Factoring 11499*1150^1150-1

[lorgix]

Does anyone know of a factor of 11499*1150^1150-1?

And no, there really isn't anything special about this number. I just got stuck on it for some reason.

And I'm posting here now hoping that someone else cares to throw a few curves at it or such. The expected size of the smallest factor is approaching 30digits right now. I'm running curves with B1=250k B2=100*B1 right now. Those limits and below should be exhausted in, well several hours.

Should anyone be interested I'll post all details.

[Mini-Geek]

Originally Posted by lorgix

I'm running curves with B1=250k B2=100*B1 right now.

Unless you are severely limited by memory usage or something, your search would be more efficient if you used GMP-ECM and its automatic B2 bound selection, instead of B2=100*B1.

[lorgix]

Yeah, I know the 100 is low.

I don't really know how to use GMP-ECM... I'm working on it.

[bsquared]

Originally Posted by lorgix

Yeah, I know the 100 is low.

I don't really know how to use GMP-ECM... I'm working on it.

What program are you using, if not GMP-ECM?

[Batalov]

Prime95, most likely?

[lorgix]

You are correct sir!

Prime95 is the only program I know well that is suited for this.

For some reason I've started factoring numbers of the form k*1150^1150+-1, where 10499<k<11501. 70 to go...

What limits should I be using?


(11484*1150^1150+1)/409607532333161 is PRP to seven bases btw.

[Mini-Geek]

Originally Posted by lorgix

You are correct sir!

Prime95 is the only program I know well that is suited for this.

If you don't already have it, download GMP-ECM from http://gilchrist.ca/jeff/factoring/index.html. It's not too hard to use, run "ecm -h" to get the help. An example for this number is:

ecm -c 430 -one 25e4 <input.txt >>ecm-results.txt

where input.txt is a file containing "11499*1150^1150-1" (without quotes). -c 430 tells it to run up to 430 curves, -one tells it to stop after finding a factor, 25e4 (250e3 or 250000 would be the same result) tells it the B1, and it automatically chooses a B2 of 183032866. Then <input.txt feeds it input.txt which contains the number and it appends its output (you can use > to overwrite any existing file or >> to append to any existing file) to ecm-results.txt.
Or just:

ecm -c 430 -one 25e4

and then type or paste in 11499*1150^1150-1 and hit enter. But make sure you do this in a way that you'll still be able to see the results when it's done (e.g. type that from a command window, don't run it with a shortcut).

Originally Posted by lorgix

What limits should I be using?

Limits that make the time for stage 1 equal to the time for stage 2 (as long as they're both run on the same hardware - otherwise it gets more complicated). I know that GMP-ECM will automatically set the B2 so the times should be equal, but I don't think Prime95 does - you'd have to tune it for that yourself.

[EdH]

Originally Posted by Mini-Geek

Limits that make the time for stage 1 equal to the time for stage 2 (as long as they're both run on the same hardware - otherwise it gets more complicated). I know that GMP-ECM will automatically set the B2 so the times should be equal, but I don't think Prime95 does - you'd have to tune it for that yourself.

Hmm.. one of my linux machines is about even: 4915291ms/4711399ms. But, two others aren't that close: 1736741ms/690039ms and 3559730ms/1981436ms.

I wonder if it has to do with memory limitations...

[Andi47]

Originally Posted by lorgix

(11484*1150^1150+1)/409607532333161 is PRP to seven bases btw.

Primo says that it's prime. (prime proving took ~95 h)

link to Primo certificate for (11484*1150^1150+1)/409607532333161, 1.24 MB

[lorgix]

Originally Posted by Andi47

Primo says that it's prime. (prime proving took ~95 h)

link to Primo certificate for (11484*1150^1150+1)/409607532333161, 1.24 MB

Nice!

I actually started that, but stopped very soon after. I realized it would have taken forever.