3+ table

[bdodson]

Originally Posted by Batalov

Sam updated tables on Dec 20, 2013 and extended the 3+- series (to 850, LM to 1700).

Also, Ryan finished 5,409+, the first hole and an MW number.

That's c271 = p132*p140, with

p132 =
179425223701718689835393837247166286352741072164031120636432
233591589571388940800769444701690976543908944390697369214917
439022818121

just missing the 2nd place p133*p139 --- a Very nice factorization!
Plus there's also the new 10,289+ C217 = p79*p139,

p79 =
113655879171065489962942957610764269402021102040803671930852
2808314581197380973

#3 on the MW list (and another 1st hole). -Bruce

[sashamkrt]

3,664+ is finally factorized.

prp71 factor: 46248186429319551122928458843101908180054613627508438977161805828761217
prp121 factor: 6508233273760318136665544671037105965536319145162733805204127851717423986349039046753960480689546841506786595588607953377
 
n: 300993985770367900297768479922592827053902966399363728992890560931679241415483740071518664179382524514248445568861310576708721702852364709054151383772702392157194559817853004526724261601779809
# norm 9.815848e-019 alpha -7.421464 e 1.477e-014 rroots 5
skew: 144509310.80
c0: -42028609313722551509841312951728328059300199168
c1: 572281859322535662139863122544907994952
c2: 18463134903420318879508513493866
c3: -204418133943886498195183
c4: -1237792108929116
c5: 293700
Y0: -15926882454590411717289490381392468307
Y1: 24298576183044118843
alim: 200000000
rlim: 150000000
lpbr: 32
lpba: 32
mfbr: 64
mfba: 96
rlambda: 2.6
alambda: 2.6

Sieved rat. sp-q 200-201m, 220-224m, 240-245m (CADO-NFS).
Sieved alg. sp-q 124-325m (with some holes, GGNFS).

[jasonp]

Nobody has commented on this, but congratulations on a huge GNFS job! Did you use CADO-NFS for more than the sieving? If yes then you should notify Paul Zimmermann, as this is the second largest factorization that has involved CADO-NFS.

[sashamkrt]

I used CADO-NFS just for rational sp-q sieving. It produced considerably less relations than GGNFS algebraic sp-q sieving and I could not to run CADO-NFS siever for I=16 (8Gb was not enough for 4 threads).

Sieving took 4 month and processing (LA+SQRT) 2 month.

Originally Posted by jasonp

Nobody has commented on this, but congratulations on a huge GNFS job! Did you use CADO-NFS for more than the sieving? If yes then you should notify Paul Zimmermann, as this is the second largest factorization that has involved CADO-NFS.

[bdodson]

Originally Posted by frmky

NFS@Home made quick work of 3,605+ using a quartic.

prp76 factor*prp109 factor

Ryan's just done 3,604+ C248, a first hole; looks like an ecm-pretest

71827350819005332008092067536489303561724233668918943900899337. 
p186 Propper ECM

a p62. Guess I can report this as a miss/near-miss. Two passes
in 2011, a short pass in 2012 and then 9t55 in Dec 2013, for a
total of 13.5t55 =c. 2.7t60, from 11087 curves with B1 = 600M
and 46323 curves with B1 = 400M. Sounds sufficient for a p62-test,
and a 2-out-of-3 chance. Anyone think finding a p62 is easy?
-Bruce

[R.D. Silverman]

Originally Posted by bdodson

Anyone think finding a p62 is easy?
-Bruce

I've been running ECM for nearly 30 years. I've never found one.........

[frmky]

Just to update it here, 3,766+ is factored. The log is attached.

[xilman]

Originally Posted by R.D. Silverman

I've been running ECM for nearly 30 years. I've never found one.........

Rob Hooft found a p62 factor of C(758) three days ago. It's in the 2014 top-ten list right now. Rob is pre-testing candidates for NFS@Home.

The c133 co-factor split as p63*p71 with GNFS here.

AFAIK, I've never managed any p6x with ECM and I've been running it almost as long as Bob.

Paul

[bdodson]

Originally Posted by bdodson

Ryan's just done 3,604+ C248, a first hole; looks like an ecm-pretest
...
a p62. Guess I can report this as a miss/near-miss. ... Sounds sufficient
for a p62-test, and a 2-out-of-3 chance. Anyone think finding a p62 is easy?
-Bruce

Nevermind, I'll settle for a p64. Another one from Bob's 2LM list, no less,
the 3rd hole. -BD

Just popped in this morning, from one of the i7's running 8 condor jobs.

[jasonp]

6319  3, 697+ c221 2430596119059914710337969915407030131984125918638366960356417238051. c154 NFS@Home gnfs
6320  3, 697+ c154 7294369040882787305640699831446454559916572630347346447156318424089217886547. p78 NFS@Home gnfs

This is the largest GNFS postprocess I know of using Msieve, and it pushed the filtering to its limits. The original filtering run failed due to the large dataset bug, but unlike all the other times this happened, a patch to the source got filtering to converge to a 93.8M matrix with weight 13830676419 (about 112 per column in the sparse part). The linear algebra finished in abut 85 hours on 2048 Teragrid cores.

Unfortunately 17 out of the 20 quadratic characters appeared in relations, so that there were only 18 dependencies found, and 16 of them failed to converge in the Newton step. Fortunately the two that did converge were enough to complete the factorization.

[R.D. Silverman]

Originally Posted by jasonp

[code]


Unfortunately 17 out of the 20 quadratic characters appeared in relations, so that there were only 18 dependencies found, .

This is a design deficiency. The quad. chars should be selected larger than the LP bound.