mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Aliquot Sequences

Reply
 
Thread Tools
Old 2009-03-05, 17:48   #1
Andi47
 
Andi47's Avatar
 
Oct 2004
Austria

2×17×73 Posts
Default serious bug in aliquot.ub

Quote:
Originally Posted by schickel View Post
Andi47, unfortunately it looks like you got bit by the ubasic bug again. The factorization of line 1581*) was recorded wrong, invalidating the succeeding lines. I re-did a couple of lines; it looks like the ubasic program has serious issues when the sequence is up that high.
*) I think this refers to sequence 181410



I will try if the bug is reproducible this weekend and additionally try to run a few lines manually.

A second case of the bug, which seems to be reproducible (at least I got the same set of nonsense factors with the same "variable area full" crash three times before I gave up) is in sequence 168804, starting with line 694 and trying to calculate line 695:

Quote:
Originally Posted by Andi47 View Post
Sequence 168804

Code:
 690 .	 50103592439273417765455674112419058395015246692752819021428935169183229439339883436403456844 = 2^2 * 3^3 * 17 * 30824336467 * 491795222813 * 13584298984407563519 * 114083094084012348329 * 1161608078703468047677953049
 691 .	 87435680928179628070197301637552945425210263223149452662176810277141381087583138264306943156 = 2^2 * 3^3 * 17 * 2137 * 369397632948149979033671706760102207657693 * 60327786293702353483825607354175652491638731
 692 .	 152696151450202438359481984205275535262156186630649022909831360423191454696892957165345341004 = 2^2 * 3^2 * 273607447 * 1154753038283154251 * 15241503701075456647 * 880806979345377075224336852056163707065321521
 693 .	 233285788348523582351079428302479202091390805818226802622805708025054796780158201758336268212 = 2^2 * 3^2 * 631 * 673 * 622807062847608389 * 24501226740208998749664531058825028416950572326161336857217774877031
 694 .	 358220989634069824870234071770131383640271305023140756549961418271453639806064943450895838028 = 2^2 * 3^2 * 7406233 * 72520355581 * 113380251588132676571 * 854327445452909216994437 * 191262337054620722724732940313
I give up. This sequence hits too many bugs (now it's a reproducible "variable error full" in iteration 695, producing funny factors), and I don't have time to run it manually.

Last fiddled with by Andi47 on 2009-03-05 at 17:49 Reason: typo
Andi47 is offline   Reply With Quote
Old 2009-03-07, 12:10   #2
Andi47
 
Andi47's Avatar
 
Oct 2004
Austria

1001101100102 Posts
Default



@Smh: As the bug sometimes (but not always) comes with an "out of variable space" crash, could it possibly help to rise word 100 to something like "word 130" in line 220, or to rise the size of the matrices in line 230?
Andi47 is offline   Reply With Quote
Old 2009-03-08, 09:48   #3
smh
 
smh's Avatar
 
"Sander"
Oct 2002
52.345322,5.52471

4A516 Posts
Default

Quote:
Originally Posted by Andi47 View Post
@Smh: As the bug sometimes (but not always) comes with an "out of variable space" crash, could it possibly help to rise word 100 to something like "word 130" in line 220, or to rise the size of the matrices in line 230?
No idea. I'm by no means an UBasic expert. I'll guess the only way to find out is try?
smh is offline   Reply With Quote
Old 2009-03-08, 10:18   #4
schickel
 
schickel's Avatar
 
"Frank <^>"
Dec 2004
CDP Janesville

2×1,061 Posts
Default

Quote:
Originally Posted by Andi47 View Post
@Smh: As the bug sometimes (but not always) comes with an "out of variable space" crash, could it possibly help to rise word 100 to something like "word 130" in line 220, or to rise the size of the matrices in line 230?
Quote:
Originally Posted by smh View Post
No idea. I'm by no means an UBasic expert. I'll guess the only way to find out is try?
Without being able to look at the ubasic source, I'd bet not. The "word" size is the bit-size of the variables. I'm assuming that the "out of variable space" means that the total amount of space available for variables is exhausted, not that any one variable has grown too big. (In fact, a test would be to set "word" low and see what happens if a variable overflows.....does it cause an error, or does it wrap around?)

My impression has been that one type of crash occurs when a composite (especially a large one) factors into "too many" primes. If space for array variables is only allocated as the array is filled, that would explain that type of crash. The limited variable space runs out when "too many" factors are found.

With "word" at 100, the integer precision is set at ~481 digits. A test you might try is setting "word" lower and see if that causes trouble with the internal ECM code. "word 50" should be adequate for current work, that's 240 integer digits....not sure where that puts the floating point precision, though.

Edit: Just tested....if a variable overflows, an "overflow in ...." error is reported. So you might try setting "word" lower and see what happens.

Last fiddled with by schickel on 2009-03-08 at 10:28 Reason: Added "edit"
schickel is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Broken aliquot sequences schickel FactorDB 18 2013-06-12 16:09
aliquot escape firejuggler Aliquot Sequences 26 2012-01-19 08:15
New σ for Aliquot JohnFullspeed Aliquot Sequences 18 2011-08-20 21:11
aliquot and mersenne firejuggler Aliquot Sequences 1 2010-06-01 23:47
Aliquot Sum Function grandpascorpion Math 5 2008-02-06 15:12

All times are UTC. The time now is 05:57.

Fri Dec 4 05:57:00 UTC 2020 up 1 day, 2:08, 0 users, load averages: 1.29, 1.12, 1.08

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.