Factoring database - Page 4

mdettweiler · 2008-12-16, 21:52

Just curious...how does the whole "assign number to workers" thing work? Is it some kind of system that automatically runs x amount of ECM curves on a given number (the specific amounts being as you described earlier in this thread)?

I noticed that when I tried taking an arbitrary C176 from the homogeneous Cunningham list and submitted it for ECM to low limits, it returned a message that says "Waiting for worker...". What exactly does this mean? Also, is it dynamic in some way? If I sit there and stare at the "Waiting for worker..." message will it eventually change to something else to show some sort of progress?

Syd · 2008-12-16, 22:07

Quote:

Originally Posted by mdettweiler

Just curious...how does the whole "assign number to workers" thing work? Is it some kind of system that automatically runs x amount of ECM curves on a given number (the specific amounts being as you described earlier in this thread)?

I noticed that when I tried taking an arbitrary C176 from the homogeneous Cunningham list and submitted it for ECM to low limits, it returned a message that says "Waiting for worker...". What exactly does this mean? Also, is it dynamic in some way? If I sit there and stare at the "Waiting for worker..." message will it eventually change to something else to show some sort of progress?

It runs a few curves on the given number on one of the connected maschines. Once you click the button, it writes to a table what should be done and it shows up as "Waiting for worker". When a worker is idle, it picks up the work, updates the status to "Assigned to worker x", and if successfull, updates the factors. Unfortunately there is no automatic reload yet, but you may click search again to see it.

mdettweiler · 2008-12-16, 22:59

Quote:

Originally Posted by Syd

It runs a few curves on the given number on one of the connected maschines. Once you click the button, it writes to a table what should be done and it shows up as "Waiting for worker". When a worker is idle, it picks up the work, updates the status to "Assigned to worker x", and if successfull, updates the factors. Unfortunately there is no automatic reload yet, but you may click search again to see it.

Hmm...when I click "Search" it just gives me the normal report page, with no status whatsoever on the workers. (Or maybe it just finished since I first loaded the page?

)

Edit: I just tried this again with the same C176, then immediately re-searched it after assigning it to a worker. It then reported back as "Assigned to Worker #2".

Edit2: And...presto! It's found a P22 factor!

Batalov · 2008-12-16, 23:10

Quote:

Originally Posted by mdettweiler

Hmm...when I click "Search" it just gives me the normal report page, with no status whatsoever on the workers. (Or maybe it just finished since I first loaded the page?

)

No, it so happens that I pre-ran those 12+11.x searches just an hour ago, because I had a couple recent NFS factors in that family (x=169 and 182) (and because I was testing the f(x),101,200 capability). Cool!

It is probably worth mentioning that the table homog.cunnighams are ECMed very hard already - nothing will most likely be found by ECM what is not already known from Paul's tables. Check the progression of the ECM factor sizes in the update and older sections, and you will see that at least 35-40-digit factors may be now expected. Syd, you may want to put a flag field in the database of how much ECM is known to have been done (by your workers or external efforts) on specific composites and skip the ECM to low/medium/high limits, if requested by a button press (or even grey-out/inactivate those buttons on such flagged numbers). (So that the workers would not be run uselessly again and again.)

P.S. It would be nice if we could add algebraic factors at least manually. Example 12^190+11^190
Report factor(s): ...and we enter... 12^38+11^38
(which is itself composite, but please evaluate it behind the scenes, look it up in the database and apply all found factors recursively, right? Just my 2 cents)

P.P.S. Added the p38 factor manually to the same 12+11.200...

fivemack · 2008-12-16, 23:48

Can we have fibonacci() and lucas() numbers? They have the same sort of divisor patterns as Cunningham numbers, and there are big tables of factors available for inhalation at http://home.att.net/~blair.kelly/mat.../fibonacci.txt and http://home.att.net/~blair.kelly/mat...acci/lucas.txt.

Would it be possible to have a notation - ## or something - for 'product of the first N primes' rather than 'product of the primes less than N'; otherwise x#+1,100,130 gets rather repetitive.

Batalov · 2008-12-17, 05:19

Syd, you may want to parse some algebraic forms.
Well, http://factorization.ath.cx/search.p...1670676383%5E2
Maybe, for starters, it would be nice to parse
a^n
a^n-b^n
a^odd+b^odd
and then progress to Aurife... ...ehh... Aurifeuillian! yes, I can. (I though I could. Chug, chug, chug...)

FactorEyes · 2008-12-17, 06:04

Quote:

Originally Posted by Batalov

and then progress to Aurife... ...ehh... Aurifeuillian! yes, I can. (I though I could. Chug, chug, chug...)

Well done! Your Aurifeuillian papers are in order.

10metreh · 2008-12-17, 07:51

I'm working on the numbers just above a googol. There are quite a few easy QS/SNFS numbers in the range from 10^100 to 10^100+100. I couldn't find a factor table for these, so I'm doing them.

CRGreathouse · 2008-12-17, 14:10

Quote:

Originally Posted by 10metreh

I'm working on the numbers just above a googol. There are quite a few easy QS/SNFS numbers in the range from 10^100 to 10^100+100. I couldn't find a factor table for these, so I'm doing them.

Amusingly, I worked on adding numbers in this range to the database a few days ago, building on my work in A076848.

Jeff Gilchrist · 2008-12-17, 18:21

Nice database. I did a search for 3^437-2^437 and then entered the remaining factors which I found. I did one at a time, but when I click on "Report" it just brings me back to the same page with no message about if the factor was successfully save or not. If I try to search again the new factors are not showing up.

So did this fail, or do you have some sort of process to verify the factors and they will appear at some later date?

10metreh · 2008-12-17, 18:55

Quote:

Originally Posted by CRGreathouse

Amusingly, I worked on adding numbers in this range to the database a few days ago, building on my work in A076848.

And you missed out a C72 and a couple of C74s.

2008-12-16, 23:48	#38
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 14356₈ Posts	A couple more cheeky requests Can we have fibonacci() and lucas() numbers? They have the same sort of divisor patterns as Cunningham numbers, and there are big tables of factors available for inhalation at http://home.att.net/~blair.kelly/mat.../fibonacci.txt and http://home.att.net/~blair.kelly/mat...acci/lucas.txt. Would it be possible to have a notation - ## or something - for 'product of the first N primes' rather than 'product of the primes less than N'; otherwise x#+1,100,130 gets rather repetitive.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Database for k-b-b's:	3.14159	Miscellaneous Math	325	2016-04-09 17:45
Factoring database issues	Mini-Geek	Factoring	5	2009-07-01 11:51
database.zip	HiddenWarrior	Data	1	2004-03-29 03:53
Database layout	Prime95	PrimeNet	1	2003-01-18 00:49
Is there a performance database?	Joe O	Lounge	35	2002-09-06 20:19

2008-12-16, 21:52	#34
mdettweiler A Sunny Moo Aug 2007 USA (GMT-5) 3×2,083 Posts	Just curious...how does the whole "assign number to workers" thing work? Is it some kind of system that automatically runs x amount of ECM curves on a given number (the specific amounts being as you described earlier in this thread)? I noticed that when I tried taking an arbitrary C176 from the homogeneous Cunningham list and submitted it for ECM to low limits, it returned a message that says "Waiting for worker...". What exactly does this mean? Also, is it dynamic in some way? If I sit there and stare at the "Waiting for worker..." message will it eventually change to something else to show some sort of progress?

2008-12-17, 05:19	#39
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 22157₈ Posts	Syd, you may want to parse some algebraic forms. Well, http://factorization.ath.cx/search.p...1670676383%5E2 Maybe, for starters, it would be nice to parse a^n a^n-b^n a^odd+b^odd and then progress to Aurife... ...ehh... Aurifeuillian! yes, I can. (I though I could. Chug, chug, chug...)

2008-12-17, 07:51	#41
10metreh Nov 2008 2322₁₀ Posts	I'm working on the numbers just above a googol. There are quite a few easy QS/SNFS numbers in the range from 10^100 to 10^100+100. I couldn't find a factor table for these, so I'm doing them.

2008-12-17, 18:21	#43
Jeff Gilchrist Jun 2003 Ottawa, Canada 7·167 Posts	Nice database. I did a search for 3^437-2^437 and then entered the remaining factors which I found. I did one at a time, but when I click on "Report" it just brings me back to the same page with no message about if the factor was successfully save or not. If I try to search again the new factors are not showing up. So did this fail, or do you have some sort of process to verify the factors and they will appear at some later date?