Cunningham ECM efforts - Page 6

frmky · 2022-11-13, 08:19

Quote:

Originally Posted by pinhodecarlos

Greg, did you manage to get 2,1091+ into LA stage?

Not yet. With the 14 day deadline I'm still waiting for more stragglers to find their way back.

pinhodecarlos · 2022-11-13, 08:26

Quote:

Originally Posted by frmky

Not yet. With the 14 day deadline I'm still waiting for more stragglers to find their way back.

In that case we will have to wait for the 17th Nov.

frmky · 2022-11-19, 22:25

2,1091+ is in LA now.

Code:

linear algebra completed 226672 of 115553837 dimensions (0.2%, ETA 78h56m)

frmky · 2022-11-23, 18:25

Quote:

Originally Posted by frmky

2,1091+ is in LA now.

And it's done. Another surprisingly smooth run.

pinhodecarlos · 2022-11-23, 20:26

Congrats! Now to the next...

swellman · 2022-11-27, 16:03

2,1097+ is currently sieving, and assuming nothing goes wrong, Greg has announced plans to tackle 2,2194LM.

Again assuming no problems with 2,2194LM, I presume Greg will consider 2,2206L, a SNFS 332 difficulty composite. By coincidence 2,2206L should be starting ECM in the coming days, so if Yoyo doesn’t find a factor perhaps Greg will take it on.

2,1109+ is now enqueued with Yoyo, follow its progress here. I expect this job to actually start ECM in late January.

I am planning to next enqueue 2,2222L with Yoyo@Home. I’ve no notion whether NFS@Home can handle a SNFS 334/GNFS 228 job but at least we can meet our due diligence wrt ECM.

Slowly working through the Gang of 31, er 20-something.

VBCurtis · 2022-11-27, 17:21

Quote:

Originally Posted by swellman

I am planning to next enqueue 2,222L with Yoyo@Home. I’ve no notion whether NFS@Home can handle a SNFS 334/GNFS 228 job but at least we can meet our due diligence wrt ECM.

Slowly working through the Gang of 31, er 20-something.

The one big trick remaining to extend nfs@home's reach is to use 3 large primes on both sides. A 50% yield improvement would add a few digits of range, making high 220s or low 230s in reach. It may even add speed, since such small lims are used for these jobs.

sweety439 · 2022-11-27, 17:31

Quote:

Originally Posted by swellman

I am planning to next enqueue 2,222L with Yoyo@Home ...

Do you mean 2,2222L?

frmky · 2022-11-27, 18:13

Quote:

Originally Posted by VBCurtis

The one big trick remaining to extend nfs@home's reach is to use 3 large primes on both sides.

Does that work with lasieve5? Last I tried it ran significantly slower and sometimes stalled, but I might have been using lasieve4 at the time. I'll try it again. I also plan to try using 36-bit LPs on one side and push msieve filtering to its limit.

charybdis · 2022-11-27, 20:05

Quote:

Originally Posted by VBCurtis

The one big trick remaining to extend nfs@home's reach is to use 3 large primes on both sides. A 50% yield improvement would add a few digits of range, making high 220s or low 230s in reach. It may even add speed, since such small lims are used for these jobs.

Quote:

Originally Posted by frmky

Does that work with lasieve5? Last I tried it ran significantly slower and sometimes stalled, but I might have been using lasieve4 at the time. I'll try it again. I also plan to try using 36-bit LPs on one side and push msieve filtering to its limit.

3LP on both sides undoubtedly will increase yield and thus stretch the upper limit of 16e, but at such a large cost in speed that increasing the lims and losing 50% of your workers might be faster.

VBCurtis · 2022-11-28, 00:27

I've been testing 3LP with much lower mfb's than usually used- say, 90 and 102 for 35/36LP job. I don't see nearly the speed hit that e.g. 103/105 would be, and I get quite a bit better yield. I think I tried this on M1277 first, then on one of the GNFS-225 polys from the recent thread; I don't have records handy, and I was testing with CADO rather than ggnfs.

There should be a transition area where mfb 90-92 is faster than mfb 70, but mfb 100 is slower.

2022-11-19, 22:25	#58
frmky Jul 2003 So Cal A24₁₆ Posts	2,1091+ is in LA now. Code: linear algebra completed 226672 of 115553837 dimensions (0.2%, ETA 78h56m)

2022-11-27, 16:03	#61
swellman Jun 2012 3843₁₀ Posts	2,1097+ is currently sieving, and assuming nothing goes wrong, Greg has announced plans to tackle 2,2194LM. Again assuming no problems with 2,2194LM, I presume Greg will consider 2,2206L, a SNFS 332 difficulty composite. By coincidence 2,2206L should be starting ECM in the coming days, so if Yoyo doesn’t find a factor perhaps Greg will take it on. 2,1109+ is now enqueued with Yoyo, follow its progress here. I expect this job to actually start ECM in late January. I am planning to next enqueue 2,2222L with Yoyo@Home. I’ve no notion whether NFS@Home can handle a SNFS 334/GNFS 228 job but at least we can meet our due diligence wrt ECM. Slowly working through the Gang of 31, er 20-something. Last fiddled with by swellman on 2022-11-27 at 18:31 Reason: Corrected typo in 2,2222L

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Recommended bases and efforts	gd_barnes	Conjectures 'R Us	189	2022-12-27 23:33
Doublecheck efforts; S66/S79 to start with	gd_barnes	Conjectures 'R Us	16	2014-08-07 02:11
Cunningham ECM Now Futile?	R.D. Silverman	GMP-ECM	4	2012-04-25 02:45
ECM efforts mistake?	10metreh	mersennewiki	1	2008-12-28 13:31
ECM Efforts	R.D. Silverman	Factoring	63	2005-06-24 13:41

2022-11-23, 20:26	#60
pinhodecarlos "Carlos Pinho" Oct 2011 Milton Keynes, UK 2³·641 Posts	Congrats! Now to the next...

2022-11-28, 00:27	#66
VBCurtis "Curtis" Feb 2005 Riverside, CA 2²·3·7·67 Posts	I've been testing 3LP with much lower mfb's than usually used- say, 90 and 102 for 35/36LP job. I don't see nearly the speed hit that e.g. 103/105 would be, and I get quite a bit better yield. I think I tried this on M1277 first, then on one of the GNFS-225 polys from the recent thread; I don't have records handy, and I was testing with CADO rather than ggnfs. There should be a transition area where mfb 90-92 is faster than mfb 70, but mfb 100 is slower.