mersenneforum.org Prime-Wiki
 User Name Remember Me? Password
 Register FAQ Search Today's Posts Mark Forums Read

2020-05-15, 03:58   #155
Happy5214

"Alexander"
Nov 2008
The Alamo City

33410 Posts

Quote:
 Originally Posted by kar_bon No, it was Riesel k=25 (see here). Borys found his first prime of this k-value in 2006 and I don't think he used the same computer since then for his calculations. It's more about a passion in prime searching, because the second last prime was from 2008. This k is not that low weighted but the gap is quite big here.
Thanks for the correction.

LLR has gotten much faster from 2006 (when it only had SSE2 code) till now (with AVX2 and AVX-512 versions). For example, my laptop is 50% faster at LLR then my desktop, despite having only 60% of the desktop's clock frequency, because the laptop has the FMA3 FFT version and my desktop only has the SSE2 FFT code.

Quote:
 Originally Posted by storm5510 About the database: I would much rather pick up a k which already exists and needs more attention rather than start a new one. 10673 was tested to n = 25,000. I reserved it.
If you've ever looked at my reservations, you'll find that almost all of my work involves filling in gaps in the old RPPDb. My primary foci right now are the RPS 9th and 10 Drive k's and Woodall and near-Woodall k's. I run them through PRPNet servers, though I only have one computer (the aforementioned laptop) currently working on them.

 2020-05-15, 07:54 #156 kar_bon     Mar 2006 Germany 23·353 Posts The 'problem' with RPS Drive 9 &10 is they started searching for n>400k, so the given max n-value is not correct in the Wiki for all those k-values yet. In the bgeinning I decided not to include a parameter for the lowest n-value, because most of them are n=1 and not necessary at all. Otherwise there's always the remark paramter to fill in such missing range. I think of two additions later: - The Riesel template (and Proth also) can determine the highest found prime value. If this is higher than the given nMax value, put that k-value into an own category. Example: k=69 : highest prime n=6838971, nMax=2416000 - Create the category and fill in this category entry for such k-values with smaller missing n-ranges by hand: Example k=17369 from RPS Drive 10: entry on the "Riesel prime 17369" page added as [[Category:Riesel prime missing range]] The latter is done in the Wiki right now. Many higher k-values in the RPPDb got missing n-ranges, because of twin search, SG search or even constant n-search. The advantage of such category is to catch those help needed k-values easily. When I'm ready with including all data for Riesel k<1002 I'm working on the drives of RPS and NPLB transferring data from my old pages.
2020-05-15, 14:10   #157
storm5510
Random Account

Aug 2009
U.S.A.

2·7·89 Posts

Quote:
 Originally Posted by Happy5214 If you've ever looked at my reservations, you'll find that almost all of my work involves filling in gaps in the old RPPDb. My primary foci right now are the RPS 9th and 10 Drive k's and Woodall and near-Woodall k's. I run them through PRPNet servers, though I only have one computer (the aforementioned laptop) currently working on them.
I have only recently began looking at some of the assignments which individuals have. I found one which has 12. I thought I was being excessive with two, and on rare occasions, three.

Quote:
 Originally Posted by kar_bon Example: k=69 : highest prime n=6838971, nMax=2416000
This particular k has far larger primes than the highest n value. Values of n must not have been important back then.

 2020-05-24, 21:06 #158 kar_bon     Mar 2006 Germany 23×353 Posts I'm currently working on trying to speed up the generating of the prime tables for Riesel primes. So I decided to extend a new parameter "RCount" to the Riesel prime template. If RCount is not given nothing changes in the handling. If RCount is given, there're two options: - RCount is the same as the automated counting -> no problem, everything ok - RCount differs from counting -> Alert shown like the attachement So please do always: - Creating new k-values now have to contain this new parameter. - Before saving any k-value use the "Show preview" button, which will show the alert before saving and the count, so you don't need to count by your own. Fill in the correct value and save. In the end when all k-values contain the RCount the template to create the tables will only use this parameter which should speed up the creation time. If this helps, other templates like Proth will changed, too. Attached Thumbnails
2020-05-25, 17:26   #159
storm5510
Random Account

Aug 2009
U.S.A.

2·7·89 Posts

Quote:
 Originally Posted by kar_bon ...Creating new k-values now have to contain this new parameter. - Before saving any k-value use the "Show preview" button, which will show the alert before saving and the count, so you don't need to count by your own. Fill in the correct value and save. In the end when all k-values contain the RCount the template to create the tables will only use this parameter which should speed up the creation time. If this helps, other templates like Proth will changed, too.
RCount is a verification of the number of items in RNList. Is this correct? I take it that the auto-count will go away? When creating a new k, RCount would be set to zero if RNList contains no entries.

The addition of RSieve will be helpful as it will indicate a k is not in process. I added this to one of mine which needs more sieving and is currently inactive. 98475.

 2020-05-25, 20:07 #160 kar_bon     Mar 2006 Germany 54108 Posts RCount will become the value which will be taken when a k-value is shown in any overview table. The automated counting is only done in the template of the k-value page to be sure taking the correct value (this will stay). RCount will not set automatically (perhaps possible later). So creating a new k-value an entry without "RCount=" does not matter now, but this is noticed. The RSieve value is meant to show there's a sieve available for this k, which can be uploaded by everyone (the filename is generated). You can show the max n-value, sieve depth and number of candiates in the RRemarks parameter (as example see here).
 2020-05-25, 21:25 #161 kar_bon     Mar 2006 Germany 23·353 Posts Again the sieve file: I've changed the handling, when creating a new k-value - new k-value for Riesel prime created (not saved yet, only "Show preview") with RSieve set (like "RSieve=y") - there is a red link "Upload me!" - right klick the link -> open in new tab - a new page pops up to upload the file with destination filename given - choose your local file - upload file - back to the k-value page - Show preview again - the link should become blue (available page) and is "(sieve)" now - save k-value page if anything is correct I've also included the parameter list with text into the Riesel template and a new category, when there're 100 n-value or more for this k-value (see here).
2020-05-25, 23:55   #162
storm5510
Random Account

Aug 2009
U.S.A.

2×7×89 Posts

Quote:
 Originally Posted by kar_bon ...The RSieve value is meant to show there's a sieve available for this k, which can be uploaded by everyone (the filename is generated). You can show the max n-value, sieve depth and number of candiates in the RRemarks parameter (as example see here).
Understood, and thank you for the reply. I removed the reference I made earlier. I always use "Show Preview" to make sure everything is proper before doing the actual save.

In your example link, there is a unit of measure, "21T." (K)ilo, (M)ega, (G)iga, (T)era, and (P)eta would be a guess. In this case, 21e12. To be honest, there are times when I would much rather sieve than test. All a person would need to do is pick a k and look at its ending n value and go from there.

 2020-05-27, 04:19 #163 sweety439     Nov 2016 3·5·137 Posts Williams prime PM table does not show the primes for bases 64, 128, 256, and 512 Done. Last fiddled with by kar_bon on 2020-05-27 at 08:24
 2020-05-29, 10:34 #164 kar_bon     Mar 2006 Germany 23·353 Posts I've included templates for Proth prime base b (example here, smallest open SR5, Sierpinski side) and Riesel prime base b (example here, found SR5 prime, Riesel side). So every combination of k/base for Proth and Riesel are possible, helpful for PrimeGrid Sierpinski Riesel Base 5 Search or CRUS. Every base value creates it's own category for those k/base combinations. Category-Tree: - Riesel prime -- Riesel prime base 5 (here a general page for such base should be located, like a table of all k-values) ---- Riesel prime 5 1396 (= Riesel primes of the form 1396*5^n-1) As a preview for including General Fermat numbers (so Top5000 comments like "Divides GF(a,b)" or "Divide xGF(a,b,c)" are possible with links in the Wiki, see here only as texts) a page for Gen.Fermat k^262144+1 (=b^2^18+1) all known primes are listed, with data as a CSV file, so expandable without editing the page itself. Smaller exponents consists of way more values, so the page for example for b^2^16+1=b^65536+1 is the same size and all data saved as file. Last fiddled with by kar_bon on 2020-06-09 at 13:19 Reason: link corr.
 2020-06-19, 10:09 #165 kar_bon     Mar 2006 Germany 23×353 Posts Some new changes: - There's now a small menu on top of every page instead of the info block. Links to helpful pages are now available for new users or even for a quick overview. Especially the "Template prototypes" should be used by every user when creating new prime pages: all current prototype are available for the implemented prime types. Using those won't miss any parameter or using false ones. Other links are still available in the left side menu. - There're two new categories for Riesel primes in the Wiki: 1) k-values without any Sophie Germain prime 2) k-values without any twin prime These only holds for k-values $k \equiv 0\ mod\ 3$ The conditions will be generated by the Riesel template, so could be take some time, if all k-values are updated by the Wiki.

 Thread Tools

 Similar Threads Thread Thread Starter Forum Replies Last Post Uncwilly mersennewiki 16 2018-10-23 23:16 kriesel mersennewiki 0 2018-08-13 13:02 Xyzzy mersennewiki 3 2011-02-18 03:31 ixfd64 mersennewiki 6 2006-05-22 12:05 delta_t PSearch 2 2006-05-21 07:05

All times are UTC. The time now is 14:13.

Fri Jul 10 14:13:29 UTC 2020 up 107 days, 11:46, 1 user, load averages: 2.53, 2.42, 2.22

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.