73 digit ECM factor

[alpertron]

Congratulations!!!

So in this year it was found the first known prime factor of F14 and a 7x-digit factor by ECM (demonstrating that the Playstation consoles are very useful for this task), which are very important discoveries in Computational Number Theory.

[bdodson]

Originally Posted by FactorEyes

...
All the p68 through p72 factors I have found are now ECM misses.

Mine too; retroactive to 1998. -bd ("Take a breath, already!")

Ah, and Congratulations on carrying through a sustained effort,
including lots of new stuff we hadn't even imagined before.

Postscript: Ooops. Hold the presses. It's only the Mersenne numbers
for which p68-p72 were misses. Perhaps we've displaced the yoyo
p68 record too quickly, as it still holds for non-Mersenne numbers?

[R.D. Silverman]

Originally Posted by bdodson

Mine too; retroactive to 1998. -bd ("Take a breath, already!")

Ah, and Congratulations on carrying through a sustained effort,
including lots of new stuff we hadn't even imagined before.

Postscript: Ooops. Hold the presses. It's only the Mersenne numbers
for which p68-p72 were misses. Perhaps we've displaced the yoyo
p68 record too quickly, as it still holds for non-Mersenne numbers?

It would be nice to know whether any other numbers were attempted
with the same level of effort. It would be an extraordinary surprise if they
succeeded with the only number they tried.

BTW, the yoyo 68-digit result does not seem to be listed on the
top-10 page.

[jyb]

Originally Posted by R.D. Silverman

BTW, the yoyo 68-digit result does not seem to be listed on the
top-10 page.

That's because it took place in 2009 (barely). It's on the 2009 top-10 page.

[bsquared]

Originally Posted by R.D. Silverman

It would be nice to know whether any other numbers were attempted
with the same level of effort. It would be an extraordinary surprise if they
succeeded with the only number they tried.

The 63 digit find from 2^1187 - 1 looks to have received similar B1 treatment, at least. It would be interesting to know how many were tried which revealed no factors.

[ixfd64]

Originally Posted by bdodson

PS -- I've posted a link to Arjen's Gif of the ps3s over in the 2- subthread at
http://www.mersenneforum.org/showthr...425#post207425

That's really impressive. It's nice to see the PS3 being put to good use!

[yoyo]

Hello,
congratulation to this big ecm factor!!!
Is the PS3 version of ecm somwhere available and described?
Is a general usable version which can be included into Boinc ;)
yoyo

[R.D. Silverman]

Originally Posted by yoyo

Hello,
congratulation to this big ecm factor!!!
Is the PS3 version of ecm somwhere available and described?
Is a general usable version which can be included into Boinc ;)
yoyo

Why should there be?

There seems to be an (unreasonable IMO) attitude among many people
that the achievements of a research group should be made available
to the general public.

Why should epfl give away the results of their intellectual efforts?

The source for GMP-ECM is available. If you want to run it on machines
of your choice, then do it yourself.

[unconnected]

I wonder how much memory they used for stage 2. Trying to repeat 'lucky' sigma run I've received this :surprised

Step 1 took 70927813ms
Estimated memory usage: 14G
GNU MP: Cannot allocate memory (size=256)

[bdodson]

Originally Posted by unconnected

I wonder how much memory they used for stage 2. Trying to repeat 'lucky' sigma run I've received this :surprised

I was wondering too, thanks. But I'm not surprized. -Bruce

[Brian Gladman]

Originally Posted by R.D. Silverman

Why should there be?

There seems to be an (unreasonable IMO) attitude among many people
that the achievements of a research group should be made available
to the general public.

This depends on who funded the work.

I don't know a great deal about EPFL but a lot of academic institutions in Europe obtain a significant proportion of their funding from taxpayers either through national governments or through the European Commission.

If the work was partly or wholly funded by taxpayers (and I am not saying that it was), then there might well be a good case for their efforts being made openly available to those who paid for the work.

Of course taxpayers in one country cannot expect to benefit from work done in other countries. But there seems to be a widespreaad recognition that the benefits of international sharing will often be sufficient to allow taxpayer funded work in one country to be openly exploited internationally for the collective benefit of us all.

Brian Gladman