mersenneforum.org Reserved for MF - Sequence 4788
 User Name Remember Me? Password
 Register FAQ Search Today's Posts Mark Forums Read

2011-07-03, 18:19   #1266
Mini-Geek
Account Deleted

"Tim Sorbera"
Aug 2006
San Antonio, TX USA

17×251 Posts

Quote:
 Originally Posted by Mini-Geek I realized I made a rather large mistake in that math: it ignores that 3 can not be a cofactor of a downdriver, since otherwise it would be the 2*3 perfect driver. ... The new and improved script is attached. More trivia: there is a 5.01% chance of it ending by 163.6 digits, and a 95.05% chance of it ending by 66.2 digits.
(the first line without the downdriver was ~10^165.0; from the time of that posting we only had a ~2.55% chance of breaking by then) Either we were very unlucky, (i.e. about 1 in 39 such runs would end so soon) or my numbers are very off. But anyway, with the breaking being 2*7^2*p, I noticed one more thing that made it a little more inaccurate: I didn't take into account that 2*n^2*p can break it, too! I accounted for this by multiplying the probability of each line breaking by 1 + (sum 1/prime(k)^2, k=3 to infinity) (about 1.091). I was looking around a little for more accurate calculations than mine, and only found the Losing Downdriver thread, which seems to have last settled on numbers very different from mine: 1/420 for 100 digits. I calculate 1.09113630893*3/2/log(10)/100~=1/141. To simply state each, they (mainly Greebley post #6 with modifications for squares by CRGreathouse post #12) say about 1/(4.2*digits) I say 1/(1.4*digits). A factor of 3 different. I think it's because I'm accounting for the fact that the downdriver's cofactor can't have a factor of 2 or 3 by definition.
Does anyone have an automated way to look through many downdrivers and see if my numbers hold up to real test data? Just looking at this one, it'd seem I'm way off, and my memory seems to also say it's lost far easier, but I'd like something far more concrete.

The new script, taking into account the squares, is attached.
Attached Files
 downdriverprobability.zip (689 Bytes, 74 views)

Last fiddled with by Mini-Geek on 2011-07-03 at 18:26

2011-07-05, 01:40   #1267
jrk

May 2008

3×5×73 Posts

Quote:
 Originally Posted by jrk Preliminary poly for line 2696:
It's obsolete now, but in case anyone happens to collect these, here's the best one I found:
Code:
# sieve with ggnfs siever 14e on alg side from Q=12M to ??? (TBD)
# aq4788:2696
n: 8755586957647299753989940018066954131853563140628725006016461530329394193683073655733471708006846371139255974755721298957099259657401475696647916915828783
# norm 6.055972e-15 alpha -7.175530 e 3.223e-12 rroots 3
skew: 2748791.27
c0:  22518556851938655674731450005532544565
c1:  49007811576301845607236669717695
c2:  12280205727986637086832355
c3: -20291335212629008203
c4: -848640062100
c5:  765072
Y0: -408993941175622024607029493362
Y1:  1200827227936261913
rlim: 24000000
alim: 24000000
lpbr: 29
lpba: 29
mfbr: 58
mfba: 58
rlambda: 2.5
alambda: 2.5

 2011-07-05, 01:42 #1268 jrk     May 2008 3·5·73 Posts +400 curves @ 8e7 for line 2703
 2011-07-07, 16:46 #1269 bchaffin   Sep 2010 Portland, OR 22×3×31 Posts I did the c142 from 2703, and four more terms. Now at 2708 with a c157. Except for one 2^2*5 it continues to decrease... I've run 1000 @ 1M and 3000 at 10M so far. I'll keep running curves but this one is too big for me to sieve on my own.
 2011-07-07, 17:44 #1270 jrk     May 2008 21078 Posts I'll start poly selection on line 2708 as usual.
 2011-08-11, 16:04 #1271 bsquared     "Ben" Feb 2007 340310 Posts Done: Code: prp64 factor: 4038843625314075842900554513884322322635236034496696017887771851 prp93 factor: 288180620682225170824303650548350247409615780738733199886385422296800168275435704019909533659 Sequence still decreasing... I've done 1000 curves at 1e6 on the C153 at iteration 2709.
 2011-08-14, 02:00 #1272 bsquared     "Ben" Feb 2007 41·83 Posts I've done 3k curves at 10M and 1800 curves at 100M on the C153 of iteration 2709. I'm run with it a while longer and do the nfs.
 2011-08-14, 02:48 #1273 jrk     May 2008 3·5·73 Posts +960 curves @ B1=8e7
2011-08-14, 03:04   #1274
schickel

"Frank <^>"
Dec 2004
CDP Janesville

2×1,061 Posts

Quote:
 Originally Posted by bsquared Done: Code: prp64 factor: 4038843625314075842900554513884322322635236034496696017887771851 prp93 factor: 288180620682225170824303650548350247409615780738733199886385422296800168275435704019909533659 Sequence still decreasing... I've done 1000 curves at 1e6 on the C153 at iteration 2709.
At least it's still alive in terms of not getting a driver or an upwardly increasing guide!

2011-08-14, 23:26   #1275
biwema

Mar 2004

3×127 Posts

Quote:
 Originally Posted by Batalov E = mc2 = , where m=2 and c=7...
No need to worry about a 2 · 7^2 ...

2011-08-15, 06:03   #1276
Andi47

Oct 2004
Austria

2·17·73 Posts

Quote:
 Originally Posted by biwema No need to worry about a 2 · 7^2 ...
There is a need to worry - it killed the downdriver.

 Thread Tools

 Similar Threads Thread Thread Starter Forum Replies Last Post RichD Aliquot Sequences 474 2021-03-07 20:28 RichD Aliquot Sequences 468 2021-01-27 01:16 kar_bon Aliquot Sequences 127 2020-12-17 10:05 prism019 GPU to 72 6 2020-09-21 22:11 petrw1 Lone Mersenne Hunters 82 2010-01-11 01:57

All times are UTC. The time now is 17:48.

Tue Apr 13 17:48:14 UTC 2021 up 5 days, 12:29, 2 users, load averages: 5.14, 5.04, 4.88

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