Holy <censored>! Wow!!!!!!!!!

[R.D. Silverman]

In case anyone hasn't noticed, Bruce Dodson found a 66 digit factor
of 3,466+ with ECM.

I've been wondering when the first 60 digit factor would be found, but
a jump from 59 to 66 digits just blows one away........

Nicely done, Bruce!

[akruppa]

I've seen Bruce's announcement but I couldn't believe my eyes. Now that PaulZ added the factor to the ECM records page and you wrote this posting it's gradually sinking in... this is nothing short of a miracle. I hardly know what to say.


Except CONGRATULATIONS, BRUCE!!!!



Alex

[dave_dm]

Whooooa! Nice one Bruce :)

*recreates RSA modulus to be 20 digits larger*

Dave

[Mystwalker]

Many congratulations for Bruce from my side, too!

But I think we should also compliment the developers of gmp-ecm. The new factors have been found with high bounds (11e7, 26e7), which wouldn't have been sensible without their work.

[akruppa]

Actually, the group order is incredibly smooth:
2^2, 3, 11243, 336181, 844957, 1866679, 6062029, 7600843, 8046121, 8154571, 13153633, 249436823

B1=14M, B2=100*B1 would have found this factor! So the FFT stage 2 with the high B2 values it allows was not the decisive element in this discovery, but rather unbelievable luck (and great persistance!) on Bruce's part.

Alex

[Vijay]

Great going Bruce!

[rogue]

What I find interesting is "[powered by GMP 4.1.90]". The current release is 4.1.4, so where did this release come from?

[Mystwalker]

Originally Posted by akruppa

B1=14M, B2=100*B1 would have found this factor! So the FFT stage 2 with the high B2 values it allows was not the decisive element in this discovery, but rather unbelievable luck (and great persistance!) on Bruce's part.

I'd rather say that we was pretty unlucky - after all, no 43M-curve got this one (ok, likeliness should be ~exp(-1) ) and the second biggest factor is just a bit too large for B1=11M...

Technically, the gmp-ecm improvements weren't that important, but I'm guess that psycologically, they convinced Bruce to do the 55-digit step.
I think it's unnecessary to do a big discussion about who's to gratulate for these records (so just give up your modesty and admit your responsible, too ).

Originally Posted by rogue

What I find interesting is "[powered by GMP 4.1.90]". The current release is 4.1.4, so where did this release come from?

Seems to be a closed-beta version of GMP 4.2 - somewhere (unfortunately, I can't remember), I've read that the release is due soon (whatever soon means). After all, Bruce was also able to test the beta of gmp-ecm6. Interesting that everytime he used a beta, he found a record factor. Coincidence? Or advertisement?

[BotXXX]

Originally Posted by Mystwalker

Seems to be a closed-beta version of GMP 4.2 - somewhere (unfortunately, I can't remember), I've read that the release is due soon (whatever soon means).

accourding to this page http://www.swox.com/gmp/ it was first planned to be releases november 2004 as how i remember, than it became december 2004 follewd by januar 2005 and now it is early 2005. But well i am curious about it.

It comes with an improved configuration system, and several functional improvements and some speed improvements.

Altho the promised 5.0 seems very very promising but also far far away.

[Mystwalker]

Right, that information is available as well. To be honest, I didn't considered them as reliable for an approx. release date, as I've seen those deadline setbacks as well.
I think it was in one of the mailing lists (probably devel or discuss). Then it has to be in march, as I searched in all april entries...

[tcadigan]

one hell of a factor! great job!