mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > Conjectures 'R Us

Reply
 
Thread Tools
Old 2007-12-14, 22:37   #1
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

27C016 Posts
Default Bases 4-32 reservations/statuses/primes

Please use this thread for posting reservations and statuses on bases 4-32 for "Conjectures 'R Us". When considering what to reserve, please take into account the size of the base. Base 30 of course tests much slower than base 9 at the same n-value. Also for bases that are powers of 2 (bases 4 & 16 for this effort), LLR will convert them to base 2 before searching, which means faster searching at effectively the same size of test. Although it will write out the 'primes' file in base 4 or 16, the results file will be in base 2.

For statuses, these are guidelines only. Nothing is set in stone. Please report a status at least once/month and preferably more frequently. The web pages will be updated most every day to include new info. and sometimes several times a day. Use your own judgment on this. If you're searching at n=5K on multiple k's and are knocking down primes left and right, then report in a little more often. Clearly if you're searching above n=100K, then once/month to let us know where your search is at is fine.

I am currently working on Sierp base 16 for all k's. I am up to n=19K and will take it up to n=25K, which should be done Sunday. I'll then post the 60+ k's remaining for everyone else to search. I'll wait to see what the response is before reserving any more k's myself. I have eliminated all k's where higher primes were found for bases 2 and 4 and could be converted to an equivalent k base 16.

Base 16 is a good base because it LLRs like base 2 for the equivalent size, i.e. k*16^n+1 LLRs at the same speed as k*2^(4n)+1. This is one where I think we'll get a lot of top-5000 primes! It only takes an exponent of around n=83.4K to make the top-5000 list and with 60+ k's left at n=25K, there will be plenty to go around.

More than one person can reserve the same k for a different n-range. k's already reserved can be reserved by a 2nd person for the lowest n-range above that which has already been reserved. I enourage this kind of collaborative effort to facilitate reserving many k's for a narrow n-range or knocking out very stubborn k's. If this happens, the k will be shown on the reservation pages with both people and their respective n-range reservations. Clearly if one person finds a prime, then the other can choose to stop searching at his own discretion. Both people should check the primes and reservations threads a little more frequently than they normally would.


Thanks,
Gary

Last fiddled with by gd_barnes on 2010-01-18 at 09:05 Reason: clarification in 1st para
gd_barnes is offline   Reply With Quote
Old 2007-12-15, 21:31   #2
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

26×3×53 Posts
Default

Quote:
Originally Posted by michaf View Post
I was under the impression that Xyzzy was only a device used by the gerbils. Actually one gerbil and a spirit of one I believe.
Ha ha. Good one Michaf! I just responded to your PM about base 24 Sierp.

For everyone's info., Michaf has started from scratch running PFGW on all k's for base 24 Sierp. Michaf, can you post something in the reservation thread about reserving base 24 Sierp to n=5K? Thanks! I just want to keep everyone in the loop of what is going on. I'll show it as reserved by you on the web pages later tonight.

Anon, thanks for info.! I never knew that before. I'll respond to your PM later this evening. My daughter has a championship soccer game to go to shortly. I'm pretty excited about it!

Citrix, I got a nice PM from Xyzzy this morning that discusses some important things. I'll coordinate through the both of you if I feel we really need to move or edit anything.

Later tonight...a new sieving thread. I'm going to post my numerous sieved files there and will ask that others post some of theirs if they only want to (or can only sieve) or if they find they've bitten off too much to test.


Gary
gd_barnes is offline   Reply With Quote
Old 2007-12-15, 22:42   #3
michaf
 
michaf's Avatar
 
Jan 2005

479 Posts
Default

I'm running Sierpinski base 24 upto n=5000 now.
I'll report here when it's done
michaf is offline   Reply With Quote
Old 2007-12-16, 06:42   #4
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

26·3·53 Posts
Default

When done doing all searches, please send results files to me at gbarnes017 at gmail dot com.

Like some of the larger projects such as Riesel Sieve and Seventeen-or-Bust, we will probably need to do some double-checks on stubborn k's where a prime cannot be found up to a very large n. Both projects recently found huge primes on double-checks. Having the residues in the results files will help us identify if there are any patterns of problems in any searches.

In LLR, it's the lresults.txt file. In PFGW, it's the pfgw.out file.


Thanks!
Gary

Last fiddled with by gd_barnes on 2010-01-18 at 07:19
gd_barnes is offline   Reply With Quote
Old 2007-12-16, 07:02   #5
Citrix
 
Citrix's Avatar
 
Jun 2003

2×33×29 Posts
Default

Gary, Could we have a separate webpage for powers of 2. I think these will be the most popular.
By the way the webpage looks really nice.
Citrix is offline   Reply With Quote
Old 2007-12-16, 08:47   #6
Siemelink
 
Siemelink's Avatar
 
Jan 2006
Hungary

4148 Posts
Default

Quote:
Originally Posted by gd_barnes View Post
When done doing all searches, please send results files to me at gbarnes017 at gmail dot com.

Like some of the larger projects such as Riesel Sieve and Seventeen-or-Bust, we will probably need to do some double-checks on stubborn k's where a prime cannot be found up to a very large n. Both projects recently found huge primes on double-checks. Having the residues in the results files will help us identify if there are any patterns of problems in any searches.

In LLR, it's the lresults.txt file. In PFGW, it's the pfgw.out file.

Thanks!
Gary
Hi Gary,

All your coordination effort is really nice. I checked out the webpages, they are solid stuff. I just had one minor thought, is it possible to keep the headers of the table in view when scrolling down?

As for your request for the lresults file, well, I don't know. While technically correct this represents a lot more administration effort.
I run this stuff manually on a handfull of computers. I check the lresults file daily to check for progress and then empty the lresults file. So I don't have the lresults anymore.

Laters, Willem

Last fiddled with by gd_barnes on 2010-01-18 at 07:19
Siemelink is offline   Reply With Quote
Old 2007-12-16, 08:55   #7
Siemelink
 
Siemelink's Avatar
 
Jan 2006
Hungary

22×67 Posts
Default

To get a feeling on how to use srfile and sr1sieve I checked 233*28^n-1 until 25000. No prime found.
Siemelink is offline   Reply With Quote
Old 2007-12-16, 08:59   #8
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

26×3×53 Posts
Default Web page reservations page changes needed

Quote:
Originally Posted by Citrix View Post
Gary, Could we have a separate webpage for powers of 2. I think these will be the most popular.
By the way the webpage looks really nice.
Thanks Citrix. I agree those will probably be the most popular. What does everyone else think? Should we have separate pages for the bases that are powers of 2? I admit I kind of like it 'sorted by base with no holes' but I can understand the reasoning behind having them separate.

Before responding to the above though, I have to bring up that there is a much bigger issue here. I think you may get part of your wish but for a different reason on base 16 for the reservations page. I'm going to need entirely new reservations pages for bases with many k's remaining. I think I've decided how to handle it but am open to hearing more on the matter. The issue is that I need to allow for 100's or even 1000's of k's remaining on a base. This is almost definitely going to be the case on bases 6 or 19 when we get into those due to the high conjectures. Base 19 Riesel has a conjecture > 1.1M! I don't want to even think about bases 3, 7, and 15!

Here is what I'm currently considering: Remove the k's remaining on the 2 main info. pages, show the # of k's remaining instead and provide a link to a NEW reservations page set up for the specific base that has reservations info. for all of the remaining k's. That way, I can leave all of the other main info. about the bases on the main pages and wouldn't have the duplicate work of keeping them synced up. This is the same thing that I do for the bases with major projects right now. I think it makes the most sense and keeps things fairly consistent and easy to follow.

I'll then create the above mentioned separate reservations pages for bases with large #'s of k's remaining like there will be on base 16. I'll leave the bases on the current main info. pages and reservations page with all other info. but for the k's remaining and reservations, I'll put a link to the page with the reservations info. like I do on the main page.

If I don't do this, the current reservations page will quickly become a mess in the near future. Right now for base 16 Sierp, it's looking like there will be around 53-57 k's remaining at n=25K (plus 3 being searched by the SOB and Sierp Base 4 projects). Imagine what the current reservations page would look like after that and after Michaf has finished base 24 Sierp to n=5K? It wouldn't be a pretty sight nor SITE! The page is already somewhat cluttered and we don't want to make it downright UGLY and hard to follow!

Don't think I'm ignoring your suggestion about powers of 2 being on a separate page. After things settle down from the more pressing issue of separate reservations pages for some bases, I would still consider it depending on what people think.

Late added note for clarity: When I say 'main info. pages', I'm referring to the 2 pages that are NOT the one reservation page. It's easy to get lost in semantics here.


Gary
gd_barnes is offline   Reply With Quote
Old 2007-12-16, 09:47   #9
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

1017610 Posts
Default

Quote:
Originally Posted by Siemelink View Post
Hi Gary,

All your coordination effort is really nice. I checked out the webpages, they are solid stuff. I just had one minor thought, is it possible to keep the headers of the table in view when scrolling down?

As for your request for the lresults file, well, I don't know. While technically correct this represents a lot more administration effort.
I run this stuff manually on a handfull of computers. I check the lresults file daily to check for progress and then empty the lresults file. So I don't have the lresults anymore.

Laters, Willem
Thanks, Willem and for coordinating through us here.

Well, I'll let you off THIS time since I just posted the request a little while ago!

Seriously, it would be best if you could copy them off and send them but I'm not going to press the issue if it's a problem.

So what you're saying is that you perhaps run a range of n=5K-25K like you did on 233*28^n-1 on several computers by splitting it up perhaps like 5-10K on one machine, 10-15K on another, etc.? Is that correct? If so, I don't mind getting multiple files for one search, regardless of how they are sorted. Many people do searches in that manner on several machines. But even if they're a jumbled mess, when we do some double checks here-and-there, we can sort them if necessary to find the appropriate residues needed.

If this presents a large issue for some people, I'll drop the request. It's just something that will be helpful in the future for the 'greater good'.

One thing I do have to have is primes files and results files on is when searching brand NEW bases from n=1. There's no choice there. We have to know what primes are found for every k to show that we've proven the conjecture.

No biggie. If you can do it, that's great and is helpful. If not, that's fine.

On the web pages headers. I also had an initial suggestion for them before I made the pages public. I'll see what I can do sometime this week. They're designed off of the 15k.org pages where the headers also do not stay stationary. It would be kind of handy for viewing.


Quote:
Originally Posted by Siemelink View Post
To get a feeling on how to use srfile and sr1sieve I checked 233*28^n-1 until 25000. No prime found.

Thanks for the info. and for helping out the effort. I'll mark it as searched to n=25K. Some questions:
1. Did you use LLR to search it or do you use some other software? That was a pretty fast sieve and search and it's got me wondering.
2. Do you want to keep this k reserved? If so, do you know how high you're going to search it?
3. Are there any others that you've started working on or that you wish to reserve at this time?

The reservations help us avoid double-work.

You already reported the large efforts that you did on base 22 in that thread. I'm still keeping those reserved for you because I'm assuming that you're continuing the search on them.

One final thought...I am going to request that the bases 22 and 23 threads be locked and coordinated through this effort. Do you see that as a problem? When I do that, I'll put one final post in them stating that they will be locked and requesting any additional info. that people have be posted here.


Thanks,
Gary

Last fiddled with by gd_barnes on 2007-12-16 at 09:52
gd_barnes is offline   Reply With Quote
Old 2007-12-16, 14:23   #10
Siemelink
 
Siemelink's Avatar
 
Jan 2006
Hungary

1000011002 Posts
Default

Hi Gary,

Here are some answers to your questions. You sure fire them off!
I am happy to keep my efforts public, your page makes the various efforts very clear.

I am not interested in keeping the lresults files. That represents effort that I'd rather not spend. Given that there are already have a few hundred k & b pairs on the page the amount of data would quickly balloon.

The run with b=28 took me five or six hours. Started with NewPgen, ran LLR until 20k, then some more sieving with sr1sieve and LLR to mop it up.

I will take all the base 22 candidates until 100k. I think I'll take two of them to 200k. I'll post an update this week sometime.

Laters, Willem.
Siemelink is offline   Reply With Quote
Old 2007-12-16, 21:18   #11
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

100111110000002 Posts
Default

Quote:
Originally Posted by Siemelink View Post
Hi Gary,

Here are some answers to your questions. You sure fire them off!
I am happy to keep my efforts public, your page makes the various efforts very clear.

I am not interested in keeping the lresults files. That represents effort that I'd rather not spend. Given that there are already have a few hundred k & b pairs on the page the amount of data would quickly balloon.

The run with b=28 took me five or six hours. Started with NewPgen, ran LLR until 20k, then some more sieving with sr1sieve and LLR to mop it up.

I will take all the base 22 candidates until 100k. I think I'll take two of them to 200k. I'll post an update this week sometime.

Laters, Willem.
Thanks for answering my NUMEROUS questions and keeping us updated Willem. No problem on the results files.


See ya,
Gary
gd_barnes is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Bases 501-1030 reservations/statuses/primes KEP Conjectures 'R Us 3701 2020-08-15 09:36
Bases 33-100 reservations/statuses/primes Siemelink Conjectures 'R Us 1667 2020-08-14 16:04
Bases 251-500 reservations/statuses/primes gd_barnes Conjectures 'R Us 2126 2020-08-05 17:01
Bases 101-250 reservations/statuses/primes gd_barnes Conjectures 'R Us 826 2020-08-05 17:00
Riesel base 3 reservations/statuses/primes KEP Conjectures 'R Us 1040 2020-07-13 07:31

All times are UTC. The time now is 16:29.

Sat Aug 15 16:29:00 UTC 2020 up 2 days, 13:04, 0 users, load averages: 2.08, 1.93, 1.85

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.