GNFS targets which need more ECM - Page 16

swellman · 2016-10-31, 12:14

Quote:

Originally Posted by fivemack

C197_141_86: 20480 curves at B1=3e8 completed with no factor found; so that's the 35K done.

Shall I set the GTX970 to run polynomial search for two weeks?

Nice job on the ECM. No objection but one suggestion - search for a short time to get a baseline GNFS poly and compare to the SNFS poly to verify GNFS is the optimal path forward. If a barely acceptable GNFS poly can't beat SNFS then two weeks additional searching is unlikely to help.

Look at C186_145_72 - SNFS is slightly better than initial GNFS efforts, and now it's been suggested to me that further poly searching is not likely to help.

fivemack · 2016-10-31, 14:12

Quote:

Originally Posted by swellman

Nice job on the ECM. No objection but one suggestion - search for a short time to get a baseline GNFS poly and compare to the SNFS poly to verify GNFS is the optimal path forward. If a barely acceptable GNFS poly can't beat SNFS then two weeks additional searching is unlikely to help.

Look at C186_145_72 - SNFS is slightly better than initial GNFS efforts, and now it's been suggested to me that further poly searching is not likely to help.

Excellent point ... the SNFS polynomial

Code:

N 19182977796570883027397612007795349768467584440992117523739592612776325655445541732585511670283636370086880565577351234162227572914093167428002460062418878069374344571514831893000598109126181591461
SKEW 0.561251022556
A6 636056
A0 19881
R0 311504538542350645715503145019022560519520256
R1 1227630238259041401156597561561

has a combined score of 8.624e-15 and an hour of GNFS-search gets me nothing more than 2.835e-15. So this is an SNFS number. I guess I'll queue it up after trial sieving ... can you suggest a target for November's ECM GPU cycles?

swellman · 2016-10-31, 14:39

C192_139_81 is just coming out of t55 via yoyo. Needs 20k curves @B1=3e8. How does that sound? I'm on the road right now but will get started upon my return this weekend.

BTW - would you also queue C186_145_72 to 15e? The poly is given in post 163 of this thread. Thank you in advance. fivemack: queued

fivemack · 2016-10-31, 16:17

Quote:

Originally Posted by swellman

C192_139_81 is just coming out of t55 via yoyo. Needs 20k curves @B1=3e8. How does that sound? I'm on the road right now but will get started upon my return this weekend.

BTW - would you also queue C186_145_72 to 15e? The poly is given in post 163 of this thread. Thank you in advance. fivemack: queued

I think C192_139_81 is probably an SNFS target; I think we're about to run out of achievable unambiguous GNFS targets at XYYXF, the only hope is if ECM efforts knock a C6x off one of the C25x difficulty>300 numbers.

This doesn't mean I'm not going to run ECM curves on C192_139_81.

swellman · 2016-10-31, 19:37

Quote:

Originally Posted by fivemack

I think C192_139_81 is probably an SNFS target; I think we're about to run out of achievable unambiguous GNFS targets at XYYXF, the only hope is if ECM efforts knock a C6x off one of the C25x difficulty>300 numbers.

Agreed.

Quote:

This doesn't mean I'm not going to run ECM curves on C192_139_81.

Excellent. I'll plug away on it as well. Even if all remaining numbers are best seived via SNFS, they still need ECM.

And thanks for queuing C186_145_72.

amphoria · 2016-11-01, 19:39

No one appears to have run any curves on C194_139_131 at B1=300M yet so I will reserve it.

amphoria · 2016-11-14, 12:37

Quote:

Originally Posted by amphoria

No one appears to have run any curves on C194_139_131 at B1=300M yet so I will reserve it.

20,000 curves with B1=300M complete with no factor found.

I also ran poly select for a day and found the following poly which is a small improvement on Max's from post #152:

Code:

# norm 6.010324e-019 alpha -7.291559 e 1.119e-014 rroots 3
skew: 314183815.04
c0: 42200690310394645731006573695629365784697161008
c1: -4174931980001203202982947992258045649412
c2: 34314236194563307930578628324180
c3: 171993435884563172649601
c4: -165572939397788
c5: 130548
Y0: -48339132221599400156975896129259741483
Y1: 76780817191247017

This is another borderline GNFS/SNFS candidate so I did some trial sieving with Yafu's best candidate for SNFS of

Code:

# 139^131+131^139, difficulty: 294.30, anorm: 2.70e+038, rnorm: -1.12e+055
# scaled difficulty: 297.07, suggest sieving rational side
# size = 6.576e-015, alpha = 0.000, combined = 2.615e-015, rroots = 0
type: snfs
size: 294
skew: 5.1292
c6: 1
c0: 18209
Y1: -4980119914244835315191043937110847348056000415691
Y0: 140063703503689367173618364344202364099995564521

However, GNFS gave a significantly better yield. The only problem is that they are both 16e jobs so are unlikely to be processed soon. Also further poly select will be required for GNFS as the msieve target is 1.16e-014 to < 1.33e-014.

pinhodecarlos · 2016-11-15, 08:41

Quote:

Originally Posted by amphoria

However, GNFS gave a significantly better yield. The only problem is that they are both 16e jobs so are unlikely to be processed soon.

Dave,

Not if you contact Greg (aka frmky) to queue it on 16e grid. Please let me know if you need my support.

Carlos

swellman · 2016-11-15, 21:03

Amphoria - nice job on the ECM of C194_139_131 and subsequent test sieving. I'll post a request for polynomial over in the msieve subforum along with your result - maybe someone will find something slightly better within the expected range of Murphy e-score. As to running NFS on it, there's Greg (NFS@Home), Ryan Propper (he's done several C19x GNFS jobs recently), possibly others with big hardware (fivemack - any interest?). Or we defer it until another time.

Do you want to tackle C193_143_93? It's next on the list.

And an update - I am currently running ECM on C192_139_81. Progress is slow but as other jobs finish I then put that machine on this effort. Worst case I should have 10,000 curves completed by mid Dec.

fivemack · 2016-11-15, 23:25

My new hardware is now installed, and can sieve a C170 inside a week, so a C192 probably within three months; though it seems slightly silly to run on second-hand four-year-old data-centre hardware when Ryan can run on rather larger quantities of rather newer data-centre hardware.

swellman · 2016-11-22, 22:19

Quote:

Originally Posted by fivemack

My new hardware is now installed, and can sieve a C170 inside a week, so a C192 probably within three months; though it seems slightly silly to run on second-hand four-year-old data-centre hardware when Ryan can run on rather larger quantities of rather newer data-centre hardware.

Update: Ryan has taken up GNFS of C194_139_131. Currently in LA, factors expected by mid-December.

Fivemack - what's your status on ECM of C192_139_81? I've run about 5,000 curves so far @B1=3e8, no joy.

2016-10-31, 14:39	#168
swellman Jun 2012 BFC₁₆ Posts	C192_139_81 is just coming out of t55 via yoyo. Needs 20k curves @B1=3e8. How does that sound? I'm on the road right now but will get started upon my return this weekend. BTW - would you also queue C186_145_72 to 15e? The poly is given in post 163 of this thread. Thank you in advance. fivemack: queued Last fiddled with by fivemack on 2016-10-31 at 15:56

2016-11-01, 19:39	#171
amphoria "Dave" Sep 2005 UK 2³×347 Posts	C194_139_131 No one appears to have run any curves on C194_139_131 at B1=300M yet so I will reserve it.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Ready GNFS targets	XYYXF	XYYXF Project	86	2020-03-07 16:23
SNFS targets which need more ECM	XYYXF	XYYXF Project	57	2017-07-04 19:15
Ready SNFS targets	XYYXF	XYYXF Project	25	2016-11-20 21:35
3,697+ (GNFS 220.9)	pinhodecarlos	NFS@Home	0	2014-12-24 19:13
3,766+ (GNFS 215.5)	pinhodecarlos	NFS@Home	34	2014-04-01 21:27

2016-11-15, 21:03	#174
swellman Jun 2012 BFC₁₆ Posts	Amphoria - nice job on the ECM of C194_139_131 and subsequent test sieving. I'll post a request for polynomial over in the msieve subforum along with your result - maybe someone will find something slightly better within the expected range of Murphy e-score. As to running NFS on it, there's Greg (NFS@Home), Ryan Propper (he's done several C19x GNFS jobs recently), possibly others with big hardware (fivemack - any interest?). Or we defer it until another time. Do you want to tackle C193_143_93? It's next on the list. And an update - I am currently running ECM on C192_139_81. Progress is slow but as other jobs finish I then put that machine on this effort. Worst case I should have 10,000 curves completed by mid Dec.

2016-11-15, 23:25	#175
fivemack (loop (#_fork)) Feb 2006 Cambridge, England 6419₁₀ Posts	My new hardware is now installed, and can sieve a C170 inside a week, so a C192 probably within three months; though it seems slightly silly to run on second-hand four-year-old data-centre hardware when Ryan can run on rather larger quantities of rather newer data-centre hardware.