3,748+ c204 Smaller-but-Needed

Batalov · 2019-06-28, 03:06

Is there any interest in 3,748+ c204 Smaller-but-Needed gnfs cofactor?

VBCurtis · 2019-06-28, 03:59

I am personally interested in any major-project candidate over 200 digits; the closer to 200 the shorter the job and the more iterations we have to learn from, so I'd be happy to run this number as the next forum project if there's sufficient interest. On CADO, it would run I=15 and 34LP, much less taxing RAM requirement (say, 3GB per client rather than 9+).

2330L is on pace to sieve in 90 thread-years, where most threads are hyper-threads; that's about 50 core-years if we pretend HT does not exist. A C204 should be about 70% that effort, 60 to 65 [hyper]thread-years.

Maybe start a new thread to gauge interest?

pinhodecarlos · 2019-06-28, 17:22

I’m in.

Dylan14 · 2019-06-28, 21:09

Do we know how much ECM has been done on this cofactor? I would assume at least to the t50 level, since on factordb there is a 51 digit factor of 3^748+1.

swellman · 2019-06-28, 22:10

Quote:

Originally Posted by Dylan14

Do we know how much ECM has been done on this cofactor? I would assume at least to the t50 level, since on factordb there is a 51 digit factor of 3^748+1.

It’s a good question - I can enqueue the number on Yoyo@Home but at what level?

How about a full t60 followed by most of a t65? But that’s just a guess.

Dylan14 · 2019-06-28, 22:39

Quote:

Originally Posted by swellman

It’s a good question - I can enqueue the number on Yoyo@Home but at what level?

How about a full t60 followed by most of a t65? But that’s just a guess.

Assuming that only a t50 has been done(*), ecm-effort.py says to do
6000@43M
2200@110M
39800@260M
24400@850M

So I would posit a full t60 + half a t65 (~40k curves).

(*) I actually assumed 9000 curves were done at t50 level (B1 = 43M)

R.D. Silverman · 2019-06-28, 23:42

Quote:

Originally Posted by Dylan14

Assuming that only a t50 has been done(*), ecm-effort.py says to do
6000@43M
2200@110M
39800@260M
24400@850M

So I would posit a full t60 + half a t65 (~40k curves).

(*) I actually assumed 9000 curves were done at t50 level (B1 = 43M)

Bruce did a full t60.

swellman · 2019-06-29, 01:15

Quote:

Originally Posted by R.D. Silverman

Bruce did a full t60.

In that case I’ll begin enqueuing 3_748+ on Yoyo@Home next week at the t65 level. It will probably take most of the summer to chew it up completely.

Dylan14 · 2019-06-29, 01:16

Quote:

Originally Posted by R.D. Silverman

Bruce did a full t60.

With this bit of information plus the table of optimal ECM parameters, ecm-effort.py now says 22600 curves @ 850M would be enough before passing on to NFS.

swellman · 2019-11-09, 01:47

Ryan Propper was kind enough to ECM 3,748+ to the t65+ level. It is ready for GNFS.

Anyone interested in a poly hunt? The record e-score for a C204 is 2.630e-15. Suggest focusing on deg 5, as deg 6 is very likely to have inferior performance on a C204.

Code:

 533439167600904850230361756102700151678687933392166847323827307497363839257031077774321424872955045754669625577486179222154434651598903112919949771321416511589029559325246084363632977829645558547714072241

R.D. Silverman · 2019-11-09, 01:51

Quote:

Originally Posted by swellman

Ryan Propper was kind enough to ECM 3,748+ to the t65+ level. It is ready for GNFS.

Anyone interested in a poly hunt? The record e-score for a C204 is 2.630e-15. Suggest focusing on deg 5, as deg 6 is very likely to have inferior performance on a C204.

Code:

 533439167600904850230361756102700151678687933392166847323827307497363839257031077774321424872955045754669625577486179222154434651598903112919949771321416511589029559325246084363632977829645558547714072241

Was a poly search for 2,2210M finished? Or will Greg do it via an SNFS quartic???

2019-06-28, 03:06	#1
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 9,829 Posts	3,748+ c204 Smaller-but-Needed Is there any interest in 3,748+ c204 Smaller-but-Needed gnfs cofactor?

2019-11-09, 01:47	#10
swellman Jun 2012 2×1,753 Posts	Ryan Propper was kind enough to ECM 3,748+ to the t65+ level. It is ready for GNFS. Anyone interested in a poly hunt? The record e-score for a C204 is 2.630e-15. Suggest focusing on deg 5, as deg 6 is very likely to have inferior performance on a C204. Code: 533439167600904850230361756102700151678687933392166847323827307497363839257031077774321424872955045754669625577486179222154434651598903112919949771321416511589029559325246084363632977829645558547714072241

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Using 16e on smaller numbers	fivemack	Factoring	3	2017-09-19 08:52
NFS on smaller numbers?	skan	YAFU	6	2013-02-26 13:57
Bernoulli(200) c204	akruppa	Factoring	114	2012-08-20 14:01
checking smaller number	fortega	Data	2	2005-06-16 22:48
Factoring Smaller Numbers	marc	Factoring	6	2004-10-09 14:17

2019-06-28, 03:59	#2
VBCurtis "Curtis" Feb 2005 Riverside, CA 149F₁₆ Posts	I am personally interested in any major-project candidate over 200 digits; the closer to 200 the shorter the job and the more iterations we have to learn from, so I'd be happy to run this number as the next forum project if there's sufficient interest. On CADO, it would run I=15 and 34LP, much less taxing RAM requirement (say, 3GB per client rather than 9+). 2330L is on pace to sieve in 90 thread-years, where most threads are hyper-threads; that's about 50 core-years if we pretend HT does not exist. A C204 should be about 70% that effort, 60 to 65 [hyper]thread-years. Maybe start a new thread to gauge interest?

2019-06-28, 17:22	#3
pinhodecarlos "Carlos Pinho" Oct 2011 Milton Keynes, UK 2²·1,259 Posts	I’m in.

2019-06-28, 21:09	#4
Dylan14 "Dylan" Mar 2017 2⁴·37 Posts	Do we know how much ECM has been done on this cofactor? I would assume at least to the t50 level, since on factordb there is a 51 digit factor of 3^748+1.