| Author |
Message |
|---|
 Michael Goetz  Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
                                           
|
|
Effective January 1st, 2018, we will be changing the criteria for determining which prime discoveries get announced with news notification alerts, PDF documents, and inclusion on the front page and wiki.
Over the years, finding large primes has become easier and easier, and what was once a rare and special event is now business as usual. We want announcements to be special occurrences and not something that most people ignore. The big change is that merely being a mega prime is no longer sufficient to warrant an announcement. Being one of the top 100 largest known primes is the new criteria. This has the advantage of scaling automatically as computing power increases. If we changed the criteria to, say, 1.5 million digits, we'd probably be facing the same problem 5 years from now.
We will be making the following changes:
- The main criteria for determining what gets an announcement will be inclusion in the top 100 primes. As of right now, that means at least 1,379,838 digits, but this number will go up over time.
- In practical terms, this means PPS-MEGA and GFN-17-MEGA will no longer get announcements. Down the road, GFN-18 and SR5 may also drop below the threshold.
- Primes that do not make the Top 100 may still be announced if they are special for some other reason, at PrimeGrid's discretion. Twin primes from SGS would still get announced, for example.
- This affects the news announcements that get sent to the Notifications section of BOINC clients, the PrimeGrid Wiki, and both lists of significant primes on the home page.
- The automatic list of all recent primes in the rightmost column on the home page, which lists every new prime, will not change.
- Conjecture pages will still list every prime that eliminates a k, even if the prime doesn't get announced. Currently all conjectures produce primes in the top 100.
- AP26 discoveries will no longer be announced, or included on the front page, but they will continue to listed on the forums.
- Should we be fortunate enough to find an AP27, there will be a full announcement. I.e., the announcement policy for AP27 is unchanged.
- The policy concerning AP25 discoveries is under review. Currently, it gets a forum post, only.
(Updated: As of 2018, AP25s will no longer get a a forum posting. Ap 26 or longer will continue to get forum posts.)
- The forum pages that list all the GFN primes are unaffected by this change.
Note that the effective date applies to the date of discovery, not when the announcement would be done. A prime found on December 31st but reported in January would be treated under the old rules.
____________
My lucky number is 75898524288+1
|
|
|
|
|
|
I think, Primegrid need "month total announcement" in which you can write.
" In this month we found b SGS primes : link to forum thread with small announcements, like for GFN currently
c Proth primes : link to forum thread
d GFN primes : link to forum thread
e AP : link to forum thread
..."
Or we find primes too often?
____________
(252^6548-1)^2-2 is prime! Small, but mine.
134137784^32768+1(DC)
107853608^8192+1(DC)
10465966^16384+1(DC)
 |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
|
I think, Primegrid need "month total announcement" in which you can write.
" In this month we found b SGS primes : link to forum thread with small announcements, like for GFN currently
c Proth primes : link to forum thread
d GFN primes : link to forum thread
e AP : link to forum thread
..."
Or we find primes too often?
I do not plan on doing this since anybody can go to our prime lists and see what's been found recently.
But here the numbers for the past 31 days (this is a one time thing):
+---------+----------+
| project | count(*) |
+---------+----------+
| GFN | 51 |
| MEG | 5 |
| PPS | 10 |
| SGS | 21 |
+---------+----------+
GFN includes all GFN primes. MEG is PPS-MEGA. PPS is both PPS and PPSE.
Here's just mega primes:
+---------+----------+
| project | count(*) |
+---------+----------+
| GFN | 1 |
| MEG | 5 |
+---------+----------+
Here's primes that would be announced under the 2018 policy:
+---------+----------+
| project | count(*) |
+---------+----------+
| GFN | 1 |
+---------+----------+
____________
My lucky number is 75898524288+1 |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
|
|
Similar numbers for all primes since January 1st, 2017:
All primes:
+---------+----------+
| project | count(*) |
+---------+----------+
| GCW | 1 |
| GFN | 354 |
| MEG | 33 |
| PPS | 339 |
| PSP | 1 |
| SGS | 533 |
| SR5 | 2 |
+---------+----------+
Mega primes:
+---------+----------+
| project | count(*) |
+---------+----------+
| GCW | 1 |
| GFN | 18 |
| MEG | 33 |
| PSP | 1 |
| SR5 | 2 |
+---------+----------+
Announceable under 2018 rules:
+---------+----------+
| project | count(*) |
+---------+----------+
| GCW | 1 |
| GFN | 6 |
| PSP | 1 |
| SR5 | 2 |
+---------+----------+
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
When there will be a sum up of discovered primes per project and month/year, an additional column with the range (min/max) of the numbers would be interesting. |
|
|
|
|
|
I like it, I think AP 25's are still rare enough to remain under current policy. However I agree with the changes. |
|
|
|
|
|
2 thoughts:
- "Mega Primes Discovered" could be part of your announcement when it will reach a modulo 100 miletone (it should be one by year)
- Add a new counter : "Top 100 Mega Primes Discovered" on the main page and add an announcement for %10 (?)
____________
|
|
|
|
|
|
at least include a monthly update for the rest of discoveries. this way i have nothing to hope for. at least whenever i see an update i always wonder if it's me. :-D
|
|
|
mikey
Send message
Joined: 17 Mar 09
Posts: 2539
ID: 37043
Credit: 1,170,632,001
RAC: 272,999
                   
|
at least include a monthly update for the rest of discoveries. this way i have nothing to hope for. at least whenever i see an update i always wonder if it's me. :-D
If you have notifications turned on for your account then you will get notified by email.
|
|
|
|
|
|
19021 989031 3 |
|
|
|
|
|
Which sub-projects are good for a prime in the top 100 ? |
|
|
Honza Volunteer moderator
Volunteer tester
Project scientist
Send message
Joined: 15 Aug 05
Posts: 2082
ID: 352
Credit: 9,803,561,179
RAC: 3,049,999
                                             
|
Which sub-projects are good for a prime in the top 100 ?
See the project front page and left column named Prime Rank.
I would say SR5 be best choice - Top100 and still quite good probability of a prime.
____________
My stats |
|
|
|
|
|
And GFN 18 on GPUs
____________
(252^6548-1)^2-2 is prime! Small, but mine.
134137784^32768+1(DC)
107853608^8192+1(DC)
10465966^16384+1(DC)
|
|
|
pschoefer Volunteer developer
Volunteer tester
Send message
Joined: 20 Sep 05
Posts: 723
ID: 845
Credit: 4,052,327,053
RAC: 1,963,411
                  
|
I do not plan on doing this since anybody can go to our prime lists and see what's been found recently.
Is there any list that can be filtered or sorted by date of discovery? I thought about doing a monthly prime report containing the missing mega prime information and the numbers of smaller primes found per project myself, but there is no easy way to extract this information from the publicly available sources:
- double checker for PPS mega primes can only be found by carefully monitoring the Top Prime Finders list
- the only good way to get the hardware information for 'small' mega primes is hoping that the finder has only one active machine and has not hidden it
- some projects are not testing the candidates strictly from smallest to largest all the time, making it hard to count them in the Primes by Project list
____________
 |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
|
Is there any list that can be filtered or sorted by date of discovery?
This is possible, and something I've thought would be useful. I may add it at some point.
- double checker for PPS mega primes can only be found by carefully monitoring the Top Prime Finders list
This could be done as well.
- the only good way to get the hardware information for 'small' mega primes is hoping that the finder has only one active machine and has not hidden it
Impossible. We don't store that information with the tasks or the primes (and that's not going to change.) Or at least it probably won't change. No promises. Even if we were to make some rather significant changes to retain the hardware information, if there's multiple GPUs in a system, only the fastest GPU gets recorded by BOINC, so what's in the data might not be the GPU that actually did the work. If there's a GPU involved, we need to eyeball the stderr.txt file to look for the output from Genefer where it says what hardware is being used.
____________
My lucky number is 75898524288+1 |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
|
Is there any list that can be filtered or sorted by date of discovery?
This is possible, and something I've thought would be useful. I may add it at some point.
This is now live. Let me know if it's useful, etc.
____________
My lucky number is 75898524288+1 |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
|
- double checker for PPS mega primes can only be found by carefully monitoring the Top Prime Finders list
This could be done as well.
Double checkers can now be optionally displayed.
Thanks for the suggestions. I think both are useful.
Please note that the default is still to show the primes in size order, and double checker information is turned off by default. Both are selectable as options.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Is there an announcement policy for divisors of Fermat numbers?
For example, if a prime from PPSE (Proth Prime Search Extended) which is not in itself particularly huge, happens to be a factor of a classical Fermat number F_m, will that lead to an official announcement?
PrimeGrid is not finding many of those.
And then we have to ask about divisors of GF(m, b) = b^(2^m) + 1 or xGF(m, a, b) = a^(2^m) + b^(2^m) as well.
/JeppeSN |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
|
|
All Fermat divisors get an announcement, regardless of size. And yes, as you can tell with the date option, we don't find them very often.
Generalized and extended generalized Fermat divisors do not get announcements.
____________
My lucky number is 75898524288+1 |
|
|
|
|
All Fermat divisors get an announcement, regardless of size. And yes, as you can tell with the date option, we don't find them very often.
Generalized and extended generalized Fermat divisors do not get announcements.
I like the decision that Fermat divisors will be announced.
Chris Caldwell treats some GF divisors specially, in that he has a Top Twenty for them. Maybe we could have a rule that a prime that will make it to any of Caldwell's Top 20 lists, will get an announcement? Most of the xGF and GF divisors we find will not fall under that category, but GF divisors from a few specific bases (3,5,6,10,12) will.
/JeppeSN |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
|
|
Chris has his policies. We have ours. I see no reason for that particular policy to change.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Just curious, but are the private messages still being sent in the event you find a prime regardless of size? Wasn't sure if the notice system had changed that as well..
Cheers |
|
|
Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 14712
ID: 53948
Credit: 1,050,921,451
RAC: 384,030
|
Just curious, but are the private messages still being sent in the event you find a prime regardless of size? Wasn't sure if the notice system had changed that as well..
Cheers
Is the email account you have associated with your BOINC account correct, and operational? The email was sent to that account yesterday and according to Jim as of right now it hasn't bounced.
____________
My lucky number is 75898524288+1 |
|
|
RafaelVolunteer tester
Send message
Joined: 22 Oct 14
Posts: 1002
ID: 370496
Credit: 1,083,511,878
RAC: 180,601
       
|
Just curious, but are the private messages still being sent in the event you find a prime regardless of size? Wasn't sure if the notice system had changed that as well..
Cheers
Is the email account you have associated with your BOINC account correct, and operational? The email was sent to that account yesterday and according to Jim as of right now it hasn't bounced.
They seem to be working, I got one not too long ago. |
|
|
JimB Honorary cruncher
Send message
Joined: 4 Aug 11
Posts: 921
ID: 107307
Credit: 1,001,476,823
RAC: 0
       
|
|
And just to be clear, this isn't our server trying to send the mail. All @primegrid.com email accounts are actually gmail. I wouldn't expect other than the occasional intermittent problem that would soon be fixed without us doing anything. |
|
|
|
|
|
Sorry should be more specific. I meant the 'private message' on my acct. It has the inbox and outbox messages and in the past I would get a message. Not my outside email.
Cheers. And again no big deal. |
|
|
JimB Honorary cruncher
Send message
Joined: 4 Aug 11
Posts: 921
ID: 107307
Credit: 1,001,476,823
RAC: 0
|
|
Those have always been emails only. Nothing has changed for primes reportable to T5K. |
|
|
|
|
|
Ah maybe that's it. The last one I got was for a gfn 15 and I did get an email on my last one.
Thanks |
|
|
