Polynomial Request Thread - Page 107

RichD · 2018-04-12, 14:33

I'll start 4-5M later today.

VBCurtis · 2018-04-12, 15:06

Quote:

Originally Posted by Dubslow

Forgetful moron here, what's the size of a GNFS-205 matrix? North or south of 32 GiB?

We don't have much data for such jobs, but RyanP did RSA-210 here: http://mersenneforum.org/showthread.php?t=18626

In the first post of that thread, his log mentions matrix size 65M, needing 26GB. That puts a C205 matrix around 50-55M and 20-23GB. I think it'll safely fit on a 32GB machine.

swellman · 2018-04-12, 15:21

So we are going for the C205? Staring into the abyss...

VBCurtis · 2018-04-12, 17:09

We're gonna do both, and the forthcoming BOINC challenge-pentathalon-thingy may help Greg agree to run the C205 on 16e. 15e has a ton of work right now, though if we quickly get something near record range for the C205 we should probably switch to the C200 in case both jobs are needed to feed the queues.

Overnight produced 1.77e-15 for C205 from GPU-msieve. C204 record is around 2.6e-15, so something in the 2.25+ range would be a respectable poly, and 2.5 would cause me to label it ready for sieving even if we only spend a week.

Anybody wanna ask Greg about 276-C205 for 16e? If he says no, we should definitely feed the C200 to 15e before we try feeding C205 to 15e! Either one can be done in conjunction with a manually-organized effort using 16f on the smallest Q values to improve average yield; something like 5M to 100M on 16f and 100M-900M on 15e can finish C200, if not C205.

swellman · 2018-04-12, 17:40

Fair enough, but does anyone have details on Max’s C200? I may just work on that first/exclusively due to my slow speed with these large GNFS jobs.

Still reserving 1 to 1e5 for the C200.

Max0526 · 2018-04-12, 18:39

Quote:

Originally Posted by swellman

Fair enough, but does anyone have details on Max’s C200? I may just work on that first/exclusively due to my slow speed with these large GNFS jobs.

Still reserving 1 to 1e5 for the C200.

Cofactor of repunit R507=(10^507-1)/9, http://factordb.com/index.php?id=1100000000192594626

ECM effort:

Code:

-------+-------+----------------------+------------+----------+-----------+----------+
10^n-1 |compos.| factors              | date       |B1/curves/|  sigma/   | contrib. |
       | /prp  |                      |            |  length  |  method   |          |
-------+-------+----------------------+------------+----------+-----------+----------+
 507   |  c200 | 05: 11/13/23/29/     |Sep 06, 2016|85e7/70000|           |YOYO@home |
 507   |       |                      |Jul 09, 2017|76e8/5000 |           |YOYO@home |

It's next in line for Kurt Beschorner's team. If our poly is good, and Greg agrees to feed it into NFS@Home queue, I bet Kurt will deal with the matrix like a pro, he just finished a C199 in December. His hardware is all tuned up for this size.
If NFS@Home can't accommodate the sieving, we'll have to wait until Kurt and the team finishes R447.
In any case, Kurt and Bo usually do test-sieving for the better polynomials that I provide. If any of you guys wishes to weigh in on test-sieving / sieving parameters, I am sure Kurt and Bo will appreciate your opinion.
The best poly I have so far has the E = 4.163e-15 which I spinned up from E = 3.87e-15. I think that anything above 3.4e-15 would be of value for us right now.

Number:

Code:

n: 46802612901958501868841642965751124516436202272587430374287737787818526612306217484595957815607978887434356766596560265669642422106894901815250576734687838908431237216708427946009827065503443998015551

I would also ask Ryan Propper to please let Kurt's team enjoy factoring this number. :-)

VBCurtis · 2018-04-12, 19:04

Thanks for the info, Max! Let's do the C200 first, then; it makes an interesting new-territory stretch for the 15e queue, and having an eager matrix-solver provides incentive to tackle it first.
Do you know if their team uses GGNFS or CADO for sieving?

I'll take 1M-2M on the C200 also.

Max0526 · 2018-04-12, 19:20

@VBCurtis
> Do you know if their team uses GGNFS or CADO for sieving?
GGNFS only.

RichD · 2018-04-12, 21:07

I'll also take the 4-5M range for this number later tonight after I finish another "smallish" number on my other GPU.

swellman · 2018-04-13, 00:33

I’ve stopped poly searching of the C197. -npr revealed nothing worth reporting. Moving on.

Taking 1-100K on Max’s C200. Sounds like Kurt and his team are well positioned for this job. Hoping we find a record poly quickly! Posting anything over 3.4e-15.

Quote:

Originally Posted by Max0526

I would also ask Ryan Propper to please let Kurt's team enjoy factoring this number. :-)

Don’t think that’s an issue - Ryan hasn’t been factoring for a few months.

Max0526 · 2018-04-13, 02:01

Quote:

Originally Posted by VBCurtis

Dear Max:
Please polish. Msieve-generated.

Code:

# norm 3.102524e-19 alpha -6.943160 e 7.217e-15 rroots 3
skew: 152039729.27
c0: 2931130217185327116601317015565420773934927200
c1: -918193698486396137048328916808242683510
c2: 46768211205451464797643011244055
c3: 40873832936202201931322
c4: 1577994290944152
c5: 991728
Y0: -116079600316758606974465208257186926517
Y1: 13552554441652526533

I tried and couldn't get anything better, sorry. :-(

2018-04-12, 14:33	#1167
RichD Sep 2008 Kansas 3,391 Posts	C205 poly I'll start 4-5M later today.

2018-04-12, 19:20	#1174
Max0526 "Max" Jun 2016 Toronto 19×47 Posts	C200 @VBCurtis > Do you know if their team uses GGNFS or CADO for sieving? GGNFS only.

2018-04-12, 21:07	#1175
RichD Sep 2008 Kansas 110100111111₂ Posts	C200 poly I'll also take the 4-5M range for this number later tonight after I finish another "smallish" number on my other GPU.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
GIMPS wiki account request thread	ixfd64	mersennewiki	169	2018-09-21 05:43
Polynomial Discriminant is n^k for an n-1 degree polynomial	carpetpool	Miscellaneous Math	14	2017-02-18 19:46
Lost Prime Raider password request thread	cheesehead	Forum Feedback	6	2009-07-28 13:02
Polynomial	R.D. Silverman	NFSNET Discussion	13	2005-09-16 20:07
Deutscher Thread (german thread)	TauCeti	NFSNET Discussion	0	2003-12-11 22:12

2018-04-12, 15:21	#1169
swellman Jun 2012 C14₁₆ Posts	So we are going for the C205? Staring into the abyss...

2018-04-12, 17:09	#1170
VBCurtis "Curtis" Feb 2005 Riverside, CA 2×2,437 Posts	We're gonna do both, and the forthcoming BOINC challenge-pentathalon-thingy may help Greg agree to run the C205 on 16e. 15e has a ton of work right now, though if we quickly get something near record range for the C205 we should probably switch to the C200 in case both jobs are needed to feed the queues. Overnight produced 1.77e-15 for C205 from GPU-msieve. C204 record is around 2.6e-15, so something in the 2.25+ range would be a respectable poly, and 2.5 would cause me to label it ready for sieving even if we only spend a week. Anybody wanna ask Greg about 276-C205 for 16e? If he says no, we should definitely feed the C200 to 15e before we try feeding C205 to 15e! Either one can be done in conjunction with a manually-organized effort using 16f on the smallest Q values to improve average yield; something like 5M to 100M on 16f and 100M-900M on 15e can finish C200, if not C205.

2018-04-12, 17:40	#1171
swellman Jun 2012 2²·773 Posts	Fair enough, but does anyone have details on Max’s C200? I may just work on that first/exclusively due to my slow speed with these large GNFS jobs. Still reserving 1 to 1e5 for the C200.

2018-04-12, 19:04	#1173
VBCurtis "Curtis" Feb 2005 Riverside, CA 2·2,437 Posts	Thanks for the info, Max! Let's do the C200 first, then; it makes an interesting new-territory stretch for the 15e queue, and having an eager matrix-solver provides incentive to tackle it first. Do you know if their team uses GGNFS or CADO for sieving? I'll take 1M-2M on the C200 also.