mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   FactorDB (https://www.mersenneforum.org/forumdisplay.php?f=94)
-   -   Factoring small composites (https://www.mersenneforum.org/showthread.php?t=24411)

hansl 2019-05-09 04:34

Factoring small composites
 
I was considering trying to factor some composites from the Downloads -> "List of 1.000 randomly chosen, small composite numbers"
Is this useful to the overall purpose of factordb to fully factor these small composites?

Also is there any particular script or program I can use to only evaluate the expressions created by this, and spit them into a simple file of decimal numbers by line, so I can more easily test with various programs?

DukeBG 2019-05-09 20:11

[QUOTE=hansl;516193]I was considering trying to factor some composites from the Downloads -> "List of 1.000 randomly chosen, small composite numbers"
Is this useful to the overall purpose of factordb to fully factor these small composites?[/QUOTE]
They end up factored by someone eventually anyway, kinda.

Nobody is specifically waiting for them – and if someone is, they are wrong to do so, because factordb isn't and shouldn't be a treated as a free factoring service. Anyone is better off factoring numbers themselves than submitting them to FDB and waiting for them to be factored.

Unfortunately, it kinda gets used like that by whoever since it's so open. There's a lot of "garbage" composite numbers that people might've not even submitted consciously, but just looked up some formulae.

I've downloaded that list right now and i see it has 91 and 92-digit numbers. You can see how many total there are [URL="http://factordb.com/stat_1.php"]here[/URL].

[QUOTE]Also is there any particular script or program I can use to only evaluate the expressions created by this, and spit them into a simple file of decimal numbers by line, so I can more easily test with various programs?[/QUOTE]

Generally you wouldn't need to – programs like yafu (which I feel is most suitable for this task) can take the expressions as input no problem.

LaurV 2019-05-10 05:15

There is a "yoyo" perl script somewhere, who does a wonderful job. This script takes a random composite from the db and factors it using yafu (or other external tool) then it reports the result to the db. It is nice in sense that it is "set it and forget it", and you can also specify the desired digit and from where the random composite is taken (like "get a 120-digits composite from the smallest 100 numbers which are 120-digit composites). The "randomness" is to avoid duplication of work - it still happens sometime because people want to factor "the smallest composite available" and fdb does not have an assignment procedure, but the probability is low. Of course, you must have installed some perl, and some factoring tool, like yafu.

You can search the forum for yoyo.pl or so, and if you can't find it, I will post it soon when I get home (I do not have it here at job).

DukeBG 2019-05-10 06:56

Small correction. FDB doesn't actually have an ordinality of numbers with the same digit count (by, erm, the numbers themselves). All the outputs in the same digit size are arbitrarily random. So you cannot get "smallest" 120-digits, for example. However, they are always essentially in the same order – the order they are read from the DB. So those scripts randomize that – the "page" from which the numbers are taken.

yoyo 2019-05-10 10:30

[QUOTE=LaurV;516354]There is a "yoyo" perl script somewhere, who does a wonderful job. This script takes a random composite from the db and factors it using yafu (or other external tool) then it reports the result to the db. It is nice in sense that it is "set it and forget it", and you can also specify the desired digit and from where the random composite is taken (like "get a 120-digits composite from the smallest 100 numbers which are 120-digit composites). The "randomness" is to avoid duplication of work - it still happens sometime because people want to factor "the smallest composite available" and fdb does not have an assignment procedure, but the probability is low. Of course, you must have installed some perl, and some factoring tool, like yafu.

You can search the forum for yoyo.pl or so, and if you can't find it, I will post it soon when I get home (I do not have it here at job).[/QUOTE]

-> [url]https://www.rechenkraft.net/wiki/Benutzer_Diskussion:Yoyo/factordb#yafu.pl[/url]

chris2be8 2019-05-10 16:02

It's at [URL]http://mersenneforum.org/showthread.php?t=19232&page=3[/URL] post 25.

Note you must change factorization.ath.cx to factordb.com since the old DNS entry doesn't work now.

Chris

chris2be8 2020-11-01 17:31

[QUOTE=unconnected;561779]Looks like someone flooded factordb with thousands of small composites.
[url]http://factordb.com/stat_1.php[/url][/QUOTE]

I'm working on them. But it's like painting the Forth Bridge. And starting to get annoying.

[quote]
One fool can ask more questions than a thousand wise men can answer.
[/quote]

Though I don't think that alone would slow down sequence processing.

Chris

garambois 2020-11-01 19:43

Page updated, but only for bases 2, 3, 20, 21, 23, 162, 439, 496.
31^36 and complete base 385 are reserved for me.

Thanks to all for your help !
Please check if all your requests have been taken into account.

The other bases will be updated in the next few weeks, as it takes a lot of time for each base .

EdH 2020-11-02 13:18

Note: Post was edited to remove thread irrelevant content from this copied post. Original content can be found [URL="https://www.mersenneforum.org/showpost.php?p=561927&postcount=663"]here[/URL].

@chris2be8: I wonder if the large number of small composites on factordb was just a result of the rebuild. I did factor a few thousand, but quickly got close to the hour's limit for a couple values. They seemed to be clearing pretty fast at the time, so I left the rest alone.

chris2be8 2020-11-02 17:13

I don't think the small composites are from the rebuild. I've seen quite a few numbers like:
[code]
(738468*49##+557)/35840141013305049257323
(2^295+94923)/13352559799576307
[/code]

So it's general junk, not components of sequences. I've been tackling the ones that can be done by SNFS in the 80-90 digit range as well as everything from 70-79 digits. But don't hold your breath waiting for me to clear them all.

Chris

EdH 2020-11-02 17:37

[QUOTE=chris2be8;561962]I don't think the small composites are from the rebuild. I've seen quite a few numbers like:
[code]
(738468*49##+557)/35840141013305049257323
(2^295+94923)/13352559799576307
[/code]So it's general junk, not components of sequences. I've been tackling the ones that can be done by SNFS in the 80-90 digit range as well as everything from 70-79 digits. But don't hold your breath waiting for me to clear them all.

Chris[/QUOTE]
I had done about 50k, 90 dd, but noticed the count fell much more than my 50k, so I figured someone with some computing power was working there. I have since trimmed my work down considerabley and moved it to 97 dd. I'm actually factoring more than are showing up, which is reassuring, even if it is pretty low ATM compared to sometimes. If you're only working to 90 dd, maybe I'll move down to 91 when I throw some work their way.


All times are UTC. The time now is 01:09.

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