3LM Table discussion

[R.D. Silverman]

Originally Posted by jyb

???

http://homes.cerias.purdue.edu/~ssw/cun/3LM

Doesn't that qualify as "Sam's page"?

Indeed it is! However, there is no official extension to the existing
tables. These were provided a while back and Sam posted them, but without
any kind of announcement that they were being added to the tables.

[jyb]

Originally Posted by R.D. Silverman

Indeed it is! However, there is no official extension to the existing
tables. These were provided a while back and Sam posted them, but without
any kind of announcement that they were being added to the tables.

Indeed. I make no claims as to whether these are part of the "real" tables, nor do I have any desire to join the holy war^H^H^H^H^H^H^H^H debate over whether these numbers are worthwhile targets for our CPU time. I was just questioning what looked like a factual error.

[bdodson]

Originally Posted by jyb

???

http://homes.cerias.purdue.edu/~ssw/cun/3LM

Doesn't that qualify as "Sam's page"?

OK; I mis-typed/composed. There's nothing on Sam's page
about a _recent_ extension of the 3LM. Not to attempt to
impose a minimal attention span pre-req on forum posts; but
I was remarking on my friend Serge's post saying

Phew!
There's a full update ...

and we're discussing a collection of candidates for factorization
just brought to the forum attention, apparently by the DB. -bd

(For the purposes of discussion of The Cunningham tables, there's
a well-defined reserved term for a Cunningham Update, initiated by
Sam. I'm saying this isn't that (as far as I know); and that I'd
very much appreciate not having the thread on the 3+ Cunningham
list not littered with a bunch of non-Cunningham factors.)

[Batalov]

Apologies for unintended sensationalism.

I simply meant that this was the last local table to be cleaned up for readability (with SNFS difficulty added, minus ECM efforts); per se, it was always there in posts #2-4, since Jan 2005. I looked at it several times in dismay, but because of the last 3+ factor, decided to roll the sleeves and clean it. And indeed, there was a lot of vacuum cleaning. The semi-official updates are also housed at Sam's pages and the latest factors are fairly fresh. 'Phew' only meant what Garo meant year ago. (Same thread.)

As I did it, I realized that some numbers are amenable to individual contributions. But let me tone it down now.
For the volunteers:
1. I totally agree that a better application of PC time is NFS@Home. Many other projects, too.
2. If the real fun is DIY, then it should be all of it, from start to finish; if not from writing new software, then at least from constructing a suitable polynomial and parameters.
3. The only place the contributor's results will go is that email pileup, for now. Apparently there are no reservations, a simple sandbox. Email Sam when done.
4. No holly wars. Peace!

[bdodson]

Originally Posted by Batalov

Phew!
There's a full update to the 3LM extension tables. (See post #2-4, merged.)
Also, added a live link to the 'massaged' FactorDB ...

Ah. The update to the Garo tables (including that 3LM extension he had,
that I forgot about). Actually, since posts #2 through #4 were replaced
with the new and improved Serge post #2, I spent some time puzzling about
the old posts #5 & #6 (which are the new #3 & #4). And the live link over
on the Cunningham DB thread.

Thanks!

Strange; the original 3LM's are entirely cleared (the ones in the "main
tables"), so it's the ones 1200-1800 that are in the extension (as I learned
from the DB link). Err, that's (3^[6(n-3)])+1 for n = 201 to 300,
corresponding to 1200-1800. The next DB page goes past that; and the
next; and the next ... And it's the Garo table restricted to the
Xilman/Leyland range 1200-1800. No shortage of numbers, it's like they're
infinite or something.

-Bruce

[Mini-Geek]

Originally Posted by Mini-Geek

I might run 3,1275M. I'm PMing Batalov for details.

Done. c144=p68*p77. 3^1275+1 is now fully factored.

Sun Nov 08 21:51:15 2009  prp68 factor: 28307046970942456068307849275575041770247062269862468071803789660351
Sun Nov 08 21:51:15 2009  prp77 factor: 25514712584982993649721349633001246398829959056614327094684863451960879218951

My largest NFS, but relatively quite small: it only took 35 hours on my dual-core Athlon. (almost 40 hours counting the time I paused it)
Already in the DB. Does this need to be emailed to Prof. Wagstaff or anybody?

[bdodson]

Originally Posted by Mini-Geek

Done. c144=p68*p77. 3^1275+1 is now fully factored.

  prp68 factor *  prp77 factor

Does this need to be emailed to Prof. Wagstaff or anybody?

Yes, that would be good. There's a link on the bottom of

http://homes.cerias.purdue.edu/~ssw/cun/

Probably a reminder that it's for his file

http://homes.cerias.purdue.edu/~ssw/cun/3LMext

would help to get it recorded there. -Bruce

[bdodson]

Originally Posted by bdodson

Yes, that would be good. There's a link on the bottom of

http://homes.cerias.purdue.edu/~ssw/cun/ -Bruce

Looks like Sam's taking this prospective extension ... ugh, oops,
not that Sirius thing again. With a notable degree of attention, like. -bd