Factoring database

[Syd]

Originally Posted by henryzz

would it be possible to have the usage table above the list of factor tables

Sure it is

[J.F.]

I don't think it is wise to have those 'is definitely prime' buttons out in the open. I'm afraid this is somehow gonna lead to pollution (unintentional, spiders, intentional) of your fine database, which would be a shame imo.

Perhaps you can mark primes from 'trusted' sources upon import? An alternative would be to add some sort of user registration and input history, so errors can be tracked. Other ideas?

About the Mersenne factorizations: did you by any chance forget to input lowM.txt? I had a look at M5000 - it contained all factors from M2500 (from factoredM.txt), but missed the two smallest ones from lowM.txt. After having inputted them, the digit count of the remaining cofactor is still off btw.

[Syd]

User registration together with input history, login and so on is a quite complex task, but eventually I can even use it together with some reservation/ECM efforts tool.

But maybe .. how complex is it to prove a number definitely prime, say up to 1000 digits with ECPP? Are any tools available?

Indeed, i forgot to import the factors from lowM.txt. In case of M5000, can you tell me which factors are missing?

[jasonp]

For primality proving, the tool of choice is Primo, though I don't know how amenable it is to batch input.

[mdettweiler]

Originally Posted by jasonp

For primality proving, the tool of choice is Primo, though I don't know how amenable it is to batch input.

Yesterday I went through a lot of the PRP's in the database and manually proved them prime with PARI/GP's isprime() function--it was able to do many of them instantaneously, and only the ones >100 digits or so took even just a few seconds apiece to do.

At any rate, from what I can tell, PARI would be much more scriptable, and it probably wouldn't be too hard to set up into the existing worker system so that primality-proving jobs could be queued up and fed to all the workers like is already done with ECM and TF work.

[kar_bon]

i just want to reach the link in post #7 but at work it's blocked by our webfilter:

URL: http://factorization.ath.cx/search.php

URL-Kategorien: Malware

any chance to change this?

try here: http://filterdb.iss.net/urlcheck/

[J.F.]

Originally Posted by Syd

In case of M5000, can you tell me which factors are missing?

Since M2500 is fully factored, they must originate from 2^2500+1, which currently shows a 665 digit composite (after I hit ECM-medium to remove a few digits).
Edgingtons data suggests that the composite can be reduced to 580 digits. I don't know where to look for the remaining 85 digits of prime(s) - the Cunningham tables only go to 2400.

[edit]
I had a look at M2398. It showed a 530-digit composite (id=21276431). However, several implied factors from M767 were missing. After having 'submitted' them, the digit count agrees with Edgington.
I suggest doing something like

   forall factors f of Mx 
      submit f to M(kx) with k=2,3,...

[Syd]

Originally Posted by mdettweiler

At any rate, from what I can tell, PARI would be much more scriptable, and it probably wouldn't be too hard to set up into the existing worker system so that primality-proving jobs could be queued up and fed to all the workers like is already done with ECM and TF work.

Thats an awesome library, I'm done with just a few lines of extra code like you said. Though it would be nice to have those primo primality certificates for download.

any chance to change this?

try here: http://filterdb.iss.net/urlcheck/

Thats weired

[10metreh]

3 x 5^5 x 11 x 17 x 31 x 41 x 101 x 251 x 401 x 601 x 1801 x 4001 x 4051 x 7001 
x 8101 x 28001 x 61681 x 96001 x 160001 x 268501 x 340801 x 1074001 x 2020001 x 2787601 
x 3775501 x 22624001 x 37177501 x 107580001 x 229668251 x 831172501 x 2259150001 
x 3173389601 x 14321535001 x 269089806001 x 1481124532001 x 619324035675566251 
x 1277297679372570001 x 4710883168879506001 x 47970133603445383501 
x 94291866932171243501 x 5519485418336288303251 x 232983411264923603130001 
x 573759820507018639639785001 x 983262166270913562877993411320001 
x 633746672116596658276533517443227501 
x 18152902839291497575027462639977160832701118299213751 
x 246053469753590746981511859818675718355368494592178751 
x 23072017280360209143853190404178284560740949254655131346786858037350918591251 x 
8877945148742945001146041439025147034098690503591013177336356694416517527310181938001 x 
167151693059203651258538223678521800139589649629368247829683865124002155584217612884019447563002501 x 
45712845832064476909251288412537780542141386629649967711230003222863831555000707453689906421700439212068044560931077031315106846707799052501 x 
2952514364581316337639834904034501808253955768505847014917395820453653708312223506491002496302532891440120688361870627195018283686594739576042156463430228261599130139157586670190936850324484407773053651741800071409643774842685094764505897386290434816137207257527045648890038980443489148266586061742881796117450662972816739833595245680317687294799762715910575282701475446274531684090484014335382380426254949454811556963459149908837253490252431670151904810696439528202733056843409785493492110147847556186485233154284233119314768888487458979069953852588599421726292742704836614070001(the C580)

are the known prime factors of M5000.

[J.F.]

Error in my previous post: I believe M2398 should be M2301. Doesn't matter much.

@10metreh: aha, may I ask where did you come across that 85 digit prime?

[Batalov]

2,500+ | 2,2500+
2,100+ | 2,500+
2,500+/2,100+ factors easily -> p85 comes from there.

p140 and p99 come from 1250L,M of course

1250L (10M,50L,250M) 37177501.831172501.633746672116596658276533517443227501.P99
    M (2,10L,50M,250L) 5*.14321535001.P140

I bet it was done years ago.