| Author |
Message |
|---|
 Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
                               
|
|
With the discovery of the AP27, we've achieved this project's goal.
The question now is what to do with AP27, and that's a discussion going on behind the scenes.
Right now, I don't know what's going to happen, but I can tell you what I do know.
1) If you have an AP27 badge goal, now is the time to get that badge. Just in case.
2) If we decide to shut down AP27, I will give you the customary 30 day notice.
3) The 30 day clock is NOT running yet.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Respecting whatever decision is made the question is why does it have to be shut down?
Cheers Rick |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
Respecting whatever decision is made the question is why does it have to be shut down?
Cheers Rick
That's one of the things being discussed.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
What about AP28, AP29 etc? |
|
|
tngSend message
Joined: 29 Aug 10
Posts: 486
ID: 66603
Credit: 47,367,929,794
RAC: 27,840,572
                                                   
|
|
If it closes, is there another GPU subproject that might be activated on BOINC?
____________
 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
What about AP28, AP29 etc?
AP28 is one of the items being discussed.
AP29 is not possible with the current app, but might be possible with some changes.
It's expected that an AP28 would take about 10 times as hard to find as the AP27 using the current app.
____________
My lucky number is 75898524288+1 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
If it closes, is there another GPU subproject that might be activated on BOINC?
I don't expect a "replacement" for AP27, but it's always been our intention to eventually update WWWW and move it to BOINC. That's unrelated to whether or not AP27 continues running.
____________
My lucky number is 75898524288+1 |
|
|
|
|
to eventually update WWWW and move it to BOINC.
WWWW? |
|
|
|
|
|
It seems to be easier to find AP2x (for x < 27) than prime numbers. It seems to me therefore that this is an attractive project and closing it owuld make primegrid as a whole a bit less attractive. |
|
|
|
|
|
Too many projects closing! |
|
|
mackerelVolunteer tester
Send message
Joined: 2 Oct 08
Posts: 2645
ID: 29980
Credit: 568,565,361
RAC: 266
                    
|
|
My opinion: unless the server resources are needed to run something else, keep it going.
Is there any change needed to search for AP28 or could it fall out of the existing project anyway? |
|
|
|
|
My opinion: unless the server resources are needed to run something else, keep it going.
Is there any change needed to search for AP28 or could it fall out of the existing project anyway?
Or to put it another way, what makes AP27 different from other projects, for example SGS? "We've found at least one SGS prime so surely that project should close down now."
Obviously some projects are different. Sieving gets to a point where it only uses up CPU and serves no higher purpose. Other projects get to the point where they are more likely to error out than complete work units, let alone find a result. These are good reasons for closing a project that noone would argue against.
AP27 gives people a chance of saying "I found something." Probably a better chance than other projects. Surely that is worth keeping going until an actual negative reason comes up.
|
|
|
|
|
to eventually update WWWW and move it to BOINC.
WWWW?
Wall-Sun-Sun primes, Wieferich primes, Wilson primes, and Wolstenholme primes. /JeppeSN |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
Or to put it another way, what makes AP27 different from other projects, for example SGS? "We've found at least one SGS prime so surely that project should close down now."
That’s simply not true. All of the conjecture projects have defined ends. GCW has a defined end, although like AP it has the possibility of setting new goals, whereas the conjectures do not.
If we got lucky and today was “PrimeGrid PrimeDay”, and we found a prime for every k in the SoB, PSP, ESP, TRP, and SR5 projects, all 5 projects would end.
____________
My lucky number is 75898524288+1 |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 3207
ID: 130544
Credit: 2,285,465,802
RAC: 791,403
                     
|
|
I say keep it going & 'upgrade' the search to 28. |
|
|
|
|
|
Clearly enough, when a conjecture is proved, its subproject can close. However, AP27 has not proved any conjecture.
But where was the "goal" of the AP27 subproject officially declared? There are different interpretations:
- "Prove the conjecture that an AP27 exists." However, the fact that APk's exist for all k, is a theorem (Green-Tao), so no conjecture is proved by exhibiting an example.
- "Find (just) one AP longer than the longest known AP (which was an AP26)." If this was the goal, the goal has been reached. Finding another AP27 will not beat the (new) length record. However, finding an AP28 will beat the (new) length record. However, the name of the subproject was AP27, so you could argue that going for an AP28 is a different project? AP28s were mentioned in the thread start for the AP27 "descriptive" thread before the AP27 was located.
- "Find AP27s." I think some people thought of this as the goal, just like "find GFN18s" seems to be the goal of the GFN18 subproject. Some GFN18s are known; the search continues. In this case, the AP27 effort can go on.
/JeppeSN
|
|
|
|
|
Or to put it another way, what makes AP27 different from other projects, for example SGS? "We've found at least one SGS prime so surely that project should close down now."
That’s simply not true. All of the conjecture projects have defined ends. GCW has a defined end, although like AP it has the possibility of setting new goals, whereas the conjectures do not.
If we got lucky and today was “PrimeGrid PrimeDay”, and we found a prime for every k in the SoB, PSP, ESP, TRP, and SR5 projects, all 5 projects would end.
So I should have added conjecture projects to my list of exceptions. That still leaves about half that are just looking for various sorts of primes. So reiterate my question. What is the difference between SGS and AP27 that we keep one and not the other? |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
Or to put it another way, what makes AP27 different from other projects, for example SGS? "We've found at least one SGS prime so surely that project should close down now."
That’s simply not true. All of the conjecture projects have defined ends. GCW has a defined end, although like AP it has the possibility of setting new goals, whereas the conjectures do not.
If we got lucky and today was “PrimeGrid PrimeDay”, and we found a prime for every k in the SoB, PSP, ESP, TRP, and SR5 projects, all 5 projects would end.
So I should have added conjecture projects to my list of exceptions. That still leaves about half that are just looking for various sorts of primes. So reiterate my question. What is the difference between SGS and AP27 that we keep one and not the other?
That's part of the discussion we're having internally.
This is probably a good point to reiterate that no decision has been made yet. This is an ongoing (and rather complicated) discussion happening behind the scenes.
The purpose of this thread was to let you all know there's a possibility we may shut down AP27, in order to give everyone the maximum amount of time to get whatever badge you might be aiming for.
While I will correct statements that I think are factually incorrect, I'm not going to argue the merits of a decision that hasn't been made yet. I can't do that; I simply do not yet know what our position on this is.
____________
My lucky number is 75898524288+1 |
|
|
RenixSend message
Joined: 26 Aug 16
Posts: 346
ID: 455383
Credit: 4,161,145,672
RAC: 7,272,401
             
|
|
My thoughts center around that searching for an AP28 can be done with the current software as is. If project resources permit and don't impact the performance of other declared project goals, and folks still want to search for a 28, then let it keep going till one is found. Sorta like the keep interest in PrimeGrid going. :-) Just declare the new goal of finding an AP28 and then stop the AP searching at that point.
It would still be like "Lucky", searching for a world record, a long long shot, but still looking for a record. :-) |
|
|
|
|
|
I argue for leaving the project open. The GPU resources going towards it cannot be used on any of the conjecture projects, so one cannot make the argument that the compute resources would be better spent elsewhere. May as well leave them able to crunch this project.
____________
 |
|
|
|
|
|
Firstly I am a fan of the AP search, since I joined in April 19 I have concentrated on it. However I do sometimes feel a bit guilty about burning so much electricity, with it's associated negative effects. If there is a real mathematical benefit for AP28 then let's keep it going. If not an indulgence in gaining peoples's interest and badge collecting may not justify the CO2 produced??
____________
|
|
|
|
|
I say keep it going & 'upgrade' the search to 28.
+1
____________
Founder of CRUNCHERS SANS FRONTIERES
www.crunchersansfrontieres.org
CSF lucky number 22872882^65536+1 |
|
|
|
|
|
I am for closing AP27.
Why should we spend resources for that will never be found (about 30 years to search). |
|
|
|
|
I am for closing AP27.
Why should we spend resources for that will never be found (about 30 years to search).
Surely there are two sorts of resources here.
There are the core resources owned by primegrid that collects and stores the data sent to it by the decentralised computers. I can see you may have an argument here, but I would be surprised if the centralised part of AP work is some enormous drain on primegrid's resources.
Then there are the decentralised machines running work units and sending the results back to primegrid. These machines are not owned by primegrid. It is up to the owners of those machines (not primegrid) to determine if looking for arithmetic progressions of primes is a waste of reources.
Of course primegrid could choose to not offer the facility. But I have not yet seen a good reason why primegrid should shut down the AP project.
|
|
|
|
|
Firstly I am a fan of the AP search, since I joined in April 19 I have concentrated on it. However I do sometimes feel a bit guilty about burning so much electricity, with it's associated negative effects. If there is a real mathematical benefit for AP28 then let's keep it going. If not an indulgence in gaining peoples's interest and badge collecting may not justify the CO2 produced??
I don't see why this applies to AP and not all other parts of primegrid. |
|
|
Dave
Send message
Joined: 13 Feb 12
Posts: 3207
ID: 130544
Credit: 2,285,465,802
RAC: 791,403
|
|
While it's not inherently a green project, one could argue if one's PG priority was to lower emissions then MSR is the best policy as this reduces the amount of GFN necessary. Regardless, who knows what beneficial offshoots there may be because of our work. There are far more damaging DC 'projects' than ours out there...! |
|
|
|
|
|
Slightly offtopic: We found an AP26 today: https://www.primegrid.com/ap.php?userid=880615&display_length=26
See our other AP26s in the thread New APs. Also see stats_ap26.
/JeppeSN |
|
|
|
|
|
Lets go for 28 & 29. In 10 months new GPU's will make it worth it :) |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
Lets go for 28 & 29. In 10 months new GPU's will make it worth it :)
29 is not possible with the current software.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Jordan Romaidis, nice new badge 26 there! /JeppeSN |
|
|
|
|
|
And the fact that the special significant discovery badge (27) was introduced as a so-called counter badge (a badge that allows "cardinality" so you can have it multiple times), shows that it is perceivable that someone finds more than one AP27. Of course, even if the AP subproject will actually be shut down soon, with his luck, I am sure Robish can manage to find another AP27 before the end. /JeppeSN |
|
|
|
|
Lets go for 28 & 29. In 10 months new GPU's will make it worth it :)
29 is not possible with the current software.
I know, but at-least 28 is. Here is hoping for 29 one day :)
Jordan Romaidis, nice new badge 26 there! /JeppeSN
Thanks! Seems like it took forever! :) |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
Lets go for 28 & 29. In 10 months new GPU's will make it worth it :)
29 is not possible with the current software.
I know, but at-least 28 is. Here is hoping for 29 one day :)
Jordan Romaidis, nice new badge 26 there! /JeppeSN
Thanks! Seems like it took forever! :)
https://www.primegrid.com/forum_thread.php?id=7012&nowrap=true#133439
Congratulations!
____________
My lucky number is 75898524288+1 |
|
|
robishVolunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 2210
ID: 126266
Credit: 7,514,885,474
RAC: 3,446,662
           
|
And the fact that the special significant discovery badge (27) was introduced as a so-called counter badge (a badge that allows "cardinality" so you can have it multiple times), shows that it is perceivable that someone finds more than one AP27. Of course, even if the AP subproject will actually be shut down soon, with his luck, I am sure Robish can manage to find another AP27 before the end. /JeppeSN
Ha. No don't think it needs to be a counter :)
but there ARE MORE out there........what's another 3 years between friends? ;) or 6?
or maybe an AP28 even .....🤓
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 |
|
|
|
|
|
I would like to see this subproject keep running. Even with me not haven't found an AP27 (yet), it was fun to find some shorter ones.
So if I could make a wish: Let's got for AP28 or AP29!
May the force be with us. Maybe we're lucky.
If AP28 or AP29 is not started, just keep AP27 running as long as it doesn't hurt the server.
I think there are enough people out the who like it, even when it has archived its primary goal. |
|
|
GregCSend message
Joined: 12 Nov 18
Posts: 52
ID: 1077873
Credit: 1,892,987,582
RAC: 2,451,854
   
|
|
If the software can't do an AP29, is an AP28 possible? If so, I vote we keep going. There are more AP27s out there still. Just sayin'. |
|
|
|
|
If the software can't do an AP29, is an AP28 possible? If so, I vote we keep going. There are more AP27s out there still. Just sayin'.
I am wondering about the nature of the limitation on 29. Would looking for AP29's require bigger work units and so may be it would be appropriate to move the project to GPU only if the required software changes were made?
|
|
|
|
|
If the software can't do an AP29, is an AP28 possible? If so, I vote we keep going. There are more AP27s out there still. Just sayin'.
I am wondering about the nature of the limitation on 29. Would looking for AP29's require bigger work units and so may be it would be appropriate to move the project to GPU only if the required software changes were made?
AP29 requires difference divisible by 29 which makes the search 29 times harder.
____________
|
|
|
|
|
If the software can't do an AP29, is an AP28 possible? If so, I vote we keep going. There are more AP27s out there still. Just sayin'.
I am wondering about the nature of the limitation on 29. Would looking for AP29's require bigger work units and so may be it would be appropriate to move the project to GPU only if the required software changes were made?
AP29 requires difference divisible by 29 which makes the search 29 times harder.
This does not really answer the question. Would looking for AP29 require
1.) a small tweak to the software
2.) a major tweak to the software on account of the software's cleverness
3.) a parameter change (presumably not based upon previous comments)?
4.) The size of the numbers involved makes it impractical even with software changes and restricting to GPU.
What would be the effects on the running of the packages if this change was done? Would the run time massively increase? (I guess x29 which would be fine many users but not me.) Would it become impractical for CPU machines?
As far as I can see previous work would not be invalidated. I suppose one might want to recheck any AP28's to see if they were actually AP29s. But currently that would not take very long. |
|
|
|
|
|
I have dreamed about WSS search coming back! To me that is the ultimate conjecture, where a prime of this form has not even been proven to exist yet. Actually finding one would be simultaneously become a proof.
____________
Proud member of Team Aggie the Pew
"Wir müssen wissen. Wir werden wissen."
"We must know, we shall know."
- David Hilbert, 1930 |
|
|
|
|
I have dreamed about WSS search coming back! To me that is the ultimate conjecture, where a prime of this form has not even been proven to exist yet. Actually finding one would be simultaneously become a proof.
Have the bugs in WWWW been fixed? Weren’t all of our results from the last search completely inaccurate?
____________
|
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
I have dreamed about WSS search coming back! To me that is the ultimate conjecture, where a prime of this form has not even been proven to exist yet. Actually finding one would be simultaneously become a proof.
Have the bugs in WWWW been fixed? Weren’t all of our results from the last search completely inaccurate?
Please move the WWWW discussion to a more appropriate location. Thank you!
____________
My lucky number is 75898524288+1 |
|
|
RenixSend message
Joined: 26 Aug 16
Posts: 346
ID: 455383
Credit: 4,161,145,672
RAC: 7,272,401
|
|
Any new hot gossip on what the future holds for the AP27 and AP28 effort??? |
|
|
|
|
Any new hot gossip on what the future holds for the AP27 and AP28 effort???
That you saw fit to give me that sinking feeling, apparently. ;-) |
|
|
BurVolunteer tester
Send message
Joined: 25 Feb 20
Posts: 515
ID: 1241833
Credit: 414,481,880
RAC: 219
         
|
|
Has anything come of the discussion whether or not to keep AP-27 running? Or looking for AP-28?
While I think it's interesting to find a second AP-27, I don't think the comparison to other subprojects is quite ok. Finding another larger SG-Prime or twin prime is an achivement, since larger SG primes are generally considered "more worthwhile".
However, APs I would say, the smaller the involved numbers, the better. Otherwise, we could all just look for the largest AP-3. So the next AP-27 would be less "good" than the current, which sets it apart from SGS and GFN type primes.
Of course, this is a sort of emotional argument, so I'd be interested how others perceive this ranking of APs.
____________
1281979 * 2^485014 + 1 is prime ... no further hits up to: n = 5,700,000 |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14011
ID: 53948
Credit: 433,187,806
RAC: 1,006,661
|
Has anything come of the discussion whether or not to keep AP-27 running? Or looking for AP-28?
There hasn't been any discussion recently.
____________
My lucky number is 75898524288+1 |
|
|
