mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Factoring

Reply
 
Thread Tools
Old 2010-01-01, 06:34   #1
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

3·419 Posts
Default FactorDB question

In the factorDB, who managed to make up ALL the entries of Cunningham numbers (upto 200000 digits), Aliquot sequences (till 1000000), Home prime sequence, and probably even (Fibonacci, Lucas numbers, Homogeneous cunninghams, factorials & primorials ± 1, smarandache numbers, quasi - repdigits, euler & bernoulli numbers, riesel & sierpinski numbers, etc... I haven't checked them up)
Extremely GREAT job by the person whoever has done so.

Remember that in my opinion, programming a factor database is VERY EASY, but making up all the entries from different projects from different corners all over the world is clearly IMPOSSIBLE. Still wondering about it, actually... :surprised

FactorDB gives out the complete list of known factors of numbers as high as, say 29901-1. But, I entered up a random number as well, such as 3615-1. For this number, it gives a 68 digit prime factor, along with a c147. How was that p68 found out? Certainly not by using ECM at all. Only one p68 is found out ever by using ECM, that is of 64×10341-1. Certainly that it should be factored off by using the Number Field Sieve algorithm only. Thus, the person whoever has found out that p68, why didn't he factor up that c147 at all?

How popular factorDB is? Do people from all over the world put up the latest factors that they have found out from different projects that are going on? Thus, keeping up the factorDB up-to-date. How is it made popular? I assume that all the factoring people, all over the world is not within the Mersenneforum at all. And then how is the Mersenneforum made popular since August 2002? How is GIMPS made popular to people all around the world? How does it reach up all the interested people all over the world to be made aware of this Mersenneforum only or the GIMPS project itself?

Last fiddled with by Raman on 2010-01-01 at 06:42
Raman is offline   Reply With Quote
Old 2010-01-01, 09:36   #2
henryzz
Just call me Henry
 
henryzz's Avatar
 
"David"
Sep 2007
Cambridge (GMT/BST)

22×1,433 Posts
Default

the p68 factor of 3^615-1 was known because it is a factor of 3^205-1 which is an algebraic factor of 3^615-1 although for some reason it is not listed by the database as an algebraic factor
henryzz is offline   Reply With Quote
Old 2010-01-01, 10:27   #3
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

2·3·52·19 Posts
Default

Quote:
Originally Posted by Raman View Post
How popular factorDB is? Do people from all over the world put up the latest factors that they have found out from different projects that are going on? Thus, keeping up the factorDB up-to-date. How is it made popular? I assume that all the factoring people, all over the world is not within the Mersenneforum at all. And then how is the Mersenneforum made popular since August 2002? How is GIMPS made popular to people all around the world? How does it reach up all the interested people all over the world to be made aware of this Mersenneforum only or the GIMPS project itself?
As i can say the FactorDB from M.Tervooren is popular to most/all of the the people at this forum doing some factoring.

I'm sure not all people all over the world know this DB nor this Forum. There're some pages with many factors not yet in the DB.
The most work on collecting factors is done here by the Aliquot Sequences thread. The Cunningham project is not 100% in the DB yet but new factors found from searchers here will be inserted (i hope).

When i begun with prime hunting, i was searching the internet, trying to find the related sites of my interests. I found this forum quite quick and those projects here. Although the information i needed were spreaded all over the net so i decided to make my own RieselPrimeDatabase with the information i could found, hoping it's useful for others.
The same with the Aliquot Sequences so i included them too.

And i think for factoring projects it's the same: there're many sites with special factorings but the FactorDB is the first one collecting them all at one place!
kar_bon is offline   Reply With Quote
Old 2010-01-02, 18:34   #4
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

3×419 Posts
Default

Just one more question: Are factordb.com, rieselprime.de, (aliquot.de, mersenneforum.org, mersennewiki.org, mersenne.org, never ending...)
hosted for free or for money? How much cost, if any? Where do you get those domain names from? (I am not actually going to host any site, just want to know about it.)

Off topic:
In February to April 2008, for 7,295- I used to log in into every Windows computer, and then run GGNFS, before night.

Since September 2009, by using Linux, several progress has been made.
1. Nohup command to keep running binaries even after logging off
2. Ssh to sit in one computer and then to connect remotely
3. Nice and Renice to monitor priorities of the processes
4. Cron to schedule task at specific time (Testing has to be done to run cron from specific folder and then to restart the GGNFS lattice siever at the left over special-q value from the command line itself, rather than adjusting the job file everytime)

Some more dreams have to come true: Which are all possible in reality? How to do so?
5. Automation of tasks, making use of perl script or linux shell script to check up if the GGNFS lattice siever is already running behind in the background, every hour or so, if not, resume the lattice siever only at the left over point, automatically, by itself
6. Use of GGNFS server client model, automatically assigning a special-q range of one lakh to each system to sieve, after it has completed a task. This makes sure that there exist no gaps at all between the overall sieving ranges, than assigning a specific range to sieve for each system. Even that the GGNFS client model, may not be used at all, rather automatically program a server client model by using the perl script or the linux shell script to adjust the special-q values within the job file automatically, and then assign a work to requesting client automatically, by using a single command line itself. I am sure that the NFSNET, NFS@Home projects make use of this only, but I have to learn about all that, how to do so, for myself.
7. The Ultimate challenge! would be to access the running computers at my college, from my home computer itself. That are outside of LAN, via Internet, by using some remote login service such as telnet, ssh, ftp, etc. Is it really possible to do this so? I have my login id (username, password) at my college, from my home computer itself, I would like to monitor the status of progress of each system within my college only.

@ 10metreh: I was writing this post right now, so that I was not able to modify the content on the Mersenne Wiki for the Cunningham Table entry on 12,247- at all.
Raman is offline   Reply With Quote
Old 2010-01-02, 18:47   #5
10metreh
 
10metreh's Avatar
 
Nov 2008

44228 Posts
Default

Quote:
Originally Posted by Raman View Post
@ 10metreh: I was writing this post right now, so that I was not able to modify the content on the Mersenne Wiki for the Cunningham Table entry on 12,247- at all.
What's the problem? Do you not want anyone else to edit those pages?
10metreh is offline   Reply With Quote
Old 2010-01-02, 19:16   #6
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

4E916 Posts
Default

Quote:
Originally Posted by 10metreh View Post
What's the problem? Do you not want anyone else to edit those pages?
Anyone can edit, but that I just wanted to let you know that I have looked up at the factors of 12,247- already, before itself. Simply, I wanted to let you know what I was doing so when you were editing that page, actually.
I got mail saying that the mersennewiki page edition has been made, while I was making up the above post itself.

Several points to be considered while editing only:
1. In the main tables, alignment have to be checked out for the new factors, once a composite is replaced by using two or more prime factors. If the remaining cofactor is composite, in case the factor is found out by using ECM, then the new prime factor should be aligned in black and then the remaining composite cofactor in red itself, or in magenta if for an already reserved number, ECM factor is got while pre-testing, and then the remaining cofactor is still composite only.
2. The #Fac value within the main tables has to be modified up accordingly, increasing by one and then retaining the plus, for an ECM factor if the cofactor is composite, and then increase by one, remove the plus for two prime factors, increase by two, remove the plus for composite splitting up into three prime factors.
3. The holes and then the remaining cunningham composites sorted by difficulty values table have to be modified only when the remaining cofactor is not composite at all. In the holes table, the entry has to be removed, adding up a <br> tag at the end of the specific table to ensure proper alignment of the remaining entries within that particular table. The count values at the top for number of composites remaining, reserved have to be changed up.
4. In the remaining Cunningham composites sorted by difficulty values table, the done entry has to be removed up, and then that count number has to be changed up right at the top.
5. Once numbers have been reserved by someone, holes table, remaining cunningham composites sorted by difficulty values table have to be edited up accordingly, and then the new reserved number within the main tables have to be changed up from red into magenta.
6. Once the new wanted lists are being issued up, within the holes table, the new wanted numbers have to be highlighted up accordingly.

I think that it is better to program a script to make sure that these changes are made automatically, but I don't know about how to program any script at all. Probably, that a submission query of factor for any number, and then determining that whether the remaining cofactor is prime or not, would be certainly enough. These things are all have to be embedded up within the Mersennewiki, at some place, somewhere. Of course, at the appropriate, right place.

Last fiddled with by Raman on 2010-01-02 at 19:18
Raman is offline   Reply With Quote
Old 2010-01-02, 19:30   #7
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

23518 Posts
Default

I am posting this up after actually checking out your changes:
Good work! But what are all you missed out is that:
#Fac was not properly aligned at all for 12,247- (9) should be moved one step to the right in order to be justified with the other entries.
The top countdown entry within the remaining holes sorted by difficulty values table has not been changed up from (411) down to (410) at all.
Be more careful the next time onwards, actually! Or that any of the counts will be missed out.

Last fiddled with by Raman on 2010-01-02 at 19:32
Raman is offline   Reply With Quote
Old 2010-01-02, 23:14   #8
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

2×3×52×19 Posts
Default

Quote:
Originally Posted by Raman View Post
Just one more question: Are factordb.com, rieselprime.de, (aliquot.de, mersenneforum.org, mersennewiki.org, mersenne.org, never ending...)
hosted for free or for money? How much cost, if any? Where do you get those domain names from? (I am not actually going to host any site, just want to know about it.)
i only can speak for rieselprime.de:

the account costs 25€ for 2009, so not to talk about. it's from www.one.com in Germany.
it includes e-mail, webhosting, FTP-access, MySQL-, PHP-support and statistics.
the next bill is for April 2010. a personel domain name was offered by one.com and i've choosen this one.
the online time was about 99.9% in 2009. the download/upload-timings are ok.
kar_bon is offline   Reply With Quote
Old 2010-01-03, 00:22   #9
ET_
Banned
 
ET_'s Avatar
 
"Luigi"
Aug 2002
Team Italia

12A016 Posts
Default

Quote:
Originally Posted by kar_bon View Post
i only can speak for rieselprime.de:

the account costs 25€ for 2009, so not to talk about. it's from www.one.com in Germany.
it includes e-mail, webhosting, FTP-access, MySQL-, PHP-support and statistics.
the next bill is for April 2010. a personel domain name was offered by one.com and i've choosen this one.
the online time was about 99.9% in 2009. the download/upload-timings are ok.
Fermatsearch.org is hosted for money. I have chosen to pay (up to) 200 euros per year because I liked the idea of managing a site of interest with Leonid Durman. When he left the project I decided to run it by myself.
I bought e-mail, web hosting, FTP access, MySQL, PHP and work my stats on it.

I'm happy with the provider (www.mclink.it).

Luigi
Luigi
ET_ is offline   Reply With Quote
Old 2010-01-27, 15:57   #10
Raman
Noodles
 
Raman's Avatar
 
"Mr. Tuch"
Dec 2007
Chennai, India

3·419 Posts
Default

1) Is the quick ECM feature of the factor database independently implemented by Syd
or does it call the GMP-ECM function? Because my first implementation of ECM after I had
studied the book of Crandall and Pomerance, was 100 to 1000 times slower than GMP-ECM.

2) Are there any projects that try to factor numbers of the form
k*b^n \pm 1
for some small values of k and b? Because in the Factor database, for the numbers of the form
2*3^n - 1
even for trivial SNFS of difficulty 90 digits are not entered into the database. Or does the factor
database does not include these factorizations till now? Numbers of form
3*2^n - 1
are atleast entered upto a decent higher level, SNFS difficulty 180. There are also other allied tables,
such as
5*2^n - 1, 7*2^n - 1, 9*2^n - 1 too,
and then their positive sides as well only, etc.

3) I have collected the list of all possible factoring projects. There are lots for which tables are possible, similar to the Cunningham tables. For each of the projects, the number of numbers are extremely large. The Cunningham project numbers are only a very tiny portion of all the numbers of the other factoring projects. Similar to the timeline of evolution of human beings since the Earth was formed up. But the numbers factored within the Cunningham project, are much advanced, the most made progress of all the factoring projects.
Each project has specific feature, for some derived (algebraic) factors exist, for some Aurifeuillian factors are possible, for some SNFS is applicable, for some only GNFS can be used, while for the RSA numbers, not even ECM is feasible actually.
For the other projects, like Fibonacci, Homogeneous Cunningham, Near Powers, Near Repdigits, etc. I would like to contribute, but they will have small ECM factors as well. I will not contribute until the small ECM factors, atleast upto 50 digits are being eliminated. Otherwise it will lead to disappointment with smaller factors. The Cunningham project itself has lot of easier numbers to do so first of all.
For aliquot sequence, I would like to contribute, but it will give series of smaller numbers, which need to be managed regularly. The factors will come quickly, and leads to waste of computing power with my resources. For the Cunningham numbers, which are much large enough, I need to run at start only, after that I will not need to turn back until the given sieve range are completed.

4) 6,355- should have been finished sieving today, and I planned to start off with 6,355+ today. But, it seems to be extending a bit. 2,935- is almost sieved, at 11 GB. I will need to sieve till 14 GB. Shortly, I will post my status of Cunningham numbers at the table at Mersenne Wiki and then update it regularly, every day.

5) Are there any projects that are trying to factor numbers like
\frac{10^{317}-1}{3} - 10^{199} or \frac{10^{317}-1}{3} - 10^{235}?
Is SNFS applicable for numbers of these type as well?
It seems that the factor database does not take care of the number of these types, as well.
Even the small factors do not pop out at all.
For example,
\frac{10^{317}-1}{3} - 10^{223} has a factor
The factor database does not mention about that at all.

There are many similar factoring projects, that can be found out at
Fibonacci Near Repdigits Smarandache

By the way, who maintains up that near repdigits page? It seems that there are many
contributions recently, and then who contribute to it up regularly? It seems that many are
from over here only?
Quote:
I am thankful to contributors 10metreh, Andreas Tete, Dmitry Domanov, Greg Childers, Ignacio Santos,
Jeff Gilchrist, Luigi Morelli, Makoto Kamada, Markus Tervooren, Max Dettweiler, Philippe Strohl,
Sander Hoogendoorn, Serge Batalov, Thomas Womack, Cedric Vonck, Chris Monico,
Phil Carmody, Shaopu Lin, ...
Really is it so? Certainly, plenty of ECM factors have to be eliminated from these tables, before
applying SNFS to any of these candidates. I am busy with the Cunningham project, anyway it
is my first choice. My subsequent choices will be Fibonacci, Lucas tables and then the Homogeneous
Cunninghams only, and then only these near repdigits tables, etc. because of the peculiar properties
of the Cunningham, Fibonacci, Lucas, Homogeneous Cunningham tables only. Any factor will be of
the form 2kn+1 or 2kn-1, existence of algebraic, Aurifeuillian factors, natural candidates, SNFS more
natural even for remaining residues, etc. For numbers of the form 2*3n-1 or 3*2n-1,
every value of n has to be tested up regularly, even fully as a SNFS candidate. Algebraic derived
factors are only being possible for some of them. For the Cunningham candidates, SNFS is naturally
being applicable even to the remaining residue cofactor numbers, and then only odd values of n need
to be cared about. Even values of n, everytime naturally split up into two factors of the same size only, right?

Last fiddled with by Raman on 2010-01-27 at 16:00
Raman is offline   Reply With Quote
Old 2010-01-27, 16:11   #11
Mini-Geek
Account Deleted
 
Mini-Geek's Avatar
 
"Tim Sorbera"
Aug 2006
San Antonio, TX USA

17·251 Posts
Default

Quote:
Originally Posted by Raman View Post
1) Is the quick ECM feature of the factor database independently implemented by Syd
or does it call the GMP-ECM function? Because my first implementation of ECM after I had
studied the book of Crandall and Pomerance, was 100 to 1000 times slower than GMP-ECM.
It calls GMP-ECM. The details have been posted somewhere, but I don't feel like digging them up right now.

As for the rest of your questions: the factor DB has factors for every form of number that people care to submit to it...so..yeah.
Mini-Geek is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Other Factordb Problems wblipp FactorDB 448 2020-10-14 16:07
A suggestion for factordb. enzocreti FactorDB 0 2018-03-02 09:12
Extending Factordb carpetpool FactorDB 6 2017-01-23 11:04
FactorDB PRP's smh FactorDB 231 2015-07-28 02:30
Automatically updating factordb chris2be8 FactorDB 9 2012-05-03 10:20

All times are UTC. The time now is 04:53.

Sun Oct 25 04:53:56 UTC 2020 up 45 days, 2:04, 0 users, load averages: 1.76, 1.52, 1.60

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.