Polynomial Request Thread

[VBCurtis]

3366_c196 best hit so far:

# norm 2.827294e-19 alpha -9.156817 e 6.881e-15 rroots 3
skew: 398246220.95
c0: 2161252598131584931988859552873325658428988114525
c1: 281777528104917614065145642986582027111335
c2: 2130193159315298801682054284702939
c3: -7127594811265851259919835
c4: 1546196148940544
c5: 10674300
Y0: -60863927839642347497925715592933457498
Y1: 37222510967963734193

I am finished 10-11M, continuing to 13M at least.

[fivemack]

Would anyone be willing to have a look at L4185A

n=8987806720543066824828777771684810837983402225562252846090077334280866398775429326543006554261754651886020042524703087822608753762793828289184008817162922382342453664386510721

It's clearly a GNFS number: a few hours of polynomial search got me 1.264e-13, which sieves significantly faster than the SNFS quartic; I forgot to stop the search and 20 hours got me

# norm 4.358323e-17 alpha -8.109162 e 1.520e-13 rroots 5
skew: 737496720.59
c0: 37366868075622921872887905169118872545735310707
c1: 33160541479075124138233527294452256766
c2: -1199812783065814588455759317492
c3: 1061028159061583553318
c4: 2514291650721
c5: 252
Y0: -32392959958129008533217117097130528
Y1: 96204575606225003

It's only a G175, so probably not worth more than 1000 core-hours to search, but I'm having difficulty running stage-1 poly search on my GPU at the moment and my CPUs are all busy. Obviously c5=252 is much too small and the skew is ugly, c5=1e6 might be a better place to hunt.

[RichD]

Originally Posted by fivemack

Would anyone be willing to have a look at L4185A

I'll start at 2M later today.

[RichD]

A couple through 2.5M. Maybe one can spin up. Continuing on toward 3M.

N: 8987806720543066824828777771684810837983402225562252846090077334280866398775429326543006554261754651886020042524703087822608753762793828289184008817162922382342453664386510721
# expecting poly E from 1.55e-13 to > 1.79e-13
R0: -5325542471148036532434643283235535
R1: 100473047717951539
A0: -7916774677545967411671056893609861285536
A1: 769746318456414744784003749018336
A2: 7215068676122915953901798804
A3: 110556152983262454371
A4: -408694036692540
A5: 2098140
skew 4602021.44, size 3.563e-17, alpha -6.226, combined = 1.527e-13 rroots = 5

R0: -5240997044352583828070231225140760
R1: 36365175670305559
A0: 221710994209052690517607471246145695157919
A1: 131595954253210535413792077523106409
A2: -88551348262343773054754374317
A3: 1082923749056012587423
A4: 350124581438126
A5: 2272920
skew 15436462.04, size 3.297e-17, alpha -7.159, combined = 1.439e-13 rroots = 5

[swellman]

Originally Posted by fivemack

Would anyone be willing to have a look at L4185A

n=8987806720543066824828777771684810837983402225562252846090077334280866398775429326543006554261754651886020042524703087822608753762793828289184008817162922382342453664386510721

I’ll search up to 2M.

[RichD]

Completed 2-3M.

N: 8987806720543066824828777771684810837983402225562252846090077334280866398775429326543006554261754651886020042524703087822608753762793828289184008817162922382342453664386510721
# expecting poly E from 1.55e-13 to > 1.79e-13
R0: -4971451444024561936083025217421892
R1: 1631264507585309
A0: -151858704644213128223140310098311770212637535
A1: 17587855350828494949364225729237510777
A2: -143736696319430339084842442675
A3: -16576413512483969381459
A4: 31376809898760
A5: 2959632
skew 46627236.03, size 3.444e-17, alpha -8.169, combined = 1.446e-13 rroots = 5

[RichD]

Now through 4.75M, not even a 6-handle in the latest run.

[Gimarel]

I have 3 different stage 1 hits, that produced a poly above 7e-17. The best polys for each stage 1 hit are:

# norm 3.768119e-19 alpha -7.174415 e 8.391e-15 rroots 5
skew: 344692012.75
c0: -564202157998984142887784858415029503603332553440
c1: 20471710419929056899196323634945502744172
c2: 1024852198685693258772746031309
c3: -366944012623409011230680
c4: -46283661662182
c5: 1277640
Y0: -93056380628397254230674038194182822521
Y1: 331483736214488911093

# norm 3.181495e-19 alpha -8.377710 e 7.461e-15 rroots 5
skew: 731768967.92
c0: -62709702069588224688577590096653381283289981294605
c1: 368850951620871891923315385346929289276297
c2: 302418384162966086044748746056779
c3: -2101912871375171293290753
c4: -1088545645305366
c5: 1008000
Y0: -97574368853537242755374104793223667514
Y1: 305123309227135907611

# norm 2.919589e-19 alpha -8.077724 e 7.013e-15 rroots 5
skew: 386451790.45
c0: 2641669456416494628584290173365361900681988003040
c1: 30302764318445907769015790745306303265656
c2: -520657525067601346807255624908998
c3: -2222778098581037087262619
c4: 3865273233695062
c5: 1015200
Y0: -97435571086275010093350427416948007363
Y1: 37574732919972977597

[swellman]

Wow! Your first/best poly is a new record for a C196. And that’s before spin!

Not sure further searching is really going to produce anything close to this level.

ETA: I just finished up 2M ~2.3M and my best poly has a 6.7 handle! Think I’m done with the C196 - congratulations Gimarel. Very nice work.

[Max0526]

Originally Posted by VBCurtis

Max-
Here's a new poly from today's npr run:

# norm 4.247414e-18 alpha -9.453619 e 3.633e-14 rroots 5
skew: 212535665.26
c0: -427171835436423635663489362007216482587801528243
c1: 4838460431917003989901984452940921240448
c2: 46647312004426277310265061167399
c3: -391513016272276498633528
c4: -1571635618437636
c5: 1335600
Y0: -623070090839342313940304884501281148
Y1: 582300282226724749

This search isn't a priority; I just started a GNFS180 on CADO, and won't be running this C186 until that is done in 6-8 weeks.

In the meantime, taking 10-12M on the C196 from Aliq 3366.

The earliest I can get to it now is September 3 when I get back home from vacation. Good thing it can wait until then.

[VBCurtis]

Originally Posted by swellman

Wow! Your first/best poly is a new record for a C196. And that’s before spin!

Not sure further searching is really going to produce anything close to this level.

Well, the previous C196 record was well below the regression for the rest of the 190's. This record is closer to what we'd expect of a C196, but is less than 15% better than the C197 record and less than 25% better than the C198 record.

It's definitely a good enough poly to send to the queue, but I don't think it's so good that we should give up. I'm going to give it another 4-6 GPU days, at least.