mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > Factoring

Reply
 
Thread Tools
Old 2021-06-07, 15:08   #111
charybdis
 
charybdis's Avatar
 
Apr 2020

5×73 Posts
Default

Quote:
Originally Posted by bur View Post
charybdis, thanks for the info. Do you have any suggestions on ncurves or I? I used ncurves0=20 and ncurves1=17 and tasks.I=13 as per VBCurtis' file.
ncurves is of minor importance; you can leave it out and CADO will pick its own values. The only concern is that you can lose relations if you set it too low. I think Curtis has done more investigation than I have.

The I value should probably be the same as for a GNFS job of a similar difficulty. Again, I'm not the best person to ask about converting quartic SNFS difficulty to GNFS difficulty.
charybdis is offline   Reply With Quote
Old 2021-06-07, 15:21   #112
Max0526
 
"Max"
Jun 2016
Toronto

19·47 Posts
Default

Quote:
Originally Posted by bur View Post
cado chokes on the number being divisible by 5, if I just divide it and use that as n it still fails:
Code:
Error:Generate Factor Base: Stderr output (last 10 lines only) follow (stored in
 file ./../factorizations/SNFS/c163/forum.makefb.stderr.2):
Error:Generate Factor Base:     Error, we found no compatible mapping that facto
rs through the given number fields. Your poly file is most probably wrong.
Sorry for always running into trouble...
I am looking into this right now (checking both polys first, generating two new polys [will be 4 from now on], live testing on CADO).

EDIT:
1) poly 1 and poly 2 were incorrect, sorry, the factor 5 should have been divided out;
2) try poly 4 maybe to see what CADO says
Code:
(10, 6) c163 / snfs164 --> poly 3
n: 2467882181617632760392274713430498852785564134165212952200359459249375570944927438733634466076889563023121950372002767078160693289161459940091235869821075619841037
# a = 139630063397928018228902803426662257004064/96442318980814750191888886570850339145357
Y0: -139630063397928018228902803426662257004064
Y1: 96442318980814750191888886570850339145357

# poly 2*x^4 - 30*x^3 + 169*x^2 - 420*x + 392
c0: 392
c1: -420
c2: 169
c3: -30
c4: 2

skew: 2.87239
# E = 2.02411475e-10

----------------------

(10, 6) c163 / snfs164 --> poly 4
n: 2467882181617632760392274713430498852785564134165212952200359459249375570944927438733634466076889563023121950372002767078160693289161459940091235869821075619841037
# a = 246139212525330982538652742856739099577364/56548015152835316412202371306089694287053
Y0: -246139212525330982538652742856739099577364
Y1: 56548015152835316412202371306089694287053

# poly 2*x^4 - 30*x^3 + 169*x^2 - 420*x + 392
c0: 392
c1: -420
c2: 169
c3: -30
c4: 2

skew: 3.94987
# E = 2.19936049e-10 <-- better than poly 1 and poly 2, still worse than poly 1
STOP!

ECM found a factor:
Code:
P35 = 14726602793483369206892252087228089
P129 = 167579869996200950127785818957768321886959759240855577131594731002945383740513830590261038852427234873121231863066676457807506933

Last fiddled with by Max0526 on 2021-06-07 at 15:53
Max0526 is offline   Reply With Quote
Old 2021-06-07, 15:29   #113
swishzzz
 
Jan 2012
Toronto, Canada

3×29 Posts
Default

Quote:
Originally Posted by EdH View Post
I ECM'd line the 119, C178 with no luck:
Code:
152@11e3 
304@5e4 
456@25e4 
912@1e6 
2432@3e6 
3344@11e6 
4080@43e6 
1080@26e7
(Sorry I'm too lazy to re-re-relearn how to determine t.)
Entering the largest 3 into http://www.wraithx.net/math/ecmprobs/ecmprobs.html and picking out the closest digit failure probability to 1/e this is roughly a t50.

Last fiddled with by swishzzz on 2021-06-07 at 15:30
swishzzz is offline   Reply With Quote
Old 2021-06-07, 16:57   #114
Plutie
 
Dec 2020
Montreal

22·5 Posts
Default

taken back, rechecked spreadsheet (already taken)

Last fiddled with by Plutie on 2021-06-07 at 17:10
Plutie is offline   Reply With Quote
Old 2021-06-07, 17:10   #115
Max0526
 
"Max"
Jun 2016
Toronto

19·47 Posts
Default

Quote:
Originally Posted by Plutie View Post
Could I reserve (10, 5)'s c128 and c137?
They are both taken by thyrex, sorry.
Maybe (10, -2) then? Line 155.
Or (10, 6)? Line 147.
Or (9, -4)? Line 117.

Last fiddled with by Max0526 on 2021-06-07 at 17:12
Max0526 is offline   Reply With Quote
Old 2021-06-07, 17:21   #116
EdH
 
EdH's Avatar
 
"Ed Hall"
Dec 2009
Adirondack Mtns

73618 Posts
Default

Quote:
Originally Posted by swishzzz View Post
Entering the largest 3 into http://www.wraithx.net/math/ecmprobs/ecmprobs.html and picking out the closest digit failure probability to 1/e this is roughly a t50.
Excellent! Thank you for the link.
EdH is offline   Reply With Quote
Old 2021-06-07, 17:44   #117
Plutie
 
Dec 2020
Montreal

22×5 Posts
Default

Quote:
Originally Posted by Max0526 View Post
They are both taken by thyrex, sorry.
Maybe (10, -2) then? Line 155.
Or (10, 6)? Line 147.
Or (9, -4)? Line 117.
I'll take both (10, -2) composites.
Plutie is offline   Reply With Quote
Old 2021-06-07, 20:33   #118
Max0526
 
"Max"
Jun 2016
Toronto

89310 Posts
Default 4 SNFS polys for bur, c162 / snfs178

Code:
(10, -3) c162 / snfs178 --> poly 1
n: 4026996679432689240601147270873947508679893119056000869085573680158349103286328458621841641176739825407849457835374955074822709172657994981592651971459711228627134568163675088092
# a = 450801066613781089738788883327812943453202826/152280662435343896113581435286061100961790077
Y0: -450801066613781089738788883327812943453202826
Y1: 152280662435343896113581435286061100961790077

# poly x^4 - 12*x^3 + 62*x^2 - 168*x + 196
c0: 196
c1: -168
c2: 62
c3: -12
c4: 1

skew: 3.44132
# E = 3.96286016e-11

------------------------

(10, -3) c162 / snfs178 --> poly 2
n: 4026996679432689240601147270873947508679893119056000869085573680158349103286328458621841641176739825407849457835374955074822709172657994981592651971459711228627134568163675088092
# a = 152280662435343896113581435286061100961790077/32200076186698649267056348809129495960943059
Y0: -152280662435343896113581435286061100961790077
Y1: 32200076186698649267056348809129495960943059

# poly x^4 - 12*x^3 + 62*x^2 - 168*x + 196
c0: 196
c1: -168
c2: 62
c3: -12
c4: 1

skew: 4.06712
# E = 5.27464924e-11 <-- best by E score

------------------------

(10, -3) c162 / snfs178 --> poly 3
n: 4026996679432689240601147270873947508679893119056000869085573680158349103286328458621841641176739825407849457835374955074822709172657994981592651971459711228627134568163675088092
# a = 158321583127594494715536857816431460393957482/23480357688549299045356040049543117118017841
Y0: -158321583127594494715536857816431460393957482
Y1: 23480357688549299045356040049543117118017841

# poly x^4 - 12*x^3 + 62*x^2 - 168*x + 196
c0: 196
c1: -168
c2: 62
c3: -12
c4: 1

skew: 4.53820
# E = 5.23859187e-11 <-- a very close second choice

------------------------

(10, -3) c162 / snfs178 --> poly 4
n: 4026996679432689240601147270873947508679893119056000869085573680158349103286328458621841641176739825407849457835374955074822709172657994981592651971459711228627134568163675088092
# a = 164362503819845093317492280346801819826124887/79160791563797247357768428908215730196978741
Y0: -164362503819845093317492280346801819826124887
Y1: 79160791563797247357768428908215730196978741

# poly x^4 - 12*x^3 + 62*x^2 - 168*x + 196
c0: 196
c1: -168
c2: 62
c3: -12
c4: 1

skew: 3.08394
# E = 4.57425598e-11
Please let me know if something is off.
Max0526 is offline   Reply With Quote
Old 2021-06-07, 20:48   #119
swishzzz
 
Jan 2012
Toronto, Canada

3×29 Posts
Default

Nearly halfway through ECM for the c165 in line 130. Is this the best snfs poly for it?

Code:
n: 395249338246469246487830723710804680638156544918234600322382035868254052724990588645820778897781462243303165391033834083657866321448102751051115362638758251280263651
c4: 1
c3: -6
c2: 17
c1: -84
c0: 196
Y1: 118957099302951841007289167805344175547512446544032
Y0: -551290229155074260762720133393713469817114576667221
skew: 4.03713
Cownoise gives a Murphy e score of 2.1025e-12. I haven't done any quartics of this size before, guessing it would be around the same difficulty as a quintic SNFS 215-220. Will have a lot of sieving parameter experimenting to do...
swishzzz is offline   Reply With Quote
Old 2021-06-07, 20:54   #120
fivemack
(loop (#_fork))
 
fivemack's Avatar
 
Feb 2006
Cambridge, England

191616 Posts
Default

For the C164 (SNFS 168.6) from (2,-8), I used

rlim=alim=6000000
sieve Q=4.6M-7.0M
lpbr=lpba=27
mfbr=mfba=54
alambda=rlambda=2.6
gnfs-lasieve4I14e

It might be worth using 3LPR even at this small scale.

Sieving was 100 thread-hours (~6.5 hours real-time, since I was using 40 threads on a a 40-thread machine with 72 other threads running on it).

I make no claim that these are optimal parameters, I haven't done degree-4 SNFS jobs of this medium size before. Sieve times looks no worse than a quintic SNFS 180 that I did years ago on much slower hardware.

Last fiddled with by fivemack on 2021-06-07 at 20:56
fivemack is offline   Reply With Quote
Old 2021-06-07, 21:05   #121
Max0526
 
"Max"
Jun 2016
Toronto

19·47 Posts
Default

Quote:
Originally Posted by swishzzz View Post
Nearly halfway through ECM for the c165 in line 130. Is this the best snfs poly for it?

Code:
n: 395249338246469246487830723710804680638156544918234600322382035868254052724990588645820778897781462243303165391033834083657866321448102751051115362638758251280263651
c4: 1
c3: -6
c2: 17
c1: -84
c0: 196
Y1: 118957099302951841007289167805344175547512446544032
Y0: -551290229155074260762720133393713469817114576667221
skew: 4.03713
Cownoise gives a Murphy e score of 2.1025e-12. I haven't done any quartics of this size before, guessing it would be around the same difficulty as a quintic SNFS 215-220. Will have a lot of sieving parameter experimenting to do...
Today we can make the SNFS polys much better
Code:
(1, -9) c165 / snfs202 --> poly 1
n: 395249338246469246487830723710804680638156544918234600322382035868254052724990588645820778897781462243303165391033834083657866321448102751051115362638758251280263651
# a = 551290229155074260762720133393713469817114576667221/118957099302951841007289167805344175547512446544032
Y0: -551290229155074260762720133393713469817114576667221
Y1: 118957099302951841007289167805344175547512446544032

# poly x^4 - 6*x^3 + 17*x^2 - 84*x + 196
c0: 196
c1: -84
c2: 17
c3: -6
c4: 1

skew: 4.03713
# E = 2.10250968e-12

------------------------

(1, -9) c165 / snfs202 --> poly 2
n: 395249338246469246487830723710804680638156544918234600322382035868254052724990588645820778897781462243303165391033834083657866321448102751051115362638758251280263651
# a = 237914198605903682014578335610688351095024893088064/78755747022153465823245733341959067116730653809603
Y0: -237914198605903682014578335610688351095024893088064
Y1: 78755747022153465823245733341959067116730653809603

# poly x^4 - 6*x^3 + 17*x^2 - 84*x + 196
c0: 196
c1: -84
c2: 17
c3: -6
c4: 1

skew: 3.46814
# E = 2.42232922e-12

------------------------

(1, -9) c165 / snfs202 --> poly 3
n: 395249338246469246487830723710804680638156544918234600322382035868254052724990588645820778897781462243303165391033834083657866321448102751051115362638758251280263651
# a = 75461831943266896733563462172336767627064790491093/38554394741355090639202298878573958685948861075174
Y0: -75461831943266896733563462172336767627064790491093
Y1: 38554394741355090639202298878573958685948861075174

# poly x^4 - 6*x^3 + 17*x^2 - 84*x + 196
c0: 196
c1: -84
c2: 17
c3: -6
c4: 1

skew: 3.03200
# E = 2.85259451e-12 <-- the best poly 

------------------------

(1, -9) c165 / snfs202 --> poly 4
n: 395249338246469246487830723710804680638156544918234600322382035868254052724990588645820778897781462243303165391033834083657866321448102751051115362638758251280263651
# a = 539761526378971268948832184300035421603284055052436/75461831943266896733563462172336767627064790491093
Y0: -539761526378971268948832184300035421603284055052436
Y1: 75461831943266896733563462172336767627064790491093

# poly x^4 - 6*x^3 + 17*x^2 - 84*x + 196
c0: 196
c1: -84
c2: 17
c3: -6
c4: 1

skew: 4.62035
# E = 2.11748989e-12
Let me know if something is off.
Max0526 is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
factoring 2ⁿ-2 equivalent to factoring 2ⁿ-1(I think) baih Miscellaneous Math 9 2020-09-21 07:11
OpenCL GPU P-1 Factoring and ECM Factoring xx005fs GPU Computing 3 2018-10-27 14:49

All times are UTC. The time now is 08:19.


Wed Jul 28 08:19:47 UTC 2021 up 5 days, 2:48, 0 users, load averages: 1.13, 1.35, 1.45

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.