Homogeneous Cunninghams now on ECMNET server
 

I have just put the remaining homogeneous Cunningham composites onto the public ECMNET server hosted by Phil Carmody.

This will provide a particularly easy way for people to contribute to factoring these numbers. All you need is the ECMNET client software, available from [url]http://home.wi.rr.com/mrodenkirch/ecmnet.zip[/url]

Then configure it to connect to port 8194 at the address 83.143.57.194

That's all there is to it!


Paul

Paul,

[B]p36 factor of 3-2_405 found by ECMNet!

[/B] A factor was found for 3-2_405 using GMP-ECM using factor method ECM
Candidate number: 14098162795245974515868320048016550988946628667369605233553276704617344008486335834548126502681469731
Factor: 205745243247457752308204807962101451
Factor Type: probable
Co-Factor: 68522424007098766742139873706964338295253467623919917733091062281
Co-Factor Type: Probable
B1: 250000
Sigma: 2576639739
Finder: [U]my email address[/U]
Found on machine: home

Paul,

[B][code] p33 factor of 3+2_410 found by ECMNet![/B]

A factor was found for 3+2_410 using GMP-ECM using factor method ECM
Candidate number: 12433321127785491846628430233459615735609499579461790244070227448919483900865014214784345674797245082929575054061
Factor: 815812353835536441048826708110181
Factor Type: probable
Co-Factor: 15240417810946737881490307111673540332631873961550698948605702075629004316597481
Co-Factor Type: Probable
B1: 250000
Sigma: 3363607725
Finder:
Found on machine: home [/code]

[QUOTE=fivemack;98877]The sieving for 5^267+3^267 completed overnight, as you'd expect; twelve dependencies all failed to produce a factorisation other than 1*N. I tried a bit more sieving and rebuilding the matrix, but the sqrt step still failed.

It turns out that 9*x^4 + 15*x^2 + 25 is one of those polynomials that is irreducible over Z but splits modulo every prime. I suspect this is problematic for the way that ggnfs uses to compute square roots :sad:

Any more ideas?[/QUOTE]

I thought I'd try it. Not looking at Bob's derivation above and instead doing it myself from scratch, I ended up using the polynomials 25 x^4 - 15 x^2 + 9 and 3^44 x - 5^44. In the sqrts, I also ended up with "Warning: getSqrtP() - f appears to split modulo all primes!" for most of the dependencies. But on the 16th one, it worked:

r1=6723789373525015879256142001890137471015973961 (pp46)
r2=1432418075124920788188560418927247123363989607131805407099380426386351 (pp70)

Total time was 3.64 hours.

Greg

[QUOTE=em99010pepe;98974]Paul,

[B][code] p33 factor of 3+2_410 found by ECMNet![/B]

A factor was found for 3+2_410 using GMP-ECM using factor method ECM
Candidate number: 12433321127785491846628430233459615735609499579461790244070227448919483900865014214784345674797245082929575054061
Factor: 815812353835536441048826708110181
Factor Type: probable
Co-Factor: 15240417810946737881490307111673540332631873961550698948605702075629004316597481
Co-Factor Type: Probable
B1: 250000
Sigma: 3363607725
Finder:
Found on machine: home [/code][/QUOTE]Yup, that one and its predecessor arrived by email here. That's one reason why I like the ECMNET server --- not only is it easy to use, I get emailed with the results without anyone having to remember to send them to me.

Keep up the good work. I'm sure there must be other accessible factors to be found.


Paul

[quote=xilman;98993]
Keep up the good work. I'm sure there must be other accessible factors to be found.

Paul[/quote]

I found a way to hide ecmclient/ecm so I added more two cores. Tomorrow I will add another 3.
Paul, just keep the server up and with work, ok?

Cheers,

Carlos

[QUOTE=em99010pepe;98997]Paul, just keep the server up and with work, ok?[/QUOTE]
That's the intention.
Another factor turned up already today.

Paul

[quote=xilman;99001]That's the intention.
Another factor turned up already today.

Paul[/quote]

It's wasn't me...
Thanks Paul.

Another one.

[code]
[B]p33 factor of 5-3_259 found by ECMNet![/B]
A factor was found for 5-3_259 using GMP-ECM using factor method ECM
Candidate number: 3907615342666299633712413326021648111766908740440423139975951596570070575764202665102674790734946182588181404266610577391283495872761352760461312970561
Factor: 117871565870533860403003953129247
Factor Type: probable
Co-Factor: 33151467139736585040965497130232562404065279070953626454461842001407082112230628905430048823759628386442544117853724063
Co-Factor Type: Composite
B1: 250000
Sigma: 3090436740
Finder:
Found on machine: home2


[/code]

Progress, and a request
 

So far, eight new factors have arrived at the server in its first 30 hours! That rate will fall off, of course, once the easy ones have been discovered.

I'd like to request that people connecting to the server try not to connect too often.

As more and more machines join in, the aggregate network load of all the connections can get quite high. I'm not sure what the default setting in the ecmclient.cfg file is these days, but I'd suggest that a connection every couple of hours is easily sufficient. The server is hosted on Phil Carmody's system and I don't want to abuse his generosity.

Anyone likely to add a fair-sized farm to the server's load, lets say more than a dozen or two machines, would be well advised to run a slave server at their site. It's easy to do, I'm more than willing to assist with setting it up, and a server can be a client of itself simultaneously. Running a slave server means that all the clients' communications are handled locally and only the aggregrate work done is communicated to the master server. Any factors found are communicated immediately they're found by a client so they do not have to wait for a scheduled client/server update. Likewise, as soon as a slave server is told of a factor by one of its clients it passes on the good news to its master server.

Reminder: the server is on port 8194 at 83.143.57.194


Paul

This may have already been factored by ECM (though I doubt it from the size of the factor), but here's another...

3,2,411-
1427745802482020164606532613800871537542769481699 (pp49)
804729923600337202368730124917351971300382419078736533207664505561844863010907 (pp78)

Greg