Factoring database

[kar_bon]

Originally Posted by klajok

[I am not sure if it is right thread for such problem]

At page http://factordb.com/search.php?id=167488215 there is listed number 57 which is obviously factored, but blocks progress of sequence: http://factordb.com/search.php?se=16...&action=last20 , additionally in text mode of the same sequence: http://factordb.com/search.php?se=16...&to=&text=Text there is similarly number 283.

"Repair sequence" doesn't work. Is there any workaround for this issue?

As mentioned, the number '57' declared as 'composite' is a Database error, 283 is ok.

As suggested before, you can use the erroneous index 43 of Invere Home Prime Base 6 (IHP6) with the corrected value and start from there.
I've done this here.

Please consider (not proven but obvious!):

1) every index of IHP6(n) is composite (divisible by 2) for even n
2) every index of IHP6(n) is composite (divisible by 3) for odd n, n == 0 MOD 3

-> Your IHP6(87) is divisible by 3 for every index and never ends in a prime!

So out of 30 n test only n's ending in 1,5,7,11,13,17,19,23,25 and 29!

Those seems to end in a prime quite quickly.

[wblipp]

Today I noticed that when a number is entered as n^2-1 the database identifies algebraic factors of n-1 and n+1, but when the number is entered as n^3-1 the database does not identify any algebraic factors. Do we know if Syd is planning to include cyclotomic algebraic factorizations?

[kar_bon]

It seems Syd has set up a new version of his FactorDB:

There's a link on top of the page New version is ready - give it a try!

So then to all: test and have a closer look!

[Syd]

Hi everyone,

I'm sorry the rewrite took so long! I had a lot of problems making everything thread-safe, cleaning up bugs, especially in the parser and so on. I think I got most of them, leaving only minor ones, but who knows .. as this is my first project to break the 500kb-sourcecode-barrier.

The rewrite uses a different database layout, this is why I had to convert the old one to the new one. It took more than 3 days on all 4 cores - a huge amount of data! The backup I used is about 5 days old - all factors added after that wont appear. Of cause you can submit them manually if you want, but its not needed, in a week or so I'll add all missing ones automatically.

There are a lot more changes. First, there is no "Quick ECM" button anymore. Instead there is a magnifying glass in the left column, this runs ECM, PM1 and PP1-Curves on the number. First click 0,3seconds, second 3s, then 30 and last one 100s, then the button disappears.

When searching terms with variable(s) there is another magnifying glass below all others - clicking this runs ECM-Curves on all unfactored terms. This may take a while ..

However, some things are not done yet.
-Login / Worker script
-Detecting merges in sequences, sequence overview
-Factor tables
-All Sequences end on a prime - even the ones that shouldn't.

Here is the URL:
http://factordb.com/new/index.php

Of cause I'll keep the old version up for a while.

Have fun

[kar_bon]

And a quick one still there:

In sequences the number is sometimes in short form like for Aliqout seq n=512:
(3)^4 = (3)^4

[Mini-Geek]

Alright, you finally got the rewrite public so we can (soon) stop worrying about the composite bug, and other things, and have new improvements. Thanks.
It should show somewhere what level of scan has been done and what will be done next time you click the button (if available). That will make it all make much more sense at a glance.

[kar_bon]

Originally Posted by Mini-Geek

It should show somewhere what level of scan has been done and what will be done next time you click the button. That will make it all make much more sense to the people.

As mentioned, it will disappear then! Try it!

[mdettweiler]

Quick question: what is the "search time" drop-down box on the Sequences page for?

[CRGreathouse]

It looks great, Syd!

Are you looking for workers, or have you scrapped that part? My cores are busy at the moment, but I may have some capacity later when I finish this project. I have the latest eecm...

[Batalov]

Is there a line length restriction in the parser?
I tried submitting the 126-digit factor of 2^919-1 ...
returns an empty page with
" factordb.com - 0 queries to generate this page (0.00 seconds) "

I bet you didn't expect factors to be that big, eh, Markus?


P.S. The report factors page worked. "2^919-1 = #factor#"

[Syd]

Originally Posted by Mini-Geek

Alright, you finally got the rewrite public so we can (soon) stop worrying about the composite bug, and other things, and have new improvements. Thanks.
It should show somewhere what level of scan has been done and what will be done next time you click the button (if available). That will make it all make much more sense at a glance.

Now it shows some "+" behind the icon, depending on the work done. I'll add the specific work done later - guess the "alt"-tag is great for that.

Originally Posted by mdettweiler

Quick question: what is the "search time" drop-down box on the Sequences page for?

When searching an unfinished sequence it spends this amount of time on every unfactored term - to avoid getting stuck too early.

Originally Posted by CRGreathouse

It looks great, Syd!

Are you looking for workers, or have you scrapped that part? My cores are busy at the moment, but I may have some capacity later when I finish this project. I have the latest eecm...

Thanks!
I plan to add the workers later, you are welcome to participate

Originally Posted by Batalov

Is there a line length restriction in the parser?
I bet you didn't expect factors to be that big, eh, Markus?

I set the limit to 80 digits .. way too low! Now its 8M digits. Should do it.