mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > And now for something completely different

Reply
 
Thread Tools
Old 2021-05-09, 19:34   #45
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

17·563 Posts
Default

Quote:
Originally Posted by ATH View Post
You can do a Lucas test with LLR ? (For some reason -q does not work for me in LLR 3.8.23)
cllr64.exe -d -oLucasPRPtest=1 input.txt

input.txt:
Code:
ABC (10^$a-1)/9
8177207
or the BPSW version of Lucas test if you use LLR 3.8.24:
cllr.exe -d -oLucasPRPtest=1 -oBPSW=1 -q"(10^8177207-1)/9"
There is a better form, and llr treats it optimally
Code:
ABC ($a*$b^$c$d)/$e
1 10 8177207 -1 9
LLR ver >= 3.8.24 will indeed create $temp.npg exactly like that, but even with earlier versions you can use the ABC file above.
Batalov is offline   Reply With Quote
Old 2021-05-09, 21:21   #46
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

C6816 Posts
Default

Thank you.

I was actually responding to your line:
"We are checking SPRP in 3-4 bases first (the best we can do for this PRP.)"

which I found curious since you can also do Lucas test + Frobenius test with LLR.


Congratulations on this huge PRP!
ATH is offline   Reply With Quote
Old 2021-05-09, 21:29   #47
paulunderwood
 
paulunderwood's Avatar
 
Sep 2002
Database er0rr

53×73 Posts
Default

Quote:
Originally Posted by ATH View Post
Thank you.

I was actually responding to your line:
"We are checking SPRP in 3-4 bases first (the best we can do for this PRP.)"

which I found curious since you can also do Lucas test + Frobenius test with LLR.


Congratulations on this huge PRP!
I'll gladly run it. Starting now.

Last fiddled with by paulunderwood on 2021-05-09 at 21:52
paulunderwood is offline   Reply With Quote
Old 2021-05-09, 21:56   #48
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

17×563 Posts
Default

Quote:
Originally Posted by ATH View Post
which I found curious since you can also do Lucas test + Frobenius test with LLR.
:
But who said that we didn't?

Q: What happens if you simply run a test with llr?
A: it runs PRP, then Lucas and finally Frobenius tests. By default, no special CLI options are needed
Quote:
Originally Posted by paulunderwood View Post
I'll gladly run it. Starting now.
By all means, sure, - run them. They have already been run, but why not run them once again?
Batalov is offline   Reply With Quote
Old 2021-05-10, 00:04   #49
paulunderwood
 
paulunderwood's Avatar
 
Sep 2002
Database er0rr

53·73 Posts
Default

Quote:
Originally Posted by Batalov View Post
But who said that we didn't?

Q: What happens if you simply run a test with llr?
A: it runs PRP, then Lucas and finally Frobenius tests. By default, no special CLI options are needed

By all means, sure, - run them. They have already been run, but why not run them once again?
Hah. I'll abort my run,

Last fiddled with by paulunderwood on 2021-05-10 at 00:06
paulunderwood is offline   Reply With Quote
Old 2021-05-10, 00:52   #50
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

225438 Posts
Lightbulb

One thing that I didn't know, is that FBase=a=<some base> is honored by llr when set as Fermat base, that is no surprise.
But! -- what follows (i.e. L+F tests) is decoupled from the choice of a, so Lucas test happens to be all the same in all runs, so one is enough.
Code:
Starting probable prime test of (10^8177207-1)/9
Using AVX-512 FFT length 1600K, Pass1=1K, Pass2=1600, clm=1, 24 threads, a = 23
(10^8177207-1)/9 is base 23-Fermat PRP! (8177207 decimal digits)  Time : 25493.017 sec.
Starting Lucas sequence
Using AVX-512 FFT length 1600K, Pass1=1K, Pass2=1600, clm=1, 24 threads, P = 6, Q = 2
_____________...and so on...___________
Using AVX-512 FFT length 1600K, Pass1=1K, Pass2=1600, clm=1, 16 threads, a = 5
10^8177207-1)/9 is base 5-Fermat PRP! (8177207 decimal digits)  Time : 25663.426 sec.
Starting Lucas sequence
Using AVX-512 FFT length 1600K, Pass1=1K, Pass2=1600, clm=1, 16 threads, P = 6, Q = 2
_____________...and so on...___________
Starting probable prime test of (10^8177207-1)/9
Using AVX-512 FFT length 1600K, Pass1=1K, Pass2=1600, clm=1, 16 threads, a = 7
(10^8177207-1)/9 is base 7-Fermat PRP! (8177207 decimal digits)  Time : 25609.258 sec.
Starting Lucas sequence
Using AVX-512 FFT length 1600K, Pass1=1K, Pass2=1600, clm=1, 16 threads, P = 6, Q = 2
_____________...and so on...___________
Starting probable prime test of (10^8177207-1)/9
Using AVX-512 FFT length 1600K, Pass1=1K, Pass2=1600, clm=1, 16 threads, a = 11
(10^8177207-1)/9 is base 11-Fermat PRP! (8177207 decimal digits)  Time : 25643.544 sec.
Starting Lucas sequence
Using AVX-512 FFT length 1600K, Pass1=1K, Pass2=1600, clm=1, 16 threads, P = 6, Q = 2
_____________...and so on...
I'll try to (as Mike likes to say) ...

...options:
FBase=
PBase=
FermatBase=
LucasBaseP=
isLucasBaseQ=
genLucasBaseQ=
generalLucasBase=
genLucasBaseP=
...
Code:
sllr  -oLucasBaseP=12 -oLucasPRPtest=1 -oBPSW=1 -d in
Starting Lucas sequence
Using FMA3 FFT length 1600K, Pass1=640, Pass2=2560, clm=1, 24 threads, P = 1, Q = 4  <<< looks like can be controlled
P = 1, Q = 4 is good: kronecker(-15,Rn) = -1
P = 6, Q = 2 is good: kronecker(28,Rn) = -1
Batalov is offline   Reply With Quote
Old 2021-05-10, 12:58   #51
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

23×397 Posts
Default

Quote:
Originally Posted by Batalov View Post
But who said that we didn't?
It was clearly a language communication issue:
"We are checking SPRP in 3-4 bases first (the best we can do for this PRP.)"

I understood "(the best we can do for this PRP.)" as "That is all we can do for this PRP, only SPRP tests and nothing else."
ATH is offline   Reply With Quote
Old 2021-05-10, 15:43   #52
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

17×563 Posts
Default

Dear Andreas, yes, it is a language communication issue.
Do you disagree with the phrase: "say anything, - and you will be misinterpreted by at least one"?
Or in other phrasing "A thought, once spoken, is a lie." (Fyodor Tyutchev, translated by Vladimir Nabokov)

At least the thread title, I hope, is unequivocal.
Batalov is offline   Reply With Quote
Old 2021-05-11, 22:46   #53
GP2
 
GP2's Avatar
 
Sep 2003

5·11·47 Posts
Default

Congratulations. I wonder, in the course of your search are you recording residues and factors for posterity?

Last fiddled with by GP2 on 2021-05-11 at 22:47
GP2 is offline   Reply With Quote
Old 2021-05-12, 12:02   #54
Dr Sardonicus
 
Dr Sardonicus's Avatar
 
Feb 2017
Nowhere

2×11×227 Posts
Default

Quote:
Originally Posted by GP2 View Post
Congratulations. I wonder, in the course of your search are you recording residues and factors for posterity?
As previously mentioned in this post, the Details section on this page, which Batalov supplied in this post, has that information.
Dr Sardonicus is online now   Reply With Quote
Old 2021-05-12, 16:32   #55
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

17·563 Posts
Arrow

Discussion of definitions (once again supporting the earlier observation that "A thought, once spoken, is a lie") is moved to a separate thread.
Batalov is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Generalized Repunit primes Bob Underwood Math 12 2020-10-11 20:01
Some CADO-NFS Work At Around 175-180 Decimal Digits EdH CADO-NFS 127 2020-10-07 01:47
Integers congruent to last two decimal digits mod 23 enzocreti enzocreti 1 2020-03-03 18:38
Twin Primes with 128 Decimal Digits tuckerkao Miscellaneous Math 2 2020-02-16 06:23
records for primes 3.14159 Information & Answers 8 2018-12-09 00:08

All times are UTC. The time now is 13:12.


Sat Oct 23 13:12:56 UTC 2021 up 92 days, 7:41, 0 users, load averages: 1.12, 1.13, 1.17

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.