mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > NFS@Home

Reply
 
Thread Tools
Old 2021-04-30, 19:16   #34
swellman
 
swellman's Avatar
 
Jun 2012

3,203 Posts
Default

Quote:
Originally Posted by frmky View Post
Yes, that's an option. The relations are still available and valid. You will likely have about 1/2 - 2/3 of the jobs below the leading edge done. You can start sieving at 75M or so and extend at the end if necessary.
So use a Q-range of 75-160M to start, with extension possible if more relations are needed.

10-75M on the -a side is presumably stored somewhere for later appending to the dataset.
swellman is offline   Reply With Quote
Old 2021-04-30, 19:52   #35
jyb
 
jyb's Avatar
 
Aug 2005
Seattle, WA

13·137 Posts
Default

Quote:
Originally Posted by swellman View Post
So use a Q-range of 75-160M to start, with extension possible if more relations are needed.

10-75M on the -a side is presumably stored somewhere for later appending to the dataset.
Well if nothing else, it's stored on my machines; I downloaded it this morning. You might as well assign this job to me for post-processing, though it will now be almost forever until it's done sieving.
jyb is online now   Reply With Quote
Old 2021-04-30, 20:13   #36
swellman
 
swellman's Avatar
 
Jun 2012

62038 Posts
Default

Quote:
Originally Posted by jyb View Post
Well if nothing else, it's stored on my machines; I downloaded it this morning. You might as well assign this job to me for post-processing, though it will now be almost forever until it's done sieving.
Job updated. I’ll move this job into sieving shortly and it will start as soon as Lucas(2250) finishes.

Hoping the pentathlon absorbs a lot of WUs!
swellman is offline   Reply With Quote
Old 2021-04-30, 22:26   #37
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

2·7·683 Posts
Thumbs up

Quote:
Originally Posted by swellman View Post
as soon as Lucas(2250) finishes.
"you got me at hello"
Batalov is offline   Reply With Quote
Old 2021-05-02, 00:18   #38
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

2·7·683 Posts
Lightbulb L(2250)

I looked at the poly chosen for L(2250), and while it is not terrible (monic at least; probably made with phi or by some LLL() calls), but with a bit of rotation I think it could have been made better:
Code:
# let's transform my shifting x by 3
? t=x-3;
? t^4+112*t^3-46*t^2-67*t+31
x^4 + 100*x^3 - 1000*x^2 + 3125*x - 3125

#let's get rid of powers of 5
? t=5*x-3;
? (t^4+112*t^3-46*t^2-67*t+31)/625
x^4 + 20*x^3 - 40*x^2 + 25*x - 5

? t=5*x+2;
? (t^4+112*t^3-46*t^2-67*t+31)/625
x^4 + 24*x^3 + 26*x^2 + 9*x + 1
I like this last poly. (we could even invert it by x := 1/x, to get x^4 + 9*x^3 + 26*x^2 + 24*x + 1
Nothing needed for this project, but just some 2 cents for the next quartic.
If I tinker with the formula a little longer, I think there exists a poly with even smaller coeff's...

P.S. Ah, right, I remembered: it is simply [1, 1, -4, -4, 1] using m=lucas(300).

[1,1,-4,-4,1] poly skew 1.00, size 1.770e-25, alpha 2.444, combined = 3.295e-15 rroots = 4
current poly . . . skew 2.36, size 9.500e-26, alpha 1.585, combined = 2.289e-15 rroots = 4
Batalov is offline   Reply With Quote
Old 2021-05-21, 11:09   #39
swellman
 
swellman's Avatar
 
Jun 2012

3,203 Posts
Default C191_4788_12590

QUEUED AS C191_4788_12590

From AS 4788:i12590

Code:
n: 13722590741694950150749977565393763479875175126696998719123024684676193081635465141226072589122026643757557895201242317939281987981699165430122175451371777818049650902679751036184312239268169
skew: 31007883.18
lss: 0
type: gnfs
c0: 110959553355656567543356858499151578955792060
c1: -38940462646278760204606159044462028312
c2: 7067916544661517449417744835111
c3: 75597863525990589208009
c4: -3827132791134024
c5: 64864800
Y0: -4175569023201610477430743076651843101
Y1: 9301402025178661764203
# norm 1.605668e-18 alpha -8.679931 e 2.105e-14 rroots 1
rlim: 134000000
alim: 134000000
lpbr: 31
lpba: 31
mfbr: 62
mfba: 91
rlambda: 2.6
alambda: 3.4
Test sieving results on the algebraic side, with Q in blocks of 20k:
Code:
60M    40740
100M   37208
150M   32379
200M   30508
250M   33704
Suggesting a Q-range of 60-210M.

Last fiddled with by swellman on 2021-05-28 at 14:38
swellman is offline   Reply With Quote
Old 2021-05-23, 20:35   #40
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

2×7×683 Posts
Default

Quote:
Originally Posted by Batalov View Post
I looked at the poly chosen for L(2250), and while it is not terrible (monic at least; probably made with phi or by some LLL() calls), but with a bit of rotation I think it could have been made better:
..
[1,1,-4,-4,1] poly skew 1.00, size 1.770e-25, alpha 2.444, combined = 3.295e-15 rroots = 4
original poly . .. skew 2.36, size 9.500e-26, alpha 1.585, combined = 2.289e-15 rroots = 4
Lucas(2250) is in LA with original poly.
Code:
weight of 32426723 cycles is about 2594183776 (80.00/cycle)
Batalov is offline   Reply With Quote
Old 2021-05-29, 15:42   #41
swellman
 
swellman's Avatar
 
Jun 2012

3,203 Posts
Default XYYXF C202_147_116

Code:
n: 5028644179505187599603747538081421960747784682907318641913584612912678359414156769454889474663347482684867898378667888076903161202663931121263093390440600128001619421159410735829104594957821918650859551
skew: 14142269.78
lss: 0
type: gnfs
c0: -727972983648726310410079785358393410334322288
c1: 283067213304939263382818921662306595494
c2: 13924751598100321978661105476159
c3: -4388759942722476709817991
c4: -241003407561175168
c5: 4237833600
Y0: -581432102790414830705315679249665723910
Y1: 24941597758289547861913
# norm 9.850693e-20 alpha -7.665176 e 3.712e-15 rroots 3
rlim: 250000000
alim: 200000000
lpbr: 33
lpba: 33
mfbr: 66
mfba: 96
rlambda: 2.8
alambda: 3.6
Sieve on -a side over Q range 30-360M using 16f_small. I’ll take the LA.

Holding off enqueuing this job until 16f digests it current workload.

ETA: now enqueued (June 18), I’ll do the LA.

Last fiddled with by swellman on 2021-06-18 at 10:51
swellman is offline   Reply With Quote
Old 2021-07-12, 19:14   #42
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

10011100000002 Posts
Default

The C201 currently sieving has yielded 586M uniques, a pretty good ratio but not enough for this size. May I have another 40MQ? That should be 100-110M raw relations, and maybe a total of 650M uniques.

Done

Last fiddled with by swellman on 2021-07-12 at 19:25
VBCurtis is offline   Reply With Quote
Old 2021-08-26, 17:20   #43
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

2·7·683 Posts
Arrow

Fibonacci(1409) sieves ("grinds slow but sure") and is en route to a successful dataset. Currently at ~81% completion.
I had (at 77% progress) toy-filtered and it looks like it will converge when 100% finished.
Batalov is offline   Reply With Quote
Old 2021-09-05, 20:12   #44
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

255A16 Posts
Arrow

Fibonacci(1409) is in LA, so we don't need to wait for 4000+ straggler rejects due to the race of lemmings...
Can be moved to 'Post processing'
Batalov is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Queue management for e_small and 15e queues VBCurtis NFS@Home 201 2021-10-17 15:54
Queue management for 14e queue VBCurtis NFS@Home 61 2021-10-17 09:16
Run down the queue on MPRIME without quitting GIMPS Rodrigo Software 7 2018-05-25 13:26
Improving the queue management. debrouxl NFS@Home 10 2018-05-06 21:05
split a prime95 queue & client installation joblack Information & Answers 1 2009-01-06 08:45

All times are UTC. The time now is 06:46.


Mon Oct 18 06:46:35 UTC 2021 up 87 days, 1:15, 0 users, load averages: 1.12, 1.13, 1.16

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.