mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Aliquot Sequences

Reply
 
Thread Tools
Old 2015-10-04, 20:48   #34
Sergei Chernykh
 
Jun 2015
Stockholm, Sweden

5316 Posts
Default

As promised, my own amicable pairs database is up and running now! I'll publish all my new pairs there from now on. Everyone is welcome to submit their findings. Submissions are processed automatically at the moment you do them, so there is no "wait-to-be-published" frustration anymore. I've already published everything I found so far - it's almost half a million new pairs.
Sergei Chernykh is offline   Reply With Quote
Old 2015-10-04, 21:44   #35
R. Gerbicz
 
R. Gerbicz's Avatar
 
"Robert Gerbicz"
Oct 2005
Hungary

101011101112 Posts
Default

Quote:
Originally Posted by Sergei Chernykh View Post
As promised, my own amicable pairs database is up and running now! I'll publish all my new pairs there from now on. Everyone is welcome to submit their findings. Submissions are processed automatically at the moment you do them, so there is no "wait-to-be-published" frustration anymore. I've already published everything I found so far - it's almost half a million new pairs.
Thanks! Made some checks what you do. And your site recognise this as a valid pair. What is a problem:

33 Anonymous 2015
310310=2*13*5*31*77
318682=2*13*7*17*103

(if 77 would be a prime then this is a valid pair). So you have to check that each factor is really prime.
(currently I don't see other problems)
R. Gerbicz is offline   Reply With Quote
Old 2015-10-04, 21:53   #36
R. Gerbicz
 
R. Gerbicz's Avatar
 
"Robert Gerbicz"
Oct 2005
Hungary

1,399 Posts
Default

Don't know what Pat did, but I would use say 20 Miller Rabin test (see https://en.wikipedia.org/wiki/Miller...primality_test) for primality checking if you want a short and fast code. Some month ago I have checked the files on Pat's site and all of the factors passed this check. (10 Rabin tests per factor could be too few).
R. Gerbicz is offline   Reply With Quote
Old 2015-10-05, 04:40   #37
Sergei Chernykh
 
Jun 2015
Stockholm, Sweden

8310 Posts
Default

Quote:
Originally Posted by R. Gerbicz View Post
Thanks! Made some checks what you do. And your site recognise this as a valid pair. What is a problem:

33 Anonymous 2015
310310=2*13*5*31*77
318682=2*13*7*17*103

(if 77 would be a prime then this is a valid pair). So you have to check that each factor is really prime.
(currently I don't see other problems)
I have Miller-Rabin test in the code, but it was turned off because it slowed down verification a lot. I was submitting the whole database this weekend. Thanks for reminding :D
Sergei Chernykh is offline   Reply With Quote
Old 2015-10-05, 05:00   #38
R. Gerbicz
 
R. Gerbicz's Avatar
 
"Robert Gerbicz"
Oct 2005
Hungary

1,399 Posts
Default

Quote:
Originally Posted by Sergei Chernykh View Post
I have Miller-Rabin test in the code, but it was turned off because it slowed down verification a lot. I was submitting the whole database this weekend. Thanks for reminding :D
OK.
Check c2_6.txt, there is a missing empty line on line number=48.
R. Gerbicz is offline   Reply With Quote
Old 2015-10-05, 05:23   #39
Sergei Chernykh
 
Jun 2015
Stockholm, Sweden

83 Posts
Default

Quote:
Originally Posted by R. Gerbicz View Post
OK.
Check c2_6.txt, there is a missing empty line on line number=48.
Fixed.
Sergei Chernykh is offline   Reply With Quote
Old 2015-10-07, 11:17   #40
AndrewWalker
 
AndrewWalker's Avatar
 
Mar 2015
Australia

5216 Posts
Default

Quote:
Originally Posted by Sergei Chernykh View Post
He seems to ignore all my submissions, at least he e-mailed to me that "I have put off breaking up your remaining files" after adding some of my 16-digit pairs in August. Anyway, I'm tired of waiting. I'm creating my own web page for amicable pairs database. It'll be ready this weekend. Submission form already works and I've uploaded all pairs up to 60 digits from Pat's site. There will be no restrictions on file size and submissions will be fully automatic
I would strongly encourage you to work with Pat so we have only one database not two. Otherwise we will
have different discovers for pairs in the databases which will just be a nightmare. I've just had some of mine from a
month or to back processed, did you resend any of yours in the smaller format?

Please, please don't make things harder for all concerned.

Maybe there's a way you can update the file of a size and send it to him so it can be directly included?

PS If you split up larger files into the correct size and zipped them together I'm sure he'd be happy to receive them by email, dropbox or otherwise if
asked beforehand. I sent a few larger lots to Pat this way.

Andrew

Last fiddled with by AndrewWalker on 2015-10-07 at 11:26
AndrewWalker is offline   Reply With Quote
Old 2015-10-07, 11:26   #41
Sergei Chernykh
 
Jun 2015
Stockholm, Sweden

5316 Posts
Default

Quote:
Originally Posted by AndrewWalker View Post
I would strongly encourage you to work with Pat so we have only one database not two. Otherwise we will
have different discovers for pairs in the databases which will just be a nightmare. I've just had some of mine from a
month or to back processed, did you resend any of yours in the smaller format?

Please, please don't make things harder for all concerned.

Andrew
Processing some recent 16-digit submissions and completely ignoring the fact that all 16-digit pairs were submitted in the middle of July (and they're still not processed) doesn't give Pat's database more credibility, does it? We need better solution than his "please wait two months" approach. My database has automatic submissions so it's by definition better and it's unbiased. I'll keep updating my database and syncing it with Pat's database regularly.

P.S. Pat should just submit all his new pairs directly to my database, together with everyone else.

Last fiddled with by Sergei Chernykh on 2015-10-07 at 11:28
Sergei Chernykh is offline   Reply With Quote
Old 2015-10-07, 11:32   #42
Sergei Chernykh
 
Jun 2015
Stockholm, Sweden

83 Posts
Default

Quote:
Originally Posted by AndrewWalker View Post
PS If you split up larger files into the correct size and zipped them together I'm sure he'd be happy to receive them by email, dropbox or otherwise if
asked beforehand. I sent a few larger lots to Pat this way.
I did this in the beginning of August, all files were in the exact format he asked for in one zip archive. No response, he didn't even try to download them.
Sergei Chernykh is offline   Reply With Quote
Old 2015-10-10, 10:13   #43
Sergei Chernykh
 
Jun 2015
Stockholm, Sweden

83 Posts
Default

Quote:
Originally Posted by R. Gerbicz View Post
Returning to speed:
With my new code I can find amicable pairs up to 1e13 in less than 8 hours on a single(!) Core-i3-2350M (2.30 GHz)
8*3600*2.3*109=6.624*1013, so it uses less than ~6.6 CPU cycles per number. I've done some new optimizations in the past few days and tested my new code on 1013:

8 threads: 3150.31 s (hyperthreading on)
4 threads: 3541.86 s (hyperthreading off)
I didn't test 1 thread, but it would be 3541.86 * 4 = 14167.44 s = 3h 56m 7.44s
Or ~5.52 CPU cycles per number.

Some tricks I used are new (at least I didn't find them in any of published atricles). I'll publish my code somewhat later this year, together with an article describing the algorithm.
Sergei Chernykh is offline   Reply With Quote
Old 2015-10-12, 07:00   #44
AndrewWalker
 
AndrewWalker's Avatar
 
Mar 2015
Australia

2·41 Posts
Default

While ever Pat still has his site I plan to submit new pairs to both. I remember Jan used to send back with each
submission a list of just the new pairs, would this be feasible (probably limited to something sensible like 5 MB if by email)?

Then I could split these up and send them to Pat (for the BDE stuff I'm doing most found are known but go up to a longer length).

I also have a record size pair over 40k digits I found with Paul Jobling a few years ago which was sent to Jan but never processed.
Sometime I'll dig it out and test both pages with it!!!

Andrew
AndrewWalker is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Search of all even-15-digit Aliquot cycles Drdmitry Aliquot Sequences 25 2016-12-16 15:26
Program for searching all odd-16-digit Aliquot cycles Drdmitry Aliquot Sequences 302 2016-05-11 02:17
Is a search for aliquot 3-cycles feasible? schickel Aliquot Sequences 7 2013-02-08 01:33
Small search of cycles with odd and even elements Drdmitry Aliquot Sequences 0 2011-12-14 13:50
Jan Munch Pedersen's Tables of Aliquot Cycles R. Gerbicz Math 0 2010-07-01 12:30

All times are UTC. The time now is 10:38.

Wed Sep 30 10:38:47 UTC 2020 up 20 days, 7:49, 0 users, load averages: 1.16, 1.27, 1.30

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.