mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Aliquot Sequences

Reply
 
Thread Tools
Old 2019-03-21, 05:54   #2872
axn
 
axn's Avatar
 
Jun 2003

37×127 Posts
Default

Quote:
Originally Posted by VBCurtis View Post
Code:
# MurphyE (Bf=4.000e+08,Bg=2.680e+08,area=2.147e+17) = 2.67e-11
I don't recall a good way to run this through msieve to find the usual score; if someone could do so we can compare it to the C196 record and decide how much msieve-GPU poly select to continue with.
Msieve v. 1.52 (SVN 883 pre-release)
Code:
commencing number field sieve (196-digit input)
R0: -88526646710138023135509930133068454122
R1: 21051283668116650162091989
A0: -14138230899246094225865032544701492909383604820
A1: -93571871123367745680754273625163957009
A2: 37510656767740839089260370637202
A3: 475303574396348223775899
A4: -8150194247457752
A5: 4386480
skew 53530918.29, size 3.836e-019, alpha -8.143, combined = 9.505e-015 rroots = 5
axn is online now   Reply With Quote
Old 2019-03-21, 11:03   #2873
swellman
 
swellman's Avatar
 
Jun 2012

2,857 Posts
Default

The skew optimizer suggests it can be improved even further:

skew: 76499475.66494
E-score: 9.56202965e-15
swellman is online now   Reply With Quote
Old 2019-03-21, 15:45   #2874
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

52×173 Posts
Default

The previous C196 record was 8.16e-15. This beats the C195 record too (which was lower than C196).

Seems the CADO poly-select is very effective on large inputs! The C206 last week in the poly select thread also found CADO highly effective.

Let's give Max a bit of time to appear and try his magic before we queue to 15e?
VBCurtis is offline   Reply With Quote
Old 2019-03-21, 22:44   #2875
Max0526
 
"Max"
Jun 2016
Toronto

701 Posts
Default C196 and C195

Quote:
Originally Posted by VBCurtis View Post
The previous C196 record was 8.16e-15. This beats the C195 record too (which was lower than C196).

Seems the CADO poly-select is very effective on large inputs! The C206 last week in the poly select thread also found CADO highly effective.

Let's give Max a bit of time to appear and try his magic before we queue to 15e?

Well, no, it doesn't (and it wasn't).
https://www.mersenneforum.org/showpo...&postcount=134 and https://www.mersenneforum.org/showpo...&postcount=101

I got swellman's PM, I'm on it but so far it doesn't look too promising.

Last fiddled with by Max0526 on 2019-03-21 at 22:45
Max0526 is offline   Reply With Quote
Old 2019-03-22, 04:10   #2876
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

52×173 Posts
Default

Thanks for the 2018 table; I've updated my copy of the chart to correct the out-of-date records.
VBCurtis is offline   Reply With Quote
Old 2019-03-23, 03:55   #2877
Max0526
 
"Max"
Jun 2016
Toronto

701 Posts
Default C196 poly

Quote:
Originally Posted by Max0526 View Post
Well, no, it doesn't (and it wasn't).
https://www.mersenneforum.org/showpo...&postcount=134 and https://www.mersenneforum.org/showpo...&postcount=101

I got swellman's PM, I'm on it but so far it doesn't look too promising.
It was a bit ahead of myself. It's better by now. I'll post it finalized before going to bed.
Max0526 is offline   Reply With Quote
Old 2019-03-23, 13:43   #2878
Max0526
 
"Max"
Jun 2016
Toronto

701 Posts
Default C196 poly -- new record

Crossed 1e-14!

They might not sieve the same, please test-sieve.

from CADO:
Code:
Y0: -88526593606906636872832313995490222337
Y1: 21051283668116650162091989
c0: -16963157469344920969930575255084128666463471380
c1: 104226072109715053258553301087375190176
c2: 40797138907888738018707441706807
c3: 393345121770100035860379
c4: -8094868342851752
c5: 4386480
skew: 76011430.42456
# lognorm 61.32, E 53.00, alpha -8.31 (proj -2.17), 5 real roots
# MurphyE(Bf=1.000e+07,Bg=5.000e+06,area=1.000e+16) = 1.00222868e-14
from msieve:
Code:
Y0: -88526652402994562936959400167280855414
Y1: 21051283668116650162091989
c0: -16945072822861140163957622147315197462146382800
c1: -113754137212796358973399404221053058157
c2: 37121473509227882514071277769078
c3: 484122945206142149130523
c4: -8156125382524952
c5: 4386480
skew: 75049313.58
# size 4.185e-19, alpha -8.314, combined = 1.002e-14 rroots = 5
Max0526 is offline   Reply With Quote
Old 2019-04-10, 11:23   #2879
swellman
 
swellman's Avatar
 
Jun 2012

2,857 Posts
Default

I test sieved both polys and found the first to be slightly better

Code:
n: 1702810278506124393622905886419417373727833799255920336877918870350397583189521830994382118639550712786413448936290016013396498804612664118174179978449082129699734624314982852315442373887230082927
skew: 76011430.42456
c0: -16963157469344920969930575255084128666463471380
c1: 104226072109715053258553301087375190176
c2: 40797138907888738018707441706807
c3: 393345121770100035860379
c4: -8094868342851752
c5: 4386480
Y0: -88526593606906636872832313995490222337
Y1: 21051283668116650162091989
type: gnfs
# lognorm 61.32, E 53.00, alpha -8.31 (proj -2.17), 5 real roots
# MurphyE(Bf=1.000e+07,Bg=5.000e+06,area=1.000e+16) = 1.00222868e-14
rlim: 268000000
alim: 400000000
lpbr: 33
lpba: 33
mfbr: 65
mfba: 94
rlambda: 3.0
alambda: 3.6
Test sieving on the -a side with Q in blocks of 2K using some advice from VBCurtis (mfbr=64 performed horribly):
Code:
Q0=50M 5285 rels
Q0=100M 4644 rels
Q0=200M 5113 rels
Q0=300M 3800 rels
Q0=400M 3234 rels
Q0=500M 3513 rels
Q0=600M 2580 rels
The big unknown to me is how many relations are needed? I would typically aim for 850-900M on a 33 bit job but that seems like overkill considering some of the parameters. Should I initially aim for 600M and slowly increase the Q range until a matrix can be built? Advice welcome.

Once a consensus is reached, I'll enqueue it in 15e.
swellman is online now   Reply With Quote
Old 2019-04-10, 13:47   #2880
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

52×173 Posts
Default

I'd go for 900M to start, and then we can decide how many more to request once we see how many unique relations we have.
You can reserve the postprocessing for me when you queue it.
VBCurtis is offline   Reply With Quote
Old 2019-09-29, 07:12   #2881
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

10000111001012 Posts
Default

The c196 is finished, factors posted. i12535 has a c187, I'm running 3000 curves at B1=2e7 overnight. If no factor, I'll run curves at 6e7 tomorrow until a t50 is complete.
VBCurtis is offline   Reply With Quote
Old 2019-09-30, 04:19   #2882
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

52×173 Posts
Default

t50 complete. I've scheduled 3000 curves at 15e7.

I lack the patience to go past a t55; if a GPUer or yoyo could do half a t60, that should be about enough.
VBCurtis is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Reserved for MF - Sequence 3366 RichD Aliquot Sequences 406 2020-06-10 03:10
Primes in n-fibonacci sequence and n-step fibonacci sequence sweety439 And now for something completely different 17 2017-06-13 03:49
ECM for c166 from 4788:2661 frmky Aliquot Sequences 36 2011-04-28 06:27
ECM work on 4788:2549.c170 schickel Aliquot Sequences 51 2011-01-05 02:32
80M to 64 bits ... but not really reserved petrw1 Lone Mersenne Hunters 82 2010-01-11 01:57

All times are UTC. The time now is 10:35.

Wed Sep 30 10:35:26 UTC 2020 up 20 days, 7:46, 0 users, load averages: 1.17, 1.26, 1.31

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.