mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Cunningham Tables (https://www.mersenneforum.org/forumdisplay.php?f=51)
-   -   Contributing to Cunningham Project (https://www.mersenneforum.org/showthread.php?t=24211)

swellman 2019-10-12 13:33

2,1084+1?
 
Should I enqueue 2,1084+1 for another round of ECM with Yoyo? If Greg is likely to enqueue it within NFS@Home soon (say by end of Oct) then I’m inclined to cease all further ECM work on 2,1084+1.

To my knowledge, 2,1084+1 has undergone almost 48,000 curves @B1=850M, plus other previous efforts by several contributors.

R.D. Silverman 2019-10-12 14:46

[QUOTE=swellman;527808]Should I enqueue 2,1084+1 for another round of ECM with Yoyo? If Greg is likely to enqueue it within NFS@Home soon (say by end of Oct) then I’m inclined to cease all further ECM work on 2,1084+1.

To my knowledge, 2,1084+1 has undergone almost 48,000 curves @B1=850M, plus other previous efforts by several contributors.[/QUOTE]

We don't know what Greg will queue next. The status page needs an update.
I believe that 2,2330M is ready. There are also the easier 2,1144+, 2,1157+, 2,2158L
(but these could use additional ECM)

If Greg is going to queue 2,1084+ next, then I would say to remove it from YoYo's queue.

frmky 2019-10-14 05:13

I'll add 2,1084+ now.

swellman 2019-10-14 19:08

[QUOTE=frmky;527955]I'll add 2,1084+ now.[/QUOTE]

Noted. No more ECM for 2,1084+. Results will trickle in for the last few hundred curves.

I’ll enqueue the following in Yoyo@Home:
[CODE]
2,1165+
2,1157+
2,1144+
2,2158L
[/CODE]

Note 2,1165+ has already completed 50% t65. It’s the last GNFS job left in the 1987 list AFAIK.

Hoping to plow through all of these in a few months.

R.D. Silverman 2019-10-14 22:44

[QUOTE=swellman;527998]

<snip>

Note 2,1165+ has already completed 50% t65. It’s the last GNFS job left in the 1987 list AFAIK.
[/QUOTE]

It is certainly the last one less than C220. Whether it is truly the "last" depends on how
high NFS@Home can reach. There are several more less than C225.

frmky 2019-10-22 23:52

I updated the status page on NFS@Home. I'm happy to change the order in which the numbers are sieved if it's more convenient.

I also ran a quick test sieve for 2,2210M. It looks like a relatively easy SNFS, but I'm going to first start the LA on 2,2150M to make sure it's really as smooth as it appears.

swellman 2019-10-23 00:40

[QUOTE=frmky;528637]I updated the status page on NFS@Home. I'm happy to change the order in which the numbers are sieved if it's more convenient.

I also ran a quick test sieve for 2,2210M. It looks like a relatively easy SNFS, but I'm going to first start the LA on 2,2150M to make sure it's really as smooth as it appears.[/QUOTE]

Ok, though I note four numbers now enqueued in NFS@Home are currently scheduled to be run to t65 by Yoyo.
[CODE]
2,1165+
2,1157+
2,1144+
2,2158L
[/CODE]

I presume any further ECM of these is counterproductive, yes?

R.D. Silverman 2019-10-23 01:46

[QUOTE=swellman;528643]Ok, though I note four numbers now enqueued in NFS@Home are currently scheduled to be run to t65 by Yoyo.
[CODE]
2,1165+
2,1157+
2,1144+
2,2158L
[/CODE]

I presume any further ECM of these is counterproductive, yes?[/QUOTE]

Yes. Note that two have not been queued.

I would have thought that 2,2210M would be faster with GNFS.... Note that we can stop
the polyselection.

R.D. Silverman 2019-10-24 12:35

[QUOTE=R.D. Silverman;528648]Yes. Note that two have not been queued.

I would have thought that 2,2210M would be faster with GNFS.... Note that we can stop
the polyselection.[/QUOTE]

Greg has queued 2,1165+ by GNFS, but I don't recall seeing any discussion about
polynomial selection. Was one selected? Did we send a polynomial for 2,2330M?

swellman 2019-10-24 15:16

[QUOTE=R.D. Silverman;528762]Greg has queued 2,1165+ by GNFS, but I don't recall seeing any discussion about
polynomial selection. Was one selected? Did we send a polynomial for 2,2330M?[/QUOTE]

Not to my knowledge. Doesn’t mean Greg didn’t find his own poly I suppose.

I am also confused about 2,2210M being run as a SNFS job. But that decision is pending LA on 2,2150M to verify smoothness(?)

Moving forward with ECM, I am planning to enqueue the following in with Yoyo:
[CODE]
2,1115+
2,1135+
2,1180+
2,1139+
3,748+
[/CODE]
Any comments or objections? The last composite is a GNFS job we can run locally if there’s interest.

VBCurtis 2019-10-24 15:53

The best 2330M poly I found, in very limited testing:
[code]Y0: -28961469478570719959140906105066840582630
Y1: 92566325806153545443
c0: 9445533148071673379778086273726321999348087566848
c1: 92405597357112380495238265590709071313716
c2: -5399213363634740995545029971716617
c3: -43667955927695773325644219
c4: 150754501738917390
c5: 39639600
skew: 204525474.24619
# size 1.383e-20, alpha -8.073, combined = 1.181e-15 rroots = 5[/code]

This was found by Gimarel. I regret that I haven't had time to fully test-sieve, and there were two or three polys that are very close in my initial testing (Q=100M, 300M, 500M, 1kq ranges).

If someone else wishes to take on the test-sieving, I'll be happy to PM them my work to build from. I believe there is only a small chance we find a substantially better poly from test-sieving, though 2-4% better is fairly likely.


All times are UTC. The time now is 14:02.

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