Team sieve #22: c166 from 3270:620

[fivemack]

Sequence 3270 is the shortest sequence with a start below 10000, having grown with a long period of 2^3*3*5; it's now at 2^3*3*7.

Iteration 620 is a C166: I spent about eight CPU-weeks finding the polynomial. We'd need to sieve from 15M to 60M using the 15e siever with this poly:

n: 1773271754188764469178068334154046464444957274952289279955825326222321328349114330973185907243392930104294678498232714048124362163786378662540085144893974571128845721
skew: 50908204.88
Y0: -144767682063889933793321068524324
Y1:  538109518954680569
c0:  2825944577227672043009611776143415421051145
c1: -96204614071321310022916080146763798
c2: -975710455700385174343979023
c3:  1385331590129139306600
c4: -3062765784988
c5:  27888
lpbr: 30
lpba: 30
mfbr: 60
mfba: 87
rlambda: 2.6
alambda: 3.6
alim: 40000000
rlim: 40000000

Sieve from Q=15 million to Q=60 million, with the 15e siever. Anyone interested?

Reservations:

 15.0M- 15.5M   Andi47 (done, 1680529)
 15.5M- 16.0M   richs (15.5-15.9 done)
 16.0M- 16.5M   Andi47 (done, 1672600)
 16.5M- 24.0M   FactorEyes (done, 24963783 relations)
 24.0M- 24.5M   Andi47 (done)
 25.0M- 30.0M   fivemack (done, 16211446 relations)
 30.0M- 32.0M   fivemack (done, 6366426 relations)
 32.0M- 40.0M   FactorEyes (done, 24922498 relations)
 40.0M- 41.0M   schickel (40.0-40.5 done)
 41.0M- 46.0M   bsquared (done, 14847229 relations)
 46.0M- 50.0M   bsquared (done, 11517802 relations)
 50.0M- 54.0M   bsquared (done, 11130575 relations)
 54.0M- 56.0M   bsquared (done, 5461598 relations)

Post-processing will be done by fivemack

Upload relations via FTP to ftp://ssh.fivemack.org and put them in the 3270.620 directory.

Relations so far: about 102 million. At least 75% done.

As of 29 Jan 2011: please, no more large reservations.
2 Feb 2011: Post-processing has begun.

[Andi47]

reserving 15M-15.5M

Edit2: I guess the -M 1 flag would be useful for this job too?

[schickel]

Originally Posted by fivemack

Sequence 3270 is the shortest sequence with a start below 10000, having grown with a long period of 2^3*3*5; it's now at 2^3*3*7.

Iteration 620 is a C166: I spent about eight CPU-weeks finding the polynomial.

I assume you feel enough ECM has been run?

And, yes, we have been talking about more joint forum projects, so maybe this could be the first one....

Actually, it looks like it's 2^5 * 3 * 7, which isn't any better than 2^3 * 3 * 5....

[fivemack]

ECM: I've done 4000 @ 3e7 and 10000 @ 1e7 - two weekends on 16 threads on the Mac Pro at the office. If anyone wants to run a bit more they're welcome.

ftp directory is 3270.620 on ssh.fivemack.org

-M 1 is indeed useful for this job

I will do the post-processing

[schickel]

Originally Posted by schickel

I assume you feel enough ECM has been run?

Originally Posted by fivemack

ECM: I've done 4000 @ 3e7 and 10000 @ 1e7 - two weekends on 16 threads on the Mac Pro at the office. If anyone wants to run a bit more they're welcome.

I wish I had that kind of ECM power!

[Andi47]

I think this thread needs to become sticky'd.

[debrouxl]

15e siever and short range of q values, so RSALS can't lend a hand in this team sieving

[EdH]

I thought I'd try to lend some effort, but my machines are either too limited or I'm doing something wrong. I get the following with the above poly:

$ gnfs-lasieve4I15e -a 3270:620.poly -o 3270:620.rels -f 15500000 -c 500000
 Warning:  lowering FB_bound to 15499999.
Segmentation fault
$

Same result on two Ubuntu machines. Is 512Mb possibly insufficient?

[Andi47]

Originally Posted by EdH

I thought I'd try to lend some effort, but my machines are either too limited or I'm doing something wrong. I get the following with the above poly:

$ gnfs-lasieve4I15e -a 3270:620.poly -o 3270:620.rels -f 15500000 -c 500000
 Warning:  lowering FB_bound to 15499999.
Segmentation fault
$

Same result on two Ubuntu machines. Is 512Mb possibly insufficient?

on my ubuntu machine it takes ~223 MB.

[Andi47]

feeling a bit lonely here.

At the current rate (I'm sieving ~80 kQ/day) we will be ready around summer 2012...

[schickel]

Originally Posted by Andi47

feeling a bit lonely here.

At the current rate (I'm sieving ~80 kQ/day) we will be ready around summer 2012...

I know how you feel. I'm still working on my c155. After that finishes, I can lend some cycles.

I'm also hoping that the New Year brings some nice deals on quad-cores!