mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Aliquot Sequences

Reply
 
Thread Tools
Old 2019-03-20, 08:37   #188
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

68410 Posts
Default

Quote:
Originally Posted by kar_bon View Post
The only way I see is to download all ELFs from open seqs and run a script over these files, scanning for those C30/C60. And perhaps all occurences of a (upgoing) C60 is relevant, not only the last one in a sequence.
Example: 158814 -> C60's at index 934, 2200 and 2722
So a merge could happen (before) the first C60 but you only be aware until the last C60, ~1800 indices later if only the last C60 will be listed.
Finally, I agree with kar_bon. I am trying to explain again how I understood the argument of the message quoted above by kar_bon, to make sure I understood correctly.
The example I propose below with C30 remains valid with C60, or C20. But indeed, the problem is that the file size increases if the number of digits of C is smaller.

Example :

Let's imagine a known S sequence stored in FactorDB:
index 200 : C30
index 1000 : C160
index 1800 : C30

1) I calculate a sequence S_1 on my own computer.
This sequence S_1 merges with S at index 190.
It is necessary to store the C30 of the index 200 of S, otherwise I have to calculate S_1 until the index 1800 to notice the fusion. It will take weeks !

2) I calculate a sequence S_2 on my own computer.
This sequence S_2 merges with S at index 1790.
It is necessary to store the C30 of the index 1800 of S, otherwise, if I have stored in my list only the first C30 of S, I will never notice the merging of S_2 with S.

Thanks to this example, I think we have to store all the C30s (or C20s or C60s... according to my choice) for an aliquot OE sequence from FactorDB.

Quote:
Originally Posted by LaurV View Post
I have my own 10, 12, 15, 20, 30, 50, 80 digits lists. That is the first x digits number that some sequence reach, for which I store the "smallest starter" and the "longest sequence". They are huge files, but they can be shared if someone needs. Better is if you have trouble finding a merge, post it and I may help...
Thanks a lot !

Quote:
Originally Posted by LaurV View Post
You always store the first. In fact, there are no merges "so high", all merges happen at very low digit count. The reason to store higher lists (people use to keep 80 and 100) is that the search is easier (less items in the list). Low-digits lists are huge. If two sequences "play yo-yo", there is still a "lowest point" where they merge, which is then (after raising) the "first" reach. It doesn't make sense to store all crosses. For example, you have two sequences that reach 10, 20, 30, 40 digits without merging, you will have stored one number for each, at 10, 20, 30, etc. If then both drop and merge, if at least one of them dropped below 7 digits, then the next 10-digits cross is already in the list, from the sequence starting with the last drop. If they drop no lower than 9 digits, then merge, and raise back (as one sequence) to10, 20, 30, 40, you will not detect this, as the new cross into the 10, etc, will not be recorded, because there is no record for sequences starting with 9 or more digits. But immediately when they cross to 50 digits, it will be detected (and recorded).
LaurV, I don't know I may have misunderstood your message, but it seems to me that you're saying that we have only to store the first C30 of FactorDB's OE sequences (But maybe I'm wrong !), is that right ?
For the reasons explained at the beginning of this post, I think, like kar_bon, that all the C30s (or C20s, or C60s...) in a sequence should be stored to establish the list that will allow us to find mergers efficiently.
garambois is offline   Reply With Quote
Old 2019-03-24, 10:47   #189
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

22·32·19 Posts
Default

OK, page updated.
Thank you to all.

My own calculations :
2^492 finished after three weeks of computation on it with 8 threads !
garambois is offline   Reply With Quote
Old 2019-04-06, 10:07   #190
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

22·32·19 Posts
Default

OK, page updated.
Thank you to all.

My own calculations :
2^491 and 2^498 finished !
garambois is offline   Reply With Quote
Old 2019-04-09, 10:12   #191
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

68410 Posts
Default

In fact, I believe that in just a few weeks, I could put online an effective database to find aliquot sequence mergers, like Wolfgang Creyaufmuller's old database (http://www.aliquot.de/aliquot.htm#Datenbanken). This database would allow all mergers with all aliquots sequences of the Blue Page to be found, up to 3,000,000.

This database would be a text file like those attached to this post to give two examples (C_60 and C_60_80).

The best option should now be chosen :
1) Should only all C60s be stored for each aliquot sequence or only all C80s (or only larger terms) for each aliquot sequences ? (only one file with C_60 for example)
2) Should all C60s and all C80s (and possibly others) be stored separately, like attached file C_60 for example ? (several files for C_60, C_80...)
3) Should all C60s and C80s (and possibly others) be stored on the same page, like attached file C_60_80 for example ? (only one file with C_60, C_80... inside)

It would certainly be necessary to update this page at least once a year, which I could do.

Do you think this work would be useful for all us ?
Do I have to do it and with which options (1, 2 or 3 or another one I wouldn't have thought of ?)
Attached Files
File Type: zip c6080.zip (31.5 KB, 67 views)
garambois is offline   Reply With Quote
Old 2019-04-14, 08:58   #192
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

10101011002 Posts
Default

OK, page updated.
Thank you to all.

My own calculations :
2^481, 2^483, 2^485, 2^487, 2^489 and 2^490 finished !
This week, I chose aliquot sequences that are easier to calculate !

garambois is offline   Reply With Quote
Old 2019-04-23, 15:31   #193
richs
 
richs's Avatar
 
"Rich"
Aug 2002
Benicia, California

101011001102 Posts
Default Drop 439^18

Quote:
Originally Posted by richs View Post
Reserving 439^18
Dropping 439^18. Added over 1000 terms and now at 135 digits. The current C130 is ECM'ed to t25.
richs is online now   Reply With Quote
Old 2019-04-28, 13:52   #194
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

22·32·19 Posts
Default

OK, page updated.
Thank you to all.

My own calculations :
3^228, 3^232 and 3^236 up to 120 digits.
garambois is offline   Reply With Quote
Old 2019-05-11, 19:50   #195
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

22·32·19 Posts
Default

Here are 3 new informations :

1) OK, page updated.
Thank you to all.

2) My own calculations :
All calculations are completed for base 3 except for 3^108 due to the reported and still uncorrected FactorDB error (https://www.mersenneforum.org/showth...t=19737&page=3 post #25).
I now reserve and calculate the aliquot sequences of base 5.

3)Excuse me, but a few weeks ago, I realized that I told some big nonsense in post #188 on this page in example 1).
If I store only the last C30 of the known S sequence on FactorDB, that is enough.
If the sequence S_1 meets S at index 190, then FactorDB will suddenly give me all the terms of S_1 until the last known term of the already known on FactorDB aliquot sequence S.
Then I just have to consider the last C30 of S_1 to identify the already known fusion sequence S.
It seemed reasonable to me to store all the latest C80s (rather than C_30s) of all known Open-End aliquots sequences up to 3,000,000.
You can find this work here :

https://www.mersenneforum.org/showthread.php?t=24423
garambois is offline   Reply With Quote
Old 2019-06-02, 09:50   #196
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

22·32·19 Posts
Default

OK, page updated.
Thank you to all.

My own calculations :
The calculations for base 5 will be completed in a few weeks.
garambois is offline   Reply With Quote
Old 2019-07-06, 08:19   #197
garambois
 
garambois's Avatar
 
"Garambois Jean-Luc"
Oct 2011
France

68410 Posts
Default

Page updated.
Thank you to all for your help !

My own calculations :

1) All aliquot sequences in base 5 have been calculated up to 120 digits.

2) I now reserve base 7.

3) I also hope to complete all calculations from base 2 up to 2^500 during the summer. But the calculations are very long ! Soon there will only be green cells in the table of base 2 !
I was hoping to find the "rare beast" : a power of 2 that would have been Open-End ! It may be for a few years from now, when I will extend the calculations to 2^600....
garambois is offline   Reply With Quote
Old 2019-08-06, 20:06   #198
Happy5214
 
Happy5214's Avatar
 
"Alexander"
Nov 2008
The Alamo City

773 Posts
Default

I have finished 21^i up to i=30 and am releasing up to that point. I'll continue from i=32 (i=31 terminates) after I clear some of my <3M backlog.
Happy5214 is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Broken aliquot sequences fivemack FactorDB 46 2021-02-21 10:46
Broken aliquot sequences schickel FactorDB 18 2013-06-12 16:09
A new theorem about aliquot sequences garambois Aliquot Sequences 34 2012-06-10 21:53
poaching aliquot sequences... Andi47 FactorDB 21 2011-12-29 21:11
New article on aliquot sequences schickel mersennewiki 0 2008-12-30 07:07

All times are UTC. The time now is 15:12.


Sat Nov 27 15:12:51 UTC 2021 up 127 days, 9:41, 0 users, load averages: 1.42, 1.33, 1.17

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, 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.