Broken sequences / duplicate entries

Syd · 2012-06-25, 21:28

Hi everyone,

sorry it took so long to repair the database.
There were a few underlying problems which caused the broken sequences, broken numbers, etc.

The major one was duplicate entries. I use separate tables for composites, status unknowns, primes, etc, and assume each ID appears in one table only. After last server crash this was not true anymore, which caused errors like "duplicate entry on PRIMARY KEY", false primes and so on. I just removed these ID´s from all tables and hope it fixed the sequences.

Then there was another bug when reporting factors which occurred under certain circumstances. Say a number x, new to the database, factors as a*b*c*d, and 3 factors are reported: a*b, b*c and a*c
This caused an "error on number" bug, which is hopefully fixed now.

And finally, large numbers, especially fermat numbers, were shown very slowly. This was caused by my modifications to gmp´s mpz_perfect_power_p function.. Now its back to default and fast again.

Please let me know if its not fixed!

I hope I´ll find some time to work on some other features soon

- Syd

firejuggler · 2012-06-25, 21:40

Thanks for the heads-up Syd

Dubslow · 2012-06-26, 03:40

Does anyone have an elf they can upload for 40320?

shickel's latest report puts at at line 1357 size 117, but if you click on the link, you can see it's a few hundred lines and 60 digits short. (When I first noticed it the DB only had it in the 600s.)

LaurV · 2012-06-26, 04:00

Quote:

Originally Posted by Dubslow

Does anyone have an elf they can upload for 40320?

shickel's latest report puts at at line 1357 size 117, but if you click on the link, you can see it's a few hundred lines and 60 digits short. (When I first noticed it the DB only had it in the 600s.)

That sequence seems to be still broken. I tried to run two cores on it, it should be faster to bring it back to index 1357 then to ask about... but aliqueit got angry with me at index 504.

The problem is the 18 digits cofactor 433850578177651157 of index 503, marked as prime. In reality this is a C18 which splits as 41283691*10509006527.

The right sigma of index 503 is i504=58434468573150132384528 and you can see there the continuation of the sequence, which is indeed a C117 at the end.

Dubslow · 2012-06-26, 20:46

Sigh... 731208 is also broken.

Code:

Verifying elf 731208
Verifying index 2799... 
ERROR: factor not prime: 504246042153376879

It's another case of floating between FF and P, and the sequence's next line is incorrect because of that.

schickel's text report shows it as "731208 3947. sz 112 2 * 3".

Dubslow · 2012-06-26, 21:50

And another.

Code:

Verifying elf 21318
Verifying index 300... 
ERROR: @index 300: value != sigma - n

Code:

 21318  1264. sz 116 2 * 3 * 5

Dubslow · 2012-06-27, 03:46

Code:

Verifying elf 641592
Verifying index 1467... 
ERROR: @index 1467: value != sigma - n

Code:

641592  1610. sz 114 2 * 3^2

Syd, did all the broken sequences need fixing by hand? Is it possible that these sequences were broken before you fixed it, meaning that the bug is in fact gone, but just that these sequences missed the cleanup somehow?

Edit: 641592 and 40320 both appear in this post (with the same problem lines) but the other two (21318 and 731208) were not previously known to be broken.

Dubslow · 2012-06-27, 05:10

Code:

Verifying elf 271680
Verifying index 440... 
ERROR: factor not prime: 336756866798891081

Code:

271680  2266. sz 117 2 * 3^3 * 5

Not previously known...

Code:

Verifying elf 331716
Verifying index 184... 
ERROR: factor not prime: 539550197360613767

Code:

331716  1117. sz 114 2 * 3^3 * 7^2

Also not previously known.

Edit: For Syd or someone else who wants to try (and is using *nix with wget and aliqueit installed) you could do this:

Code:

#! /bin/bash
for seq in $(cat AllSeqs.txt); do
    wget http://factordb.com/elf.php?seq=$seq&type=1
    aliqueit -t alq_$seq.elf >> aliquot.log
done

The problem is that it would take up to a week to run completely, but it would give a definitive list of all problems.

Dubslow · 2012-07-02, 22:53

Alright, the following should be a complete list of the 22 sequences with errors that occurred at a size <110. It's likely that there are at least a few sequences with errors at sizes >=110 that I haven't found.

The first three lines are the aliqueit check output, and the fourth is what AllSeq.txt reports is the correct status of the sequence. They're in no particular order, and this includes the others reported above.

Code:

Verifying elf 722250
Verifying index 149... 
ERROR: factor not prime: 576443222838292213

722250  1550. sz 113 2 * 3 * 5



Verifying elf 583308
Verifying index 145... 
ERROR: factor not prime: 494644839362210107

583308  1185. sz 118 2^4 * 3 * 31



Verifying elf 43974
Verifying index 326... 
ERROR: @index 326: value != sigma - n

 43974  4367. sz 115 2^2 * 3 * 7 * 103 * 14621



Verifying elf 434040
Verifying index 428... 
ERROR: @index 428: value != sigma - n

434040  1664. sz 115 2^2 * 3^2 * 5 * 331



Verifying elf 116424
Verifying index 712... 
ERROR: @index 712: value != sigma - n

116424  1398. sz 113 2^2 * 7



Verifying elf 506322
Verifying index 113... 
ERROR: factor not prime: 574402480684664617

506322  2985. sz 113 2^2 * 3 * 7 * 11



Verifying elf 936390
Verifying index 2353... 
ERROR: @index 2353: value != sigma - n

936390  4102. sz 117 2^4 * 3^3 * 31 * 271



Verifying elf 14598
Verifying index 1129... 
ERROR: factor not prime: 559829707982058169

 14598  1580. sz 116 2^2 * 3 * 7 * 13 * 251



Verifying elf 745824
Verifying index 1378... 
ERROR: @index 1378: value != sigma - n

745824  3334. sz 112 2^2 * 7



Verifying elf 976248
Verifying index 1465... 
ERROR: @index 1465: value != sigma - n

976248  1615. sz 119 2^3 * 3^3



Verifying elf 700128
Verifying index 231... 
ERROR: @index 231: value != sigma - n

700128  2244. sz   2 43



Verifying elf 406608
Verifying index 236... 
ERROR: @index 236: value != sigma - n

406608   965. sz 115 2^4 * 3



Verifying elf 239520
Verifying index 90... 
ERROR: factor not prime: 465271162599626183

239520   970. sz 114 2^3 * 3



Verifying elf 228956
Verifying index 1498... 
ERROR: @index 1498: value != sigma - n

228956  3476. sz 117 2^3 * 3 * 5 * 7



Verifying elf 246888
Verifying index 692... 
ERROR: factor not prime: 428238728474329177

246888  2044. sz 114 2 * 3



Verifying elf 566712
Verifying index 1093... 
ERROR: @index 1093: value != sigma - n

566712  1239. sz 111 2 * 3^2



Verifying elf 731208
Verifying index 2799... 
ERROR: factor not prime: 504246042153376879

731208  3947. sz 112 2 * 3



Verifying elf 21318
Verifying index 300... 
ERROR: @index 300: value != sigma - n

21318  1264. sz 116 2 * 3 * 5



Verifying elf 641592
Verifying index 1467... 
ERROR: @index 1467: value != sigma - n

641592  1610. sz 114 2 * 3^2



Verifying elf 271680
Verifying index 440... 
ERROR: factor not prime: 336756866798891081

271680  2266. sz 117 2 * 3^3 * 5



Verifying elf 331716
Verifying index 184... 
ERROR: factor not prime: 539550197360613767

331716  1117. sz 114 2 * 3^3 * 7^2



Verifying elf 40320
Verifying index 504...
ERROR: factor not prime: 433850578177651157

40320  1357. sz 117 2^3 * 3^2 * 5 * 2657

LaurV · 2012-07-03, 05:43

Good work man! We waiting for Syd now, hoping he will have some free time to fix them.

Dubslow · 2012-07-03, 05:51

Actually, I'm fairly sure there are two more sequences with errors that I missed because they show as terminated due to the errors. Long story short, I wasn't handling that properly, and it will take another week for me to find those. (700128 was a case of this, except that I happened to be watching at the right time; as a matter of fact, me accidentally catching 700128 is how I figured out I was mishandling the wrong-terminations.)

2012-06-25, 21:28	#1
Syd Sep 2008 Krefeld, Germany 11100110₂ Posts	Broken sequences / duplicate entries Hi everyone, sorry it took so long to repair the database. There were a few underlying problems which caused the broken sequences, broken numbers, etc. The major one was duplicate entries. I use separate tables for composites, status unknowns, primes, etc, and assume each ID appears in one table only. After last server crash this was not true anymore, which caused errors like "duplicate entry on PRIMARY KEY", false primes and so on. I just removed these ID´s from all tables and hope it fixed the sequences. Then there was another bug when reporting factors which occurred under certain circumstances. Say a number x, new to the database, factors as abcd, and 3 factors are reported: ab, bc and ac This caused an "error on number" bug, which is hopefully fixed now. And finally, large numbers, especially fermat numbers, were shown very slowly. This was caused by my modifications to gmp´s mpz_perfect_power_p function.. Now its back to default and fast again. Please let me know if its not fixed! I hope I´ll find some time to work on some other features soon - Syd Last fiddled with by Syd on 2012-06-25 at 21:41 Reason: typo

2012-06-26, 20:46	#5
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 16065₈ Posts	Sigh... 731208 is also broken. Code: Verifying elf 731208 Verifying index 2799... ERROR: factor not prime: 504246042153376879 It's another case of floating between FF and P, and the sequence's next line is incorrect because of that. schickel's text report shows it as "731208 3947. sz 112 2 * 3". Last fiddled with by Dubslow on 2012-06-26 at 20:49

2012-06-26, 21:50	#6
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 16065₈ Posts	And another. Code: Verifying elf 21318 Verifying index 300... ERROR: @index 300: value != sigma - n Code: 21318 1264. sz 116 2 * 3 * 5 Last fiddled with by Dubslow on 2012-06-26 at 21:50

2012-06-27, 03:46	#7
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 7221₁₀ Posts	Another broken sequence Code: Verifying elf 641592 Verifying index 1467... ERROR: @index 1467: value != sigma - n Code: 641592 1610. sz 114 2 * 3^2 Syd, did all the broken sequences need fixing by hand? Is it possible that these sequences were broken before you fixed it, meaning that the bug is in fact gone, but just that these sequences missed the cleanup somehow? Edit: 641592 and 40320 both appear in this post (with the same problem lines) but the other two (21318 and 731208) were not previously known to be broken. Last fiddled with by Dubslow on 2012-06-27 at 03:53

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Broken aliquot sequences	fivemack	FactorDB	46	2021-02-21 10:46
Graphs broken for aliquot sequences on factordb	fivemack	Aliquot Sequences	4	2018-03-23 11:54
bugreport: duplicate entries	maxal	FactorDB	1	2017-11-14 15:38
Broken aliquot sequences	schickel	FactorDB	18	2013-06-12 16:09
Wrong entries in database?	Unregistered	Information & Answers	4	2008-06-28 01:50

2012-06-25, 21:40	#2
firejuggler "Vincent" Apr 2010 Over the rainbow 2²×7×103 Posts	Thanks for the heads-up Syd

2012-06-26, 03:40	#3
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 1C35₁₆ Posts	Does anyone have an elf they can upload for 40320? shickel's latest report puts at at line 1357 size 117, but if you click on the link, you can see it's a few hundred lines and 60 digits short. (When I first noticed it the DB only had it in the 600s.)

2012-07-03, 05:43	#10
LaurV Romulan Interpreter "name field" Jun 2011 Thailand 3·23·149 Posts	Good work man! We waiting for Syd now, hoping he will have some free time to fix them.

2012-07-03, 05:51	#11
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 16065₈ Posts	Actually, I'm fairly sure there are two more sequences with errors that I missed because they show as terminated due to the errors. Long story short, I wasn't handling that properly, and it will take another week for me to find those. (700128 was a case of this, except that I happened to be watching at the right time; as a matter of fact, me accidentally catching 700128 is how I figured out I was mishandling the wrong-terminations.)