mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Cunningham Tables

Reply
 
Thread Tools
Old 2019-04-17, 03:59   #23
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

114216 Posts
Default

The previous record for C207 deg5 is 1.63e-15. For C206, it's 2.111e-15 (note: we found this a month ago, and it nearly broke the record for C205; it's a really good score for C206).
In general, if I ever break the record for one digit smaller than my candidate, I cease poly select and head straight to sieving.

If I've done more than half the "reasonable" poly select time and I break the record for my candidate size, I'm quitting and sieving.

For this job, I'd take a 1.85 right now, a 1.75 at the end of the month, and I'll hold out for a 1.65 for quite a while. Each digit of difficulty should be roughly 15% better score than one digit tougher; a poly in the low 1.8's would be "as good" as the 2.111 was last month for the C206. The first digit of this number is 3, so we're not super likely to crush the previous record (if the leading digit was 1, I'd add 5% = almost a tenth to all these hoped-for scores).

Max may have other ideas/hopes.
VBCurtis is offline   Reply With Quote
Old 2019-04-17, 04:24   #24
Max0526
 
"Max"
Jun 2016
Toronto

701 Posts
Default expected E score

I say, let's beat 1.63e-15 first. My best lucky spin would be 18% up (raw score times 1.18). So keep posting anything above 1.38e-15 for now.
Max0526 is offline   Reply With Quote
Old 2019-04-17, 08:19   #25
DukeBG
 
Mar 2018

3×43 Posts
Default

Is there a guarantee that a higher Murphy's E score means (faster and) better sieving? I haven't ran that many jobs (and not at this size anyway), but for those I tried testsieving a bunch of polys (with ggnfs) i didn't see 1:1 correspondence in order by rels/s and order by e score. I'ld say maybe start test sieving polys as they come asap before finding the blessed champion score poly?
DukeBG is offline   Reply With Quote
Old 2019-04-17, 08:59   #26
axn
 
axn's Avatar
 
Jun 2003

4,723 Posts
Default

Quote:
Originally Posted by DukeBG View Post
Is there a guarantee that a higher Murphy's E score means (faster and) better sieving?
As a first order approximation, yes. However, if that was all there to it, no one would do any test sieving, but we do. However, we don't do test sieving on all polys, only the top x with the highest Murphy E scores (maybe those that are within 10% of the largest).
axn is online now   Reply With Quote
Old 2019-04-17, 13:59   #27
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

2·472 Posts
Default

Quote:
Originally Posted by DukeBG View Post
Is there a guarantee that a higher Murphy's E score means (faster and) better sieving? I haven't ran that many jobs (and not at this size anyway), but for those I tried testsieving a bunch of polys (with ggnfs) i didn't see 1:1 correspondence in order by rels/s and order by e score. I'ld say maybe start test sieving polys as they come asap before finding the blessed champion score poly?
CADO's method of calculating E-score has a stronger correspondence with sieve performance than msieve's. I usually test-sieve any poly within 5% of the best-scoring poly, but when CADO generates the polys I haven't seen variances from score higher than 3% (perhaps due to small sample- I've only tested a dozen or so from CADO, but hundreds from msieve).

For this job, I'll be willing to test-sieve any poly with 0.1 of the best-scoring poly. Others are invited to chip in their own testing!
VBCurtis is offline   Reply With Quote
Old 2019-04-17, 16:08   #28
fivemack
(loop (#_fork))
 
fivemack's Avatar
 
Feb 2006
Cambridge, England

2·29·109 Posts
Default

I don't know whether CADO has done more work on efficient sieving with very large skews.

I trial-sieved the 1.414, 1.424, 1.496 polynomials that appeared here, using gnfs-lasieve4I16e, and indeed the speeds (about 3.3s/rel, about 2.9s/rel, about 2.6s/rel) are in the same order as the E-values, but my suspicion is that that's mostly to do with the spinning process having made the skew significantly smaller.

I tend to trial-sieve the top 32 or 40 E-values that I see, and the best-sieving one has always been from the top-10 but isn't that often the top-1.

Last fiddled with by fivemack on 2019-04-17 at 16:11
fivemack is offline   Reply With Quote
Old 2019-04-18, 02:50   #29
pinhodecarlos
 
pinhodecarlos's Avatar
 
"Carlos Pinho"
Oct 2011
Milton Keynes, UK

112138 Posts
Default

Yoyo ECM at B1 = 850,000,000; 8,630 / 9,000 curves.
pinhodecarlos is online now   Reply With Quote
Old 2019-04-19, 07:24   #30
pinhodecarlos
 
pinhodecarlos's Avatar
 
"Carlos Pinho"
Oct 2011
Milton Keynes, UK

10010100010112 Posts
Default

Could we have a step by step guide on how to install and then point the cores to the CADO server please? Much appreciated.

Carlos
pinhodecarlos is online now   Reply With Quote
Old 2019-04-19, 08:03   #31
vebis
 
vebis's Avatar
 
Oct 2015

22·17 Posts
Default

Code:
n: 334377437706404684733884220732190564550147039308686365251932207592010868660163280944362994773042289837865752943086095333357861608005044927356022538834108845220872215597190097852981295940487497385774178731621
skew: 533132.085
c0: 13400871167573270650087884944996454257736480
c1: 170662016202584202246103429104502541792
c2: -106489302432833541662856860257198
c3: -1862164478135527803969537503
c4: 375011464350515249437
c5: 1161035722389120
c6: 848502000
Y0: -1694165626316187446160079002429297
Y1: 428100302665617359981
# MurphyE (Bf=3.436e+10,Bg=1.718e+10,area=5.469e+16) = 1.45e-08
# found by revision 9436c2ff8
# f(x) = 848502000*x^6+1161035722389120*x^5+375011464350515249437*x^4-1862164478135527803969537503*x^3-106489302432833541662856860257198*x^2+170662016202584202246103429104502541792*x+13400871167573270650087884944996454257736480
# g(x) = 428100302665617359981*x-1694165626316187446160079002429297
My best (deg-6)

Last fiddled with by VBCurtis on 2019-04-19 at 13:57 Reason: Clarification; VBCurtis: changed quote tags to code tags
vebis is offline   Reply With Quote
Old 2019-04-19, 10:39   #32
R.D. Silverman
 
R.D. Silverman's Avatar
 
Nov 2003

26×113 Posts
Default

Quote:
Originally Posted by pinhodecarlos View Post
Yoyo ECM at B1 = 850,000,000; 8,630 / 9,000 curves.
Work on this number has slowed almost to a stop and there seem to be no
new numbers in the queue.
R.D. Silverman is offline   Reply With Quote
Old 2019-04-19, 11:42   #33
swellman
 
swellman's Avatar
 
Jun 2012

2·3·479 Posts
Default

Quote:
Originally Posted by R.D. Silverman View Post
Work on this number has slowed almost to a stop and there seem to be no
new numbers in the queue.
Yoyo is aware of the required ECM work for this composite, and has it in queue for another 9000 curves. The server automagically activates the next job when appropriate. It can be maddening to watch the last few curves drip in for a 99% completed task and the next number just won’t seem to advance. A characteristic of BOINC? Fewer WUs to hand out so slower progress towards the end of a task? I don’t know.

But it will get there.

(Note that DONE status indicates the ECM work has started on a number. The graphic indicator sometimes lags this event a few hours.)

Last fiddled with by swellman on 2019-04-19 at 11:43
swellman is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Coordination thread for redoing P-1 factoring ixfd64 Lone Mersenne Hunters 53 2020-10-30 19:39
big job planning henryzz Cunningham Tables 16 2010-08-07 05:08
Sieving reservations and coordination gd_barnes No Prime Left Behind 2 2008-02-16 03:28
Sieved files/sieving coordination gd_barnes Conjectures 'R Us 32 2008-01-22 03:09
Special Project Planning wblipp ElevenSmooth 2 2004-02-19 05:25

All times are UTC. The time now is 07:24.

Sat Oct 31 07:24:55 UTC 2020 up 51 days, 4:35, 2 users, load averages: 2.33, 2.04, 1.81

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.