mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Msieve (https://www.mersenneforum.org/forumdisplay.php?f=83)
-   -   Polynomial Request Thread (https://www.mersenneforum.org/showthread.php?t=18368)

RichD 2019-09-21 14:05

C179 Poly
 
I'll start at 4M.

swellman 2019-09-22 00:12

C179
 
The [url=https://www.mersenneforum.org/showpost.php?p=503952&postcount=134]top score[/url] for a C179 is 1.226e-13. I will report anything above 1e-13.

Eta: Best poly so far had a very unimpressive 8-handle.

RichD 2019-09-22 18:40

C179 Poly
 
The best in the 4M range was 8.0. Continuing with 5M.

RichD 2019-09-23 20:33

C179
 
Slightly better with an 8.1 in the 5M range. Continuing with 6M.

swellman 2019-09-24 14:32

[CODE]# norm 1.773481e-17 alpha -7.401255 e 8.847e-14 rroots 5
skew: 40647429.83
c0: -18595888596752473177165527758317903718329125
c1: 1059949905692221719571614352991024280
c2: 392078004850643944894698126859
c3: 4556416357333536663008
c4: -249788702485670
c5: 472056
Y0: -42167478289890691142116552045704416
Y1: 32870750612783513
[/CODE]

Best poly I found in 0-2M per msieve-GPU. Maybe it can be spun, but likely nowhere near enough.

Searching 2-4M next. Also giving CADO a try.

VBCurtis 2019-09-24 15:37

The essential change to default params for CADO poly searches is to use nq=5^6 = 15625, rather than default 5^5 = 3125.
It's possible 5^7 = 78125 is better yet, I haven't experimented enough yet.
Edit: Search time is linear in nq, so this will take 5x longer than default. Reducing admax is wise, but user discretion.

RichD 2019-09-25 03:01

C179 Poly
 
Still can't reach the bottom baseline. Continuing with 7M range.
[CODE]R0: -24741249053470004438219697698619541
R1: 226079261790988943
A0: 2152923289452829795026397093503954308214840
A1: 906820272218650478206589559652613616
A2: -93390438080236902856198560840
A3: -7293261983756385180071
A4: 293986866487826
A5: 6788496
skew 18475609.33, size 1.433e-17, alpha -6.755, combined = 8.836e-14 rroots = 5
[/CODE]

RichD 2019-09-25 03:22

C179 Poly
 
FWIW, the second best.
[CODE]R0: -24987693698702075254348067265644565
R1: 172231128369965383
A0: -50419592101198480310780901572126814289657048
A1: 9740807826328463118506156299719068348
A2: 505327805862328112441986143930
A3: -34184832538254122038085
A4: -537448637663882
A5: 6460272
skew 33292697.09, size 1.412e-17, alpha -8.022, combined = 8.678e-14 rroots = 5
[/CODE]

swellman 2019-09-25 16:42

C179
 
Currently searching 2-4M, should have results this weekend.

Shifting my focus to CADO - it’s searching now. Hoping to get a poly by this weekend, as the search space was limited < 200K. Kind of curious to see how CADO compares to msieve on a C179. Hoping to try a few variations on a couple of the parameters.

RichD 2019-09-25 20:43

C179 Poly
 
Slightly better. Continuing with 8M.
[CODE]R0: -24072095668185497587750658027965575
R1: 82054100518914191
A0: 15048522563377264495445178085842295992994456
A1: 788071549401826285185191715096496830
A2: -169693031484058324571175089342
A3: -22022540312474255104209
A4: 214330489212070
A5: 7785960
skew 27437684.24, size 1.461e-17, alpha -7.830, combined = 8.930e-14 rroots = 3[/CODE]

Max0526 2019-09-25 22:10

C179 polys
 
None of the 4 polys above seems to spin up, sorry.


All times are UTC. The time now is 22:02.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.