Author |
Message |
|
Hi @all,
I'm collecting for about 3 1/2 years now all primes of the form k*2^n-1 (Riesel-side, see also http://www.rieselprime.de).
As PG is crunching and finding more and more the Proth-side (k*2^n+1) I've decided to collect and create pages for those, too.
For the first step I've used W.Kellers page www.prothsearch.com for 1<k<1200 and n<640k.
PG now has done much work and also doublechecked a lot of ranges and k-values but I'm missing some detailed information here.
At http://www.primegrid.com/stats_pps_llr.php it would be helpful to see:
- n-range of search done
- primes found (n-values)
- prime is new or confirmed (doublechecked)
After reading this forum I found the lowest range for 1<k<1200 at n=200k. All other k-values the n-range seems quite higher (~500k).
The search for lower primes n>=100k for k>1200 at PPSE is also rare of result-information:
- Which n-ranges were searched?
- List of primes found
- New / confirmed primes
I know about 13400 Top5000 primes of PG, about 6200 conf./new at Proth-type.
Question:
If a non-Top5000 prime is found, where can I get the information?
I seek for all primes wheater Top5000 or not!
I'm also aware of the test-server for n>50k for 1200<k<10000 and just testing all n<50k by myself!
Thanks a lot and I'm hoping to get more results for those Proth-list.
K.Bonath |
|
|
Honza Volunteer moderator Volunteer tester Project scientist Send message
Joined: 15 Aug 05 Posts: 1893 ID: 352 Credit: 3,142,312,174 RAC: 0
                             
|
Well, there needs to be a transport of primes found on PRPNet that incorporates them into Primegrid database.
(primes found on PRPNet do not appear on Top Prime Finders list, I've been asking for this several times).
____________
My stats
Badge score: 1*1 + 5*1 + 8*3 + 9*11 + 10*1 + 11*1 + 12*3 = 186 |
|
|
John Honorary cruncher
 Send message
Joined: 21 Feb 06 Posts: 2875 ID: 2449 Credit: 2,681,934 RAC: 0
                 
|
At http://www.primegrid.com/stats_pps_llr.php it would be helpful to see:
- n-range of search done
On that page you can see the max n completed. For k<1200, n min is 1. For 1200<k<10000, n min that was consecutively tested in BOINC is n=500K. Below that there are many gaps as ranges were skipped.
As for the manual and PRPNet testing of 1200<k<10000, a lot of the early work was lost due to a HD crash. Therefore, that's why you see the new PRPNet 4.0.0 test server starting out at n=50K.
- primes found (n-values)
a nice feature to have (which we don't) is to be able to click on k and have it list n's where it is prime. :)
Question:
If a non-Top5000 prime is found, where can I get the information?
Currently, there's no available public access to this information for primes found in BOINC. As for smaller primes in PRPNet, all you have to do is access the primes page of each port for a list of all the primes.
Most of the primes found in BOINC are listed in the Prime Pages. However, there were a few times where the leading edge of the search fell below the 5000th prime and thus was not reported.
Thanks a lot and I'm hoping to get more results for those Proth-list.
Creating a public proth primes database has been a pipe dream for the past TWO years. Hopefully now with your help, it can become a reality. :)
____________
|
|
|
|
As for smaller primes in PRPNet, all you have to do is access the primes page of each port for a list of all the primes.
Could you please list the link you mean?
Most of the primes found in BOINC are listed in the Prime Pages. However, there were a few times where the leading edge of the search fell below the 5000th prime and thus was not reported.
So the non-Top5000 primes are all lost?
Creating a public proth primes database has been a pipe dream for the past TWO years. Hopefully now with your help, it can become a reality. :)
I've done some scripts to generate the html-pages automatically but sure I need the information first.
For a first sight the data from W.Kellers page can be found under Proth k=1-1200.
But not yet all information (primes linked to Top5000, current reservations) are shown. More work to do the next time.
Karsten
|
|
|
John Honorary cruncher
 Send message
Joined: 21 Feb 06 Posts: 2875 ID: 2449 Credit: 2,681,934 RAC: 0
                 
|
As for smaller primes in PRPNet, all you have to do is access the primes page of each port for a list of all the primes.
Could you please list the link you mean?
http://prime2u.com:11000/user_primes.html
http://uwin.mine.nu:10000/user_primes.html
NOTE: the decimal length is buggy.
Most of the primes found in BOINC are listed in the Prime Pages. However, there were a few times where the leading edge of the search fell below the 5000th prime and thus was not reported.
So the non-Top5000 primes are all lost?
No, they are in the PG database, just not listed at the Primes Page.
____________
|
|
|
|
I've uploaded a text-file with all PG primes found so far to http://www.rieselprime.de/dl/PG_Primes.txt (800 kB).
15556 primes including:
- all Top5000 primes upto 2010-10-16
- all primes from PG-Database like http://www.primegrid.com/primes/?section=decimal&primeid=1 up to id 13518
- all small primes from port 11000
Not yet included:
- primes from port 10000 (not available)
- some listed small primes in PRPnet-thread
Karsten |
|
|
Honza Volunteer moderator Volunteer tester Project scientist Send message
Joined: 15 Aug 05 Posts: 1893 ID: 352 Credit: 3,142,312,174 RAC: 0
                             
|
Another reason for having primes from PRPNet (and elsewhere) in one place is to eliminate duplication. This week I've re-discovered prime on PG already found by someone else (outside PG).
Good stuff, Karsten.
____________
My stats
Badge score: 1*1 + 5*1 + 8*3 + 9*11 + 10*1 + 11*1 + 12*3 = 186 |
|
|
|
Another reason for having primes from PRPNet (and elsewhere) in one place is to eliminate duplication. This week I've re-discovered prime on PG already found by someone else (outside PG).
It's annoying to find a prime and be aware it's only rediscoverd.
But: A doublecheck is important, because of bugs in older versions of primeproving programs.
Example: S.Harvey found 2 missing 'small' primes in 2009:
29*2^347896-1 and 27*2^282700-1 when the search ranges were n>1M in those days!
For the current Proth-testing the Top5000 Database got these already known primes (1200<k<10000 and n>560k):
2607*2^570948+1
6465*2^2574035+1
2607*2^601176+1
and no other for n<602000.
|
|
|
John Honorary cruncher
 Send message
Joined: 21 Feb 06 Posts: 2875 ID: 2449 Credit: 2,681,934 RAC: 0
                 
|
Another reason for having primes from PRPNet (and elsewhere) in one place is to eliminate duplication. This week I've re-discovered prime on PG already found by someone else (outside PG).
It's annoying to find a prime and be aware it's only rediscoverd.
But: A doublecheck is important, because of bugs in older versions of primeproving programs.
Double checking is fine. However, what Honza describes is actually a very annoying triple check. Everyone agrees that double checks are a necessary step in prime finding. It's the triple check that's a waste.
A "unified" primes DB would go a long ways in helping us avoid those triple checks. The elusive two year dream is finally starting to take shape. :)
____________
|
|
|
|
Here's another example for the need of such Database I've just found:
The Top5000 Database from Mr. Caldwell contains a doubled entry, namely:
155*2^67973+1:
http://primes.utm.edu/primes/page.php?id=62875
http://primes.utm.edu/primes/page.php?id=8508
So it's undiscoverd for ten years now!?
I've mailed Mr.Caldwell.
|
|
|
|
By now all Top5000 links for k<1000 are now available at http://www.rieselprime.de/Data/P00001.htm.
Statistics for k<1200:
- 19375 primes listed
- 5535 Top5000 links
Next updates to come:
- all twins for k<1200 marked (perhaps today)
- all primes / Top5000 links for k<10000 ! (next week, hope so)
(When I'm ready with testing all 1200<k<10000 and n<50000; about 15% done)
- all PG PRPnet primes for n>50000 (as so far available/known) (next week)
- current reservations (next weeks)
- changing menu / adding stats to http://www.rieselprime.de for Proth-side (next week)
Karsten
|
|
|
HAmsty Volunteer tester
 Send message
Joined: 26 Dec 08 Posts: 132 ID: 33421 Credit: 12,510,712 RAC: 0
                
|
http://vcn94.homelinux.org/PRPNet/primelist.html << *HINT* *HINT* :)
____________
|
|
|
|
Does that (the grand unified database) mean the non-BOINC primes would find their way into the top prime finder stats as well?
____________
There are only 10 kinds of people - those who understand binary and those who don't
|
|
|
Sysadm@Nbg Volunteer moderator Volunteer tester Project scientist
 Send message
Joined: 5 Feb 08 Posts: 1188 ID: 18646 Credit: 490,016,651 RAC: 0
                    
|
http://vcn94.homelinux.org/PRPNet/primelist.html << *HINT* *HINT* :)
Karsten may use this compilation, but beware: there are not only proth primes in the list
Included in the list are all primes, which are published by PrimeGrid on the single PRPNet-ports (user_primes.html)
As mentioned by John (this post): decimal length is buggy
____________
Sysadm@Nbg
my current lucky number: 3749*2^1555697+1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/
|
|
|
HAmsty Volunteer tester
 Send message
Joined: 26 Dec 08 Posts: 132 ID: 33421 Credit: 12,510,712 RAC: 0
                
|
Does that (the grand unified database) mean the non-BOINC primes would find their way into the top prime finder stats as well?
Yeah, that would be a really cool feature. Currently i've no overview of my found primes :(
____________
|
|
|
|
Yep, I know that link already, thanks.
See link to text-file in this post.
Yes, I realized there're not only Proth primes in there, no problem.
And no, this 'Proth Database' does not mean automatically updating the PG prime finder stats.
This page(s) will only be a summary of all (sometimes) known Proth primes!
Karsten |
|
|
|
Next updates to come:
- all twins for k<1200 marked (perhaps today)
Done!
357 Twins (k*2^n-1 and k*2^n+1 prime) for k<1200 marked.
|
|
|
|
Next update:
The summary page for Proth primes lists now:
- all known primes for k<10000 (115,224 so far)
-- including all Top5000 primes (18,200 links inserted)
-- including all PG primes (upto yesterday)
--- from PPSE 10000 server
--- some from these forum threads
--- some from older PPSE 10000 and 11000 server (most were lost)
- all known twins for k<10000 (2271 so far)
Notes:
- the search-range for k>1200 is set to n=50k although not all are done (hope next week)
- same for dates (set to 2010-01-10, will changed into last prime found later)
- for colors and definitons see here
- Nash weights were calculated with psieve3.exe
- Notes (see example for k=3, red star) could be inserted for factors of F / GFN
ToDo:
- current reservations (mostly for k<1200 from ProthSearch.net)
- splitting and inserting html-pages like for Riesel-primes
Please everyone who finds an error or missing prime, post here or PM me,
the more information I get the better that page(s) become.
If anybody got some old result-files, they're welcome, too.
Thanks.
Karsten
PS: Top5000 links for 1000<k<1200 not yet there. Coming next!
|
|
|
|
I think that this prime list Karsten is creating is neat for several reasons. It will be good to have a primes list for primes that are no longer on the top 5000 list.
In the future, though, I think that it would be great if, as Honza said, the PRPNet primes made it onto the Top Prime Finders list within PrimeGrid. After all, PRPNet is a subproject of PrimeGrid.
And of course, if tripple checking can be stopped within the Proth Prime Search and John's two year old pipe dream could be fulfilled, that would be great as well. Tripple checking is conterproductive, to state the obvious.
Well anyway keep at it and good luck!
____________
May the Force be with you always.
|
|
|
|
I've just uploaded the newest testpage for Proth primes.
New:
- creating reservations / ranges done from data file
(for now: if no reservation available:
k<1200: range done at 640k
k>1200: range done at 50k)
- all Proth primes from Top5000 Database
I'm also done 1/3 of the 1200<k<10000 and n<50000 tests (~90,000 primes found).
The newest attempt is to create a page of all persons who found any Proth prime ever from Top5000. This could be useful for PG's participants.
It should be doable including non-Top5000 primes, too.
But there's nothing to show yet.
Karsten |
|
|
|
Update:
- newest primes (from 2 hours ago from PPSE10k and mine for n<50k) are online
- many reservations/ranges for k<1000 updated
- begun to include comments for primes (red star) like "Divides F(x) ... GF(x)
(so far only for k=3 and only info. from Top5000 Database
- display of ranges now with abbrev. "K" or "M"
- a bit less than 1/2 of range n<50k done
Stats:
- 5000 k-values (k=1-9999, more will come after splitting page)
- 122789 primes
- 19396 Top5000 links
- 2271 twins
Next works:
- getting more info for F/GF-factors
- more reservations
- creating user-primes-page
- updating Top5000 primes for last 2 weeks
Future:
- creating page with stats
- creating page with all reservations from prothsearch.com (like range-tables there)
- more k-values/primes for k>10000
Karsten
____________
RieselPrime Database (... and now Proth Primes, too)
|
|
|
|
It's done!
The (almost) whole data for k<10000 are online.
I've created pages like for the Riesel-side, so for k<300, 300<k<2000, 2000<k<4000,...
New Menu set up including a small Statistics-page.
See here for more datails.
Just now NPLB is rallying, so the n<50k have to wait some days.
I'm trying to update every day all new primes from PG port 10000.
Karsten
____________
RieselPrime Database (... and now Proth Primes, too)
|
|
|
|
Another update:
I've included Status pages for reservation / completion for 3 different k-ranges with corresponding n-ranges.
All data taken from W.Keller's page (prothbooks= like guestbook with reservations).
So far the timeframe from Feb.2008 to Nov.2010 is shown, more older will follow.
Holding the mouse over a contributor in the table will show 2 dates (if available):
- first date is reservation
- second date is completion
- some comments (red star) mostly for ranges
It's possible to inlcude the primes, too (as tooltip or comment) if needed.
The look and ranges may change, I'm still trying to find the best fit.
____________
RieselPrime Database (... and now Proth Primes, too)
|
|
|
|
It's done!
The range 1200<k<10000 and n<50000 I've completed:
104849 primes found and listed on my pages.
A few days and all primes for n<100000 are known with PPSE10k effort!
____________
RieselPrime Database (... and now Proth Primes, too)
|
|
|
|
Hi all,
after a long time now, trying to display Proth primes on my pages, too, it's still hard to find appropriate information here:
- Status of ranges tested by PrimeGrid are only available on the "Subproject status" link, but
- for a running range only lowest and highest testing n-value
- for finished ranges only highest n tested and number of primes found
- found primes only "listed" at Top5000 (if they are in) or
for example for PPSE from the user-status tables
Again my request:
It would be very nice and helpful to find a page like this on the Proth Search:
- giving all primes found and ranges tested by PrimeGrid.
There're many pages for user stats and badges and challenges but not such simple display of the results found here!
Please, make such overview.
Thanks
Karsten
____________
RieselPrime Database (... and now Proth Primes, too) |
|
|
|
Yes, I would also appreciate a page like that. It would be a nice contribution to human knowledge. |
|
|
Honza Volunteer moderator Volunteer tester Project scientist Send message
Joined: 15 Aug 05 Posts: 1893 ID: 352 Credit: 3,142,312,174 RAC: 0
                             
|
I know it is not what you asked for but a little help
Recent progress can be found in PPS (LLR) Milestones, Goals, and Progress thread.
Good part about it is that doesn't cover only what was done but also where we are heading.
____________
My stats
Badge score: 1*1 + 5*1 + 8*3 + 9*11 + 10*1 + 11*1 + 12*3 = 186 |
|
|
John Honorary cruncher
 Send message
Joined: 21 Feb 06 Posts: 2875 ID: 2449 Credit: 2,681,934 RAC: 0
                 
|
after a long time now, trying to display Proth primes on my pages, too, it's still hard to find appropriate information here:
3 emails over the past 3 months have all gone unanswered by you. It's difficult to provide information when there's no response.
The page you want has been a request for a few years now. In the meantime, here's some piecemeal data:
2012 Goals - Progress update gives you the min/max for ranges. For k<1200, trailing edge is n min. For 1200<k<10000, leading edge is n max.
n min for 1200<k<10000 is n=310k. 310k<n<450k is being working on in two PRPNet ports: PPSElow and PPSElow2. Primes found in those ports can be found here: PPSElow primes and PPSElow2 primes.
These two ports are working on filling the following gaps:
310k<n<360k
363k<n<380k
384k<n<450k
Everything else n<725k for k<10000 has been completed.
____________
|
|
|
|
Thanks for the links again, I know them all.
But:
a) 2 servers for PPSElow -> 2 pages with (unsorted) primes, only listed by date
b) no page for found primes of n greater PPSElow-range -> get those only from Top5000
c) goals for 2012: one post (changing) with current information
Problems:
a) I have to download 2 pages and input them into a textporcessor to get those data I need
b) I have to download every new prime and input them into a textporcessor to get those data I need
c) I can't even use a textprocessing tool to get those information.
In short:
It's too much manually work to get all information to update my pages.
On the other hand I think it's a must have for a big project like PrimeGrid to show their results in a userfriendly manner but here is nothing since beginning... it's not the exercise of somebody like me.
So please, find somebody to create such pages, and for the other subprojects, too.
Thanks
____________
RieselPrime Database (... and now Proth Primes, too) |
|
|
John Honorary cruncher
 Send message
Joined: 21 Feb 06 Posts: 2875 ID: 2449 Credit: 2,681,934 RAC: 0
                 
|
On the other hand I think it's a must have for a big project like PrimeGrid to show their results in a userfriendly manner but here is nothing since beginning... it's not the exercise of somebody like me.
So please, find somebody to create such pages, and for the other subprojects, too.
I'll see if I can find a volunteer willing to help.
____________
|
|
|