-   Conjectures 'R Us (
-   -   News (

gd_barnes 2007-12-24 04:51

This thread is to report any general news about the Conjectures 'R Us effort. Items of interest include team drives, monster prime finds, conjecture proofs (i.e. finding a prime for the final k of a base), and anything else that would be interesting for the team to hear about.

The first news now is that I would like to start team drive #1 to find primes on all 56 remaining Sierp base 16 k's. I am currently in the process of sieving them all to n=200K. Everyone is still welcome to reserve individual k's and I'll be glad to send you the applicable portion of the big sieved file as a starting point for the individual seach.

We'll still leave some k's out of the team drive for individual searching in addition to those that are reserved when we are ready to start searching.

If anyone is interested in helping with some sieving, let me know. I will be up to P=200G by Monday morning (CST US) running on 2 cores. I'm now estimating an approximate optimal sieve depth of around P=4T.


gd_barnes 2007-12-26 18:06

Poll complete; split results; we'll do both!
The poll on the direction of the effort for Conjectures 'R Us is now closed. It was an even split at 3-3 between filling in the bases <=32 and focusing on bases that are powers of 2 with no votes for extending the bases up to 50 or 100. I did not vote. Although I would lean slightly towards filling in the bases <=32, since there are many differing views here, we can do BOTH!

Powers of 2:
1. The sieving is continuing on a team drive for base 16 Sierp.

2. Powers of 2 have been extended up to 256 on the pages. Bases 32, 64, and 128 were easily proven or shown to be virtually impossible to prove with current knowledge and technology with one exception: 40*128^n+1. It has been effectively searched to n=642.8K (converted from 5*2^n+1 shown on the Proth pages as completely searched to n=4.5M). So if you want to commit to a very long search, that one is available.

3. I will start base 256 on both sides later this week. The Sierp conjecture is low and so should probably be quick and not worthy of a team effort. I think the Riesel side will yield many k's to search and should be an interesting one for those interested in powers of 2.

Fill in for b<=32:
1. Riesel base 6 has now been completed to n=25K. 19 k's are remaining and ready to be worked on. This is the lowest remaining available base that is not currently being worked on. Jeanne Penne is currently work on all remaining base 4 k's with the exception of 64494*4^n+1.

2. Michaf continues to work on Sierp base 24 and has now passed n=10K on all remaining 200+ k's. A lot of work will come out of that one.

3. I will start Sierp base 6 sometime next week after base 256. With a high conjecture of k=174308, plenty of work should come from that one.


gd_barnes 2007-12-26 23:45

Major site update
I have updated the web pages to separate the Riesel and Sierpinski conjecture reservation pages and include a column for sieved files. Currently only Riesel base 13 and Sierp base 17 are uploaded. Later on, I'll upload k=64494 for Sierp base 4 as well as all of the Sierp base 16 k's after the sieve has completed to P=400G later tonight.

There were a lot of changes for the links to various pages so if you see any issues with them, just post them here.


gd_barnes 2008-01-04 20:01

Team drive #1 has started
Team drive #1 for Sierpinksi base 16 has started. Reserve those ranges while they're HOT! :smile:


gd_barnes 2008-01-08 09:15

Team drive #2 has started
Team drive #2 for Riesel base 16 has started. Come and get 'em! :smile:


kar_bon 2008-01-10 13:40

i'm going to collect all primes of Riesel power of 2-bases on the data pages of [url][/url].
the first one is k=17135: i tested this k upto n=100k (17 primes) and marked n=78064 with a comment "Riesel base 16 prime: 17135^16^19516-1" (mouse cursor over red star).
see here: [url][/url]
more to come and an extra pages with all those primes too.

gd_barnes 2008-01-10 18:48

[quote=kar_bon;122566]i'm going to collect all primes of Riesel power of 2-bases on the data pages of [URL][/URL].
the first one is k=17135: i tested this k upto n=100k (17 primes) and marked n=78064 with a comment "Riesel base 16 prime: 17135^16^19516-1" (mouse cursor over red star).
see here: [URL][/URL]
more to come and an extra pages with all those primes too.

Very good!

I am in the process of putting a list together of k's for all bases and their associated lowest first primes for historical proof or near-proof of the conjectures.

When I'm complete, I'll post spreadsheets on my web pages of all of them with links to them in this forum. For Riesel bases that are a power of 2, that will give you plenty of extra info. for the Rieselprime site.

BTW, did you buy the domain name? It's a great name! I think we should make that the 'official online' Riesel primes site instead of A new person coming to prmie searching would know exactly what it is. Also, it allows you to update it instantly and you can have the much improved k < 300 page on it.


gd_barnes 2008-01-14 05:44

Changes being made for multiples of base and GFN's
News to report...

You may have noticed that I haven't updated the site in a couple of days. It's because I've been working to do the following:

1. Search and add k's that are multiples of the base (MOB) to the pages.

2. Remove k's that make Generalized Fermat #'s (GFn's).

3. Change the definitions of what is excluded at the top of the pages as well as in the project description.

4. Create a list of additional k's to search.

This has been quite a large undertaking because not only do I need to find additional k's that are remaining, I have to see if which one make a different prime than k / b and if so, see if the new prime makes the top-10 list for largest primes for each base. Part of #1 is done. All of the searches are done and I've now updated all of the Riesel pages for MOB (but not loaded them yet). I just now started working on the Sierpinski pages. They are more involved because unlike Riesel's, they have GFn's that need to be removed (with an appropriate explanation) that make a difference in whether the conjectures are proven or not.

I hope to be done in the next few hours but if not, I'll post only the updates for the Riesel side and save the Sierp side for Monday evening, which includes all of the primes and reservation changes in the last 2 days.


gd_barnes 2008-01-14 12:20

Multiples of base included / GFn's excluded
The project definition and web pages have now been udpated to include multiples of the base (MOB) that make a different prime than k divided by the base. Additionally after a small discussion with Jean Penne, it was decided that all k's that make generalized Fermat #'s (GFn's) would be excluded because n must be q^2 in order to make a prime leaving few candidates to test. On the right side of the main Sierp conjectures page, it shows the GFn's that have no known prime.

I haven't quite searched all the bases for MOB but only 3-4 are left. With this change, the impact has been surprisingly minimal with only 4 additional k's that need primes for all bases combined so far. All conjectures that were already proven stayed that way. One conjecture, Sierp base 32 went from not proven to proven upon the exclusion of GFn k=4.

Since some of you have reserved all of a particular base, here is a list of additional k's that are MOB that we now need primes for:

19464*4^n-1 (already reserved by Jean Penne)
594*27^n-1 (other 2 k's on Riesel base 27 reserved by Simelink)

90546*6^n+1 (Sierp base 6 reserved by geoff)
17496*24^n+1 (Sierp base 24 reserved by micha)


gd_barnes 2008-01-15 05:47

We are now officially recognized as a "Prime Search Project" here at mersenneforum! :george:Woohoo! Thanks Mike!

In other news, I have requested info. from Prof. Caldwell at the top-5000 site for creating an official project code for reporting primes there.

I'll let you know when we have one.


gd_barnes 2008-01-16 08:07

Project ID and first top-5000 prime!
We now have a project ID at the top-5000 site. It is CRUS. For big primes found here, please create a new prover code that includes this project.

In related news, we just had reported our first top-5000 prime. :w00t:

Found by tnerual; it is 34543*16^91225+1.

Way to go!


All times are UTC. The time now is 03:08.

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