mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Aliquot Sequences

Reply
 
Thread Tools
Old 2011-11-29, 09:52   #1
schickel
 
schickel's Avatar
 
"Frank <^>"
Dec 2004
CDP Janesville

84A16 Posts
Default Any answers on missing factors?

Quote:
Originally Posted by LaurV View Post
Trying to dld different elf files from factorDB. PLENTY of factors missing from the elf files! What's going on?

See for example my sequence, 996666 term T1479. Where are the factors 65449*282281???????

For comparison: attached my elf file and the elf file just dld-ed (few times, in different computers, it is not from the computer!) few minutes ago.

Tried with other sequenced, same problem. Is someone "stealing" factors from the data base? :P (as we just argued on a parallel thread about the correctitude of the data base....)
Hmmm.....looking over some recent posts, did you ever get an answer to this one?
schickel is offline   Reply With Quote
Old 2011-11-29, 10:04   #2
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

2·4,481 Posts
Default

That problem was temporary, and related most probably to some internal things, optimization, tidying up, I don't know what the DB owner was doing in that time, the issue was solved next day, or next 24 hours, or so. Never happened since, at least for the sequences I have the .elf files as local copies (about 10 files). If I download them from the DB (in case somebody else did some work faster then me and I need to update my files), then I ALWAYS check (file compare byte by byte, use totalcmd) them before overwriting my local files with the DB files. Viceversa I don't do, I would just send last (newer) lines from my files (copy/paste them to "report factors" form) and never send a complete file.

Last fiddled with by LaurV on 2011-11-29 at 10:05
LaurV is offline   Reply With Quote
Old 2011-11-29, 10:13   #3
schickel
 
schickel's Avatar
 
"Frank <^>"
Dec 2004
CDP Janesville

41128 Posts
Default

Quote:
Originally Posted by LaurV View Post
That problem was temporary, and related most probably to some internal things, optimization, tidying up, I don't know what the DB owner was doing in that time, the issue was solved next day, or next 24 hours, or so. Never happened since, at least for the sequences I have the .elf files as local copies (about 10 files). If I download them from the DB (in case somebody else did some work faster then me and I need to update my files), then I ALWAYS check (file compare) them before overwriting my local files with the DB files. Viceversa I don't do, I would just send last (newer) lines from my files (copy/paste them to "report factors" form) and never send a complete file.
Glad to see that it was temporary.

Do you just do a text comparison, or do you check the actual integrity of the file? If you just do a text comparison, spacing differences would throw things off. If you want to check the integrity of an .elf file, aliqueit does a quick check when it starts up. I think there's also a switch to quit after checking a sequence. Clifford also has a couple UBASIC programs that can check .elf files. (You can run UBASIC through DOSBox.) These have been obsoleted by the DB and aliqueit.exe.....
schickel is offline   Reply With Quote
Old 2011-11-29, 11:34   #4
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

2·4,481 Posts
Default

That is why I specified totalcmd, to avoid the question :D AraxixMerge would do it too :D

In fact it does not matter, assuming my files are right and they are, because they are (were) checked every time when I restart aliqueit, then a "character" comparison would also do it. But I was talking about more then that. They WERE wrong on the DB, and I could easily see.

The attached example is artificially created now, just to show what I was saying (and seeing).

(sorry for the quality, the forum did not like its resolution - too big - and I had to resize it 50%, it is still understandable)
Attached Thumbnails
Click image for larger version

Name:	Capture.JPG
Views:	99
Size:	118.9 KB
ID:	7324  
LaurV is offline   Reply With Quote
Old 2011-11-29, 11:40   #5
schickel
 
schickel's Avatar
 
"Frank <^>"
Dec 2004
CDP Janesville

2×1,061 Posts
Default

Quote:
Originally Posted by LaurV View Post
That is why I specified totalcmd, to avoid the question :D AraxixMerge would do it too :D

In fact it does not matter, assuming my files are right and they are, because they are (were) checked every time when I restart aliqueit, then a "character" comparison would also do it. But I was talking about more then that. They WERE wrong on the DB, and I could easily see.

The attached example is artificially created now, just to show what I was saying (and seeing).

(sorry for the quality, the forum did not like its resolution - too big - and I had to resize it 50%, it is still understandable)
I didn't see the totalcmd reference.....one of the dangers of replying while someone is editing their post.

I wonder what the problem was. I download all the open sequences weekly or so and now I'm wondering if I should figure out something to verify them when I do that.....
schickel is offline   Reply With Quote
Old 2011-11-29, 11:56   #6
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

23×3×7×17 Posts
Default

Quote:
Originally Posted by schickel View Post
Do you just do a text comparison, or do you check the actual integrity of the file? If you just do a text comparison, spacing differences would throw things off. If you want to check the integrity of an .elf file, aliqueit does a quick check when it starts up. I think there's also a switch to quit after checking a sequence. Clifford also has a couple UBASIC programs that can check .elf files. (You can run UBASIC through DOSBox.)
On Clifford's page (last third of above link) there're several tools, also as WIN32-exe so chk.exe is the one. It will test, if the factors and aliquot sums are correct.
I also use TotalCmd and the "Compare by Content": it's something intelligent like missing lines are spotted (and compare after such normally further), ignore repeated spaces or other tiny goodies which are more than only comparing char by char.

Quote:
Originally Posted by schickel View Post
These have been obsoleted by the DB and aliqueit.exe.....
It's better to check twice with different programs as to run false/useless calculations.
For example: NewPGen and small Riesel-primes with n<100.
kar_bon is offline   Reply With Quote
Old 2011-11-29, 12:01   #7
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

23×3×7×17 Posts
Default

Quote:
Originally Posted by schickel View Post
I wonder what the problem was. I download all the open sequences weekly or so and now I'm wondering if I should figure out something to verify them when I do that.....
There was the problem, that the downloaded elf-files contained 'prime'-factors which were not prime but to next step was ok. I haven't checked that for a long time, but should be spot-checked again.
kar_bon is offline   Reply With Quote
Old 2011-11-29, 12:05   #8
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

2×4,481 Posts
Default

This is impossible in the middle of an aliquot sequence. Next terms would not be computed right. And we are talking about "old" sequences, terms computed years ago, not something new "in process" of factoring. The factors were MISSING, not replaced by some composites. Exactly like in the "fake" example above, see my image snap.
LaurV is offline   Reply With Quote
Old 2011-11-29, 12:24   #9
schickel
 
schickel's Avatar
 
"Frank <^>"
Dec 2004
CDP Janesville

2×1,061 Posts
Default

Quote:
Originally Posted by LaurV View Post
This is impossible in the middle of an aliquot sequence. Next terms would not be computed right. And we are talking about "old" sequences, terms computed years ago, not something new "in process" of factoring. The factors were MISSING, not replaced by some composites. Exactly like in the "fake" example above, see my image snap.
The problem previously was a "Set prime" button that Syd had made available to "trusted" users. Someone was setting composites as prime and that was screwing up the seqeunces that were generated with those numbers. This did include sequences of interest that we downloaded.

I think Syd has more checks in place now, including some internal integrity checks on the sequences. When he switched over to the new schema, it did take a long while for everything to get verified. And, more importantly, the treacherous button has been removed.

I think you called it right: something was going on internally that was causing a problem with downloading sequences......nothing that we have to worry about, since most of us (I assume) use aliqueit to extend sequences in easy territory. (I sure wish I had more there.....)
schickel is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Missing factors at the 'Known Factors' page MatWur-S530113 PrimeNet 11 2009-01-21 19:08
NewPGen missing factors? lavalamp Software 6 2009-01-14 13:46
Dumb question about missing srsieve factors robo_mojo Riesel Prime Search 4 2008-04-22 04:37
Missing factors henryzz Math 10 2007-12-08 12:45
GGNFS questions and Answers VJS Factoring 100 2006-01-15 08:34

All times are UTC. The time now is 22:54.

Thu Dec 3 22:54:19 UTC 2020 up 19:05, 1 user, load averages: 1.52, 1.45, 1.62

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.