conspiracy among the elites to suppress the knowledge until they think the rest of the world can handle it without panic. :evil:

All the talk about power failures, or connection issues are just a ruse to throw us off the scent, right?

:razz:

greenskull 2021-10-03 16:19

[QUOTE=retina;589294]I think the exponent has turned out to make a Mersenne prime, and there is a huge conspiracy among the elites to suppress the knowledge until they think the rest of the world can handle it without panic. :evil:

All the talk about power failures, or connection issues are just a ruse to throw us off the scent, right?

:razz:[/QUOTE]
Why are you revealing all the secrets?

Prime95 2021-10-03 18:52

[QUOTE=greenskull;589262]
But it increases the validity of my forecast for November 16th.
And interference with the natural course of events makes the game unfair.[/QUOTE]

If you or Ken were to poach that would be unfair in your entertaining prediction game. Historically speaking, it is likely one or two users have already noticed the stalled exponent are busily poaching as we converse.

greenskull 2021-10-03 19:40

[QUOTE=Prime95;589317]If you or Ken were to poach that would be unfair in your entertaining prediction game. Historically speaking, it is likely one or two users have already noticed the stalled exponent are busily poaching as we converse.[/QUOTE]

Please no!

kriesel 2021-10-03 20:16

1 Attachment(s)
[QUOTE=Prime95;589317]If you or Ken were to poach[/QUOTE]greenskull has no motivation to poach. Getting an assignment under another alias and stalling it, or getting such deceptive help from a friend, would help move the milestone toward his prediction of 22 November or revised to 16 November. Another expiration/assignment cycle would be needed to reach them.
If [M]57773081[/M] expires, gets reassigned, and runs to completion in the usual Cat0 ~7 days, that would be ~2021-10-30. If it's resumed, or poached, it could be completed any day now. George knows, but some others may not realize, that such a run is ~7 hours on a Radeon VII.

[QUOTE=greenskull;589276]It seems to me that this is a mortal sin.[/QUOTE]Not likely. Unless you know of skullduggery that we don't. Assignments can be and have been transferred from one participant to another by mutual agreement. [URL]https://en.wikipedia.org/wiki/Mortal_sin[/URL]

[QUOTE=greenskull;589288]I consider it unacceptable to break or change the rules during the game, especially when the bets are made.[/QUOTE]What game, what bets? You stated you did not bet a kidney.

For the 56M milestone, it went from 21 left to 0 in 12 days. For Mp#48*, it was at 21 left at ~2021-09-19. By that measure, the milestone is overdue (2021-10-01). To have one or two stall out near the finish line is unusual.

greenskull 2021-10-03 20:21

There are an infinite number of unverified exponents.
The sky will not fall to the ground if you choose any of them that are free.

We are cultured civilized people with ethical principles. Why behave like cavemen, eating each other for handful of shells?

Uncwilly 2021-10-04 06:07

If I were to poach the exponent, I might wait to turn in the results until the date exactly between the 2 predictions.....

greenskull 2021-10-04 09:17

[QUOTE=Uncwilly;589359]If I were to poach the exponent, I might wait to turn in the results until the date exactly between the 2 predictions.....[/QUOTE]
Do not do this pls :)
If you are itching to manipulate it, then try to hit exactly on November 16.

kriesel 2021-10-04 12:26

1
 
3 Attachment(s)
[QUOTE=Uncwilly;589359]I might wait to turn in the results until the date exactly between the 2 predictions.....[/QUOTE]Waiting longer than you might usually, would be rigging the outcome. (I suppose that could apply to the rest of us too...)

Now only AndersM's assignment on [M]57773081[/M] remains, which is still stalled at 26%(since 2021-09-27 last update a week ago).

From [URL]https://www.mersenne.org/report_milestones/:[/URL]
[B]Progress toward next GIMPS milestones [I](last updated 2021-10-04 12:15:09 UTC, updates every 15 minutes)[/I][/B]

[LIST][*] All exponents below [COLOR=blue]57 773 081[/COLOR] have been tested and verified.[*] All exponents below [COLOR=blue]104 815 069[/COLOR] have been tested at least once.[*] Countdown to first time checking all exponents below 105 million: [URL="https://www.mersenne.org/assignments/?exp_lo=104815069&exp_hi=105000000&execm=1&exp1=1&extf=1&exdchk=1&excert=1"]2[/URL][*] Countdown to first time checking all exponents below 106 million: [URL="https://www.mersenne.org/assignments/?exp_lo=104815069&exp_hi=106000000&execm=1&exp1=1&extf=1&exdchk=1&excert=1"]73[/URL][*] Countdown to verifying all tests below M(57 885 161): [URL="https://www.mersenne.org/assignments/?exp_lo=57773081&exp_hi=57885161&execm=1&exp1=1&extf=1&exfirst=1"]1[/URL][*] Countdown to verifying all tests below 58 million: [URL="https://www.mersenne.org/assignments/?exp_lo=57773081&exp_hi=58000000&execm=1&exp1=1&extf=1&exfirst=1&excert=1"]3[/URL][*] Countdown to verifying all tests below 59 million: [URL="https://www.mersenne.org/assignments/?exp_lo=57773081&exp_hi=59000000&execm=1&exp1=1&extf=1&exfirst=1&excert=1"]1 624[/URL] [COLOR=black](994 available)[/COLOR][*] Countdown to verifying all tests below M(74 207 281): [COLOR=blue]269 225[/COLOR][*] Countdown to verifying all tests below M(77 232 917): [COLOR=blue]324 747[/COLOR][*] Countdown to verifying all tests below M(82 589 933): [COLOR=blue]423 450[/COLOR][/LIST]

kriesel 2021-10-06 08:56

zero, milestone reached
 
4 Attachment(s)
1 LL and 1 PRP on [M]57773081[/M]. Server called it done. Milestone reached.
So just for good measure, I subsequently reported in an LLDC run earlier, that confirmed Sven Laubner's LL res64.
M[M]57885161 [/M]is confirmed as the 48th Mersenne prime.

On to [URL="https://www.mersenne.org/report_exponent/?exp_lo=74207281&exp_hi=&full=1"]Mp#49*[/URL]. That will take a while.

greenskull 2021-10-06 09:06

Oh, no!
The game was not fair :evil: But it was excitingly interesting :banana:

What are the timing assumptions for the next exponent: 74,207,281?

firejuggler 2021-10-06 09:15

First post on march with 20k test left.
so that's about 6 month.
number of test left to MP#49 is around 270 k
Thats 13.5 * 6month... lets add a bit of time because tests take longer.. and we're looking at 7 or 8 years.

kriesel 2021-10-06 09:38

[QUOTE=greenskull;589605]The game was not fair[/QUOTE]Nonsense. The data you graphed and used to make your predictions included the effects of whatever poaching occurred along the way to those data points. You chose an obvious but inferior metric, and analysis that failed to account for the nonlinearity built into the assignment rules. Statistical analysis and a linear fit, not understanding of the actual mechanisms. Don't whine, learn from it. Read and understand posts [URL="https://mersenneforum.org/showpost.php?p=582097&postcount=18"]18[/URL]-20 and several following explanatory posts.

greenskull 2021-10-06 10:38

Let's start a new thread about M74207281.

[QUOTE=kriesel;589610]Nonsense. The data you graphed and used to make your predictions included the effects of whatever poaching occurred along the way to those data points. You chose an obvious but inferior metric, and analysis that failed to account for the nonlinearity built into the assignment rules. Statistical analysis and a linear fit, not understanding of the actual mechanisms. Don't whine, learn from it. Read and understand posts [URL="https://mersenneforum.org/showpost.php?p=582097&postcount=18"]18[/URL]-20 and several following explanatory posts.[/QUOTE]

Yes, I have hyperopia.
I see worse up close, but very good into the distance :)

firejuggler 2021-10-06 10:55

[url]https://mersenneforum.org/showthread.php?t=27190[/url]

greenskull 2021-10-06 11:14

[QUOTE=firejuggler;589613][url]https://mersenneforum.org/showthread.php?t=27190[/url][/QUOTE]
:thumbs-up:
I'm already there and in the game.

kriesel 2021-10-06 14:31

[QUOTE=greenskull;589611]I see worse up close, but very good into the distance :)[/QUOTE][URL="https://idioms.thefreedictionary.com/there+is+none+so+blind+as+those+who+will+not+see"]"very good"[/URL] until irrefutably proven otherwise. Yet the denial continues.

greenskull 2021-10-06 14:34

[QUOTE=kriesel;589637][URL="https://idioms.thefreedictionary.com/there+is+none+so+blind+as+those+who+will+not+see"]"very good"[/URL] until irrefutably proven otherwise. Yet the denial continues.[/QUOTE]

I'm not interested in competing in demagoguery. I'm interested to see your M#49 forecast :)

kriesel 2021-10-06 16:31

[QUOTE=greenskull;589639]I'm not interested in competing in demagoguery. I'm interested to see your M#49 forecast :)[/QUOTE]Projections for Mp#49*-51* were previously [URL="https://www.mersenneforum.org/showpost.php?p=502995&postcount=11"]posted[/URL] 2021-08-26 and you were notified of that. See page 5 of second attachment for projected dates and background. I don't see a point to spending any time attempting to refine it after so brief an interval. Nor a point in engaging with you further on the topic, as it is unproductive.

Prime95 2021-10-06 17:22

[QUOTE=kriesel;589604]1 LL and 1 PRP on [M]57773081[/M]. Server called it done. Milestone reached.[/QUOTE]

As I predicted, a poacher was destined to intervene before the exponent expired.

Congrats to all the hard-working DCers!!

kriesel 2021-10-06 18:40

Mp48 milestone appears to have been delayed relative to trend of recent DC milestones
 
5 Attachment(s)
Comparing the last 12 days before reaching the 54M, 55M, 56M, 57M, and Mp48, respectively, indicates a trend toward more DC results submitted in the final days each time, except for Mp48 there was a decline. Some possibilities why that may be so:
Higher Mp means higher run time. Not clear why that would matter at Mp48 but not before.
Increased emphasis on not poaching for the Mp48 milestone.
Possibly change in number of users or systems doing LLDC generally.
Probably there are more.

To follow up on deltas of lowest remaining exponent below a milestone, the last was 112080.

lycorn 2021-10-06 20:41

[QUOTE=Prime95;589661]As I predicted, a poacher was destined to intervene before the exponent expired.

[/QUOTE]

Quite frankly, I can´t see the point in poaching an exponent that had been released to the legit user only 13 days ago. Aren´t the expiry rules supposed to be followed?
If I were user AndersM I would be really p***ed off. I find these poaching initiatives more detrimental than beneficial to the project, but maybe it´s just me :alien:

kriesel 2021-10-06 21:30

No matter where the expiration duration is set, or how well people resist poaching, there will be triple matching results on occasion. Take a look at the attachments of post 130. Most of the last 10 results are usually third matching res64. In the future PRP/GEC/proof will help reduce that because certs issue with priority and complete very fast.

lycorn 2021-10-07 21:34

Correct, but you apparently missed my point. We have a set of expiration rules, so we should abide by them, as not following the agreed rules is worse than not having rules at all. We may be driving people away from the project by doing this sort of things. That's what I classify as "detrimental", not some time occasionally wasted on a unneeded TC.

sweety439 2021-10-17 15:58

OEIS sequence [URL="https://oeis.org/A000043"]https://oeis.org/A000043[/URL] has not been updated by adding a proven term a(48) = 57885161

kriesel 2021-10-17 16:44

[QUOTE=sweety439;590862]OEIS sequence [URL]https://oeis.org/A000043[/URL] has not been updated by adding a proven term a(48) = 57885161[/QUOTE]That's been mentioned a few times in another thread, most recently [url]https://mersenneforum.org/showpost.php?p=590548&postcount=34[/url]

kriesel 2021-10-17 16:48

[QUOTE=lycorn;589803]Correct, but you apparently missed my point. We have a set of expiration rules, so we should abide by them, as not following the agreed rules is worse than not having rules at all. We may be driving people away from the project by doing this sort of things.[/QUOTE]Apparently you missed my point. People do not perfectly follow rules. And even if they did, due to other factors, there would be occasional unneeded tests completed or other mechanisms of cycles lost, or other points of irritation. I get your point that it may cause some to leave the project. So may a high level of enforcement or complaint about imperfect enforcement of a rule set, cause some others to leave. It's unclear where the optimum lies.

tuckerkao 2021-11-02 04:40

[QUOTE=kruoli;589937]Yes, we will never know what would have happened if the two (!) offenders (:wink:) would not have stepped in. But as George said, it was likely that someone intervened here (unfortunately?).[/QUOTE]
We've already known the original result too because the legit user turned in his assignment late.

If you check M[M]57773081[/M] closely, ViliamF trial factored all the way up to 2^80, then Kriesel turned in the LL result 2 days later, they were in the same team. 3 different groups were working on the same exponent during Oct 2021 because a lot of people wanted to move the milestone.

Old Group from year 2012: Craig Meyers, Sven Laubner
Group 1: ViliamF, Kriesel
Group 2: Tucker Kao, timothee.chaumier
Group 3: Jan S., AndersM

[QUOTE=greenskull;589605]Oh, no!
The game was not fair :evil: But it was excitingly interesting :banana:[/QUOTE]
Nov 16 2021 is still far off from Oct 21, 2021


All times are UTC. The time now is 21:50.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.