[QUOTE=Batalov;295357]11e6 is done. I have 800 of 43e6s, and we'll need some more and a poly and the circle spins again.[/QUOTE]True, but the good thing righ tnow is the downdriver is still alive.

And if you look at the slope, it looks nice and steep still.....in fact, to get much steeper, we would need a couple of lines to factor as 2 * [TEX]p[/TEX], but I shudder to think of what could happen then.....

[QUOTE=schickel;295363]True, but the good thing righ tnow is the downdriver is still alive.

And if you look at the slope, it looks nice and steep still.....in fact, to get much steeper, we would need a couple of lines to factor as 2 * [TEX]p[/TEX], but I shudder to think of what could happen then.....[/QUOTE]

The only thing which I would prefer over 2^1*3^0*... would be 2^0*3^0*... :wink:

would that be bad? we would haev to find another sequence....

[QUOTE=firejuggler;295389]would that be bad? we would haev to find another sequence....[/QUOTE]That [I]would[/I] mean finding another sequence, but it would also be good, since the 2 disapearing means that the parity has flipped and there is probably a quick (very) termination in the offing.

That would set the bar almost unbearably high, though, in terms of a record to break. The highest sequence terminated so far was [URL="http://factordb.com/aliquot.php?type=1&aq=921232&big=1"]921232[/URL], with a max height of 127 digits....very nice extreme ski slope-style finish to that one.

I see. You would prefer that it drop to around 1M before going up again?

I'm running poly selection on line 2968.

I have a raw poly with 8.079e-12:
[CODE]# norm 3.040993e-14 alpha -8.695524 e 8.079e-12 rroots 3
skew: 17037119.78
c0: -233604520796377704517742605390238324160
c1: -100608549191218862407256856948168
c2: -28411607816936368074938894
c3: -3499091303190109585
c4: 167235429052
c5: 756
Y0: -82569070740492951856116142199
Y1: 2528842373881907
[/CODE]

Here's another poly with Murphy=8.702e-12:
[code][color=blue]# sieve with ggnfs lasieve4 I14e on alg side from Q=7M to 26.5M[/color]
# est ~29M raw relations (avg. 0.073 sec/rel C2D @ 3.4GHz)
# aq4788:2968
n: 2901427496531126932639137570733764567154363493488725447618931456921289899532150462227507091080986296697887849811911080911699389211293401117366922879
# norm 3.308923e-14 alpha -7.684938 e 8.702e-12 rroots 5
skew: 1747235.30
c0: 284458236924058524332508604536609825
c1: 3568209986038099775080055048948
c2: 4155153429785771566461909
c3: -2518588749056401782
c4: -1043665541956
c5: 298200
Y0: -24981650660011224403799863808
Y1: 25826128680664607
rlim: 14000000
alim: 14000000
lpbr: 28
lpba: 28
mfbr: 56
mfba: 56
rlambda: 2.5
alambda: 2.5
[/code]

It's good to queue at RSALS, I guess.
I have 5000 43e6 curves and 120 110e6s.

Oops.
[CODE]Input number is 2901427496531126932639137570733764567154363493488725447618931456921289899532150462227507091080986296697887849811911080911699389211293401117366922879 (148 digits)
Using B1=110000000, B2=776278396540, polynomial Dickson(30), sigma=3026907363
Step 1 took 382668ms
Step 2 took 122660ms
********** Factor found in step 2: 2687210366475282566041791827478627115964912605604341
Found probable prime factor of 52 digits: 2687210366475282566041791827478627115964912605604341
Probable prime cofactor has 97 digits
[/CODE]

Looks like a c138 in i2976; will do all 11e6s. After that need a poly and a healthy volunteer.