Additional root-sieving with several different stage2_norms of the original hit

[CODE]720720 -357325846797204 41053980542990764159304 9859340025664288755841215208 -516231229064345785376424349291585222 -48575959902721326204002670786995815372151 842380064171287632571034526607045112670993928819 17935131075306692003 -513064529635387021712586371244522278 -2.79 4.209315e+26[/CODE]result in a small improvement:

[CODE]# norm 7.151360e-16 alpha -12.006819 e 3.361e-16 rroots 6
skew: 46533633.31
c0: 1633953437485972808835372375174240820486209483635575
c1: -317884823360112421091161933951538547419251225
c2: -1508622122179706870108700788906748783
c3: 1588308725083082927652305632285
c4: 19349278589117372499224
c5: -300230794703124
c6: 720720
Y0: -513064529398585109953329741896064546
Y1: 17935131075306692003
[/CODE]The score is higher, but trial-sieving suggests almost identical yield.

The best I've found is
# norm 4.963167e-16 alpha -10.377733 e 2.535e-16 rroots 6
so your's is significantly better. :smile:

A BIG Challenge - C184
 

Aliquot Sequence 4788 has hit a big voluptuous composite ready for GNFS.

The last term is [URL="http://www.factordb.com/sequences.php?se=1&aq=4788&action=last"]here[/URL].

And the number is:
[CODE]8654667943202786897300764569717846447540644853988047717553660237260758755867606179363858184334795799027449419922377599045512414795305835001129150322772945060908427265269477818993677811[/CODE]

A nice poly is requested.

AS3408:i1385 - C160
 

Aliquot Sequence 3408 has completed ECM work on a C160.

The last term is [URL="http://factordb.com/sequences.php?se=1&aq=3408&action=range&fr=1385&to=1385"]here[/URL].

The number is:
[CODE]4085684988171507780203327299922394806290879077936133966700748091011780927055862909049779448089479186599092123609908892268850213732448904830398588602432892817233[/CODE]A nice polynomial requested.

Poly requested for c150
 

I've got a c150 from aliquot sequence 572000:i3135 that needs a nice poly. The number:[code]178840658547056398340661035575701801836589428489697812907837151638999043827732425417122362163295769975015657207507501412348850500995123605989554924733[/code]The top two scores that I found with some CPU searching:[code]# norm 1.781488e-014 alpha -8.131404 e 5.941e-012 rroots 5
skew: 17076867.32

# norm 1.644440e-014 alpha -7.152138 e 5.702e-012 rroots 5
skew: 2349990.71[/code]

I'm sorry RichD and schickel, I haven't visited the forum in a while....
RichD, for your C184, I suppose you want a polydegree=6?
I'll start with the C184.

[QUOTE=firejuggler;365338]I'm sorry RichD and schickel, I haven't visited the forum in a while....
RichD, for your C184, I suppose you want a polydegree=6?
I'll start with the C184.[/QUOTE]

I ran polynomial selection on a GNFS C210 with degree 5 and degree 6. Even for that large of a composite, 5 and 6 were closely matched. And actually, I ended up using a degree 5 poly for my sieving.

I think a C184 is more in degree 5 territory, but it would also be good to get some more comparisons between degree 5 and 6 to see how they perform for different sized composites.

*very early* result for the C184
expecting poly E from 3.85e-014 to > 4.42e-014
[code]
R0: -308681613820375315889646450600
R1: 993755618144567
A0: 3733330456931509623832083088632908837419
A1: -79667494484874901279813232433056475
A2: -1781530284909785888686735505082
A3: 4662437233109115905765554
A4: 15834659559757558703
A5: -8624679856399
A6: 10004280
skew 337480.79, size 1.630e-013, alpha -8.183, combined = 1.833e-014 rroots = 4
[/code]
As you can see with the skew I seem to be in the right zone.

[QUOTE=firejuggler;365352]*very early* result for the C184
expecting poly E from 3.85e-014 to > 4.42e-014

skew 337480.79, size 1.630e-013, alpha -8.183, combined = 1.833e-014 rroots = 4

As you can see with the skew I seem to be in the right zone.[/QUOTE]

Degree 6 polys produced by msieve have much lower skew than degree 5- you could search at one million and get skews not too much bigger. Not that you have to- I don't think it matters.

I'll take a shot at Schickel's 572000 C150 over the weekend, and post a poly Monday morning.

30hr of GPU searching on Schickel's 572000:3135 produced a best of 5.44e-12:
[code]skew 1216179.20, size 1.488e-014, alpha -7.096, combined = 5.445e-012 rroots = 3
R0 -41207794771228609052293393945
R1 26799623765140837
A0 -875622949540118343384186028802881896
A1 3079460511310800368252184814314
A2 -2749209751414539770426771
A3 -3137388845678318314
A4 3050080541380
A5 1505112 [/code]

The only way this poly is useful is if the CPU-search was on a older version of msieve that overestimated E-score. I don't think the error was more than 5%, but I enclose the poly in case it's helpful.

I will take a shot at RichD's C160 from 3408 in a couple of days. I should post something by Thursday.

[QUOTE=VBCurtis;365437]30hr of GPU searching on Schickel's 572000:3135 produced a best of 5.44e-12:
[code]skew 1216179.20, size 1.488e-014, alpha -7.096, combined = 5.445e-012 rroots = 3
R0 -41207794771228609052293393945
R1 26799623765140837
A0 -875622949540118343384186028802881896
A1 3079460511310800368252184814314
A2 -2749209751414539770426771
A3 -3137388845678318314
A4 3050080541380
A5 1505112 [/code]

The only way this poly is useful is if the CPU-search was on a older version of msieve that overestimated E-score. I don't think the error was more than 5%, but I enclose the poly in case it's helpful.

I will take a shot at RichD's C160 from 3408 in a couple of days. I should post something by Thursday.[/QUOTE]Thanks for the run. I'll do some testing versus my finds to see which sieves better. (I searched with v1.49....)