Search the web
Sign In
New User? Sign Up
primenumbers · Prime numbers and primality testing
? Already a member? Sign in to Yahoo!

Yahoo! Groups Tips

Did you know...
Show off your group to the world. Share a photo of your group with us.

Best of Y! Groups

   Check them out and nominate your group.
Having problems with message search? Fill out this form to ensure your group is one of the first to be migrated to the new message search system.

Messages

  Messages Help
Advanced
Another request   Message List  
Reply | Forward Message #20577 of 21096 |
Re: Small prime divisors of very large numbers

--- In primenumbers@yahoogroups.com,
"David Broadhurst" <d.broadhurst@...> wrote:

> it's notable that the pop-in, pop-out, pop-in prime
> 3686183, in Richard Heylen's message
> http://tech.groups.yahoo.com/group/primenumbers/message/20545
> is one for which one needs to be more careful

So let's go very slowly and hope not to fool ourselves.
I shall use "znorder" in preference to "eulerphi",
since then we shall see just how pretty Richard's prime is:

p=3686183;
a=Mod(137,p);q=znorder(a);
b=Mod(137,q);r=znorder(b);
c=Mod(137,r);s=znorder(c);
d=Mod(137,s);t=znorder(d);
e=Mod(137,t);u=znorder(e);
print([e,u])

[Mod(137, 2194), 137]

Aha! The order of 137 modulo 2194 is 137,
since the smallest integer u for which
137^u = 1 mod 2194
is u = 137. How nice. No problem living with that.

Now let's use Richard's constant, k = 2136727. He claims that

137^(137^(137^137)) + k is divisible by p
137^(137^(137^(137^137))) + k is NOT divisible by p
137^(137^(137^(137^(137^137)))) + k is divisible by p

We can quickly check those 3 claims:

k=2136727;
print(a^lift(b^lift(c^137))+k)
Mod(0, 3686183)

print(a^lift(b^lift(c^lift(d^137)))+k)
Mod(1803441, 3686183)

print(a^lift(b^lift(c^lift(d^lift(e^137))))+k)
Mod(0, 3686183)

But what happens at subsequent powerings?
It's clear: e = Mod(137, 2194) has order u = 137. Hence
lift(e^137) = 1
lift(e^(137^137)) = 1
lift(e^(137^(137^137))) = 1
and so on, ad infinitum.

Here there is no "gcd problem".
Richard hit a infinite gold mine of proofs:
for any number of 137's greater than 5,
the "powering" gives the same residue, modulo p, namely
a^lift(b^lift(c^137)),
obtained when the number of 137's is n=4.

Note that p did not really "pop in" at n=4.
We deliberately chose k to make p a divisor at n=4.
Then it *must* pop back at n=6, and stay for *ever* because
137^137 = 1 mod t, where
t = z(z(z(z(p)))), with
z(x) = znorder(Mod(137,x))

Check:

p=3686183;
z(x)=znorder(Mod(137,x))
t=z(z(z(z(p))));
if(Mod(137,t)^137==1,print("Alles klar"))
Alles klar

Finally, let's prove that "my" prime stays for ever:

print(z(z(z(z(z(z(67525153)))))))
1

David





Fri Jul 3, 2009 3:40 pm

djbroadhurst
Offline Offline
Send Email Send Email

Forward
Message #20577 of 21096 |
Expand Messages Author Sort by Date

... In the middle of working on something quite different I had an odd feeling of H&W frowning at me from the bookshelf. I chose base 137 hoping to miss the...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 3, 2009
12:10 pm

... So let's go very slowly and hope not to fool ourselves. I shall use "znorder" in preference to "eulerphi", since then we shall see just how pretty...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 3, 2009
3:40 pm

... The next step is to look at cases with a Hardy-Wright "gcd problem" that cannot be solved so easily and then solve them with the Chinese Remainder Theorem...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 4, 2009
12:49 am

... Isn't it irritating to work out the right way of doing something and not being able to show the wrong way failing for the base of interest! What's...
richard_in_reading
richard_in_r...
Offline Send Email
Jul 4, 2009
5:48 am

... Simply ask Pari-GP for Mod(137,1+808*137^138)^(137^137) Puzzle: Find last 10 decimal digits of the residue of ...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 4, 2009
10:15 am

... Ok. I'm probably going to blot my copybook here but I think the last digits of 137^^k mod p are k=1 0000000137 k=2 7801462217 k=3 0547325344 k=4 5375909947...
richard_in_reading
richard_in_r...
Offline Send Email
Jul 4, 2009
4:58 pm

... The right answer by a bogus method :-) My mathematically legal method involves 3 applications of the CRT, in this "znorder" chain: z(x)=znorder(Mod(137,x))...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 4, 2009
11:15 pm

... {mypmod(b,n,p)=local(m=[p],f=0); while(n=n-1,m=concat(eulerphi(m[1]),m)); for(p=n=1,#m,n=lift(Mod(b,m[p])^n); ...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 5, 2009
11:59 am

... For information, here is the GP version that I used: GP/PARI CALCULATOR Version 2.4.2 (development) amd64 running linux (x86-64 kernel) 64-bit version ...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 5, 2009
2:14 pm

... Richard gave 3 nice examples, on whose cyclotomy I here remark, using Phi(n,x) to denote the n-th cyclotomic polynomial in x. 1) Base-2 powerings modulo...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 5, 2009
11:01 pm

... To find a gap of 5 (what Richard calls a difference of 6), one might build on the prime of Western and Seelhoff (from way back in 1886) ...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 6, 2009
12:22 am

... I gave the wrong URL. Here is the correct one http://primes.utm.edu/primes/page.php?id=6773 with apologies to Gennady Nickolaevich Gusev. Moreover, there...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 6, 2009
1:02 am

... Indeed it does, Richard. Simply use this easily found seed: p=5572*3^33+1; print(factor(znorder(Mod(3,p)))) Mat([3, 29]) noting that 29 > 3^3. David...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 6, 2009
2:28 am

... To get the same gap in base 5, use print(factor(znorder(Mod(5,14*5^3483+1)))) Mat([5, 3483]) noting that 3483 > 5^5. David...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 6, 2009
3:55 am

... It is not hard to find a bigger gap for base 137: Mod(137,78*137^192+1) has order 137^192 > 137^137 It is not hard to find a bigger gap for base 1033: b =...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 6, 2009
8:11 pm

... With b = 3581, n = 8001, the 28436-digit prime p = 2*b^n + 1: http://primes.utm.edu/primes/page.php?id=5735 found by René Dohmen in March 2000, neatly...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 8, 2009
5:01 am

... I think it gives that 2^^x+5 is divisible by 7 for all odd x > 1. [I stopped at x=1000.] Mike...
Mike Oakes
mikeoakes2
Offline Send Email
Jul 3, 2009
8:25 am

... Good old PFGW tells me 2^^x+5 is divisible by 7 for all x such that x = 3 * k + 1 with k>=0 Lélio, still illiterate in Pari-GP [Non-text portions of this...
Lélio Ribeiro de P...
lelio_73
Offline Send Email
Jul 3, 2009
10:10 am

... Dear Lélio You have been misled by an unusual notation that I have tried to avoid. The question was about the divisibility by 7 of this sequence 2 + 5 2^2...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 3, 2009
11:45 am

... The proof below can be used to show, for instance, that the seven primes (2, 3, 5, 29, 821, 23339, 67525153) that divided both 137^(137^(137^(137^137))) +...
Mark Underwood
marku606
Offline Send Email
Jul 2, 2009
11:02 pm

... Theorem: if k + m has a factor p and if k^n + m has the same factor p, then k^n^n + m will have the same factor of p. In other words, we theorize that if...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 3, 2009
1:06 am

... Maybe I'm looking at this the wrong way up, but I don't see how your Theorem (which I agree is true) helps prove that the seven primes that divide...
Richard FitzHugh
mad37wriggle
Offline Send Email
Jul 3, 2009
3:55 am

... Richard and my messages got crossed in a private email exchange just minutes ago I think. And my last message to the board in response to his private email...
Mark Underwood
marku606
Offline Send Email
Jul 3, 2009
4:40 am

... Richard and my messages just got crossed in private email. Here is what I sent him just a minute ago: ************* Hi Richard, You were not getting lost...
Mark Underwood
marku606
Offline Send Email
Jul 3, 2009
4:59 am

... Yes. I underwrote the theorem after careful reading. ... Yes. Thanks to Richard F for spotting that. Casualties so far: 1) Mark R, Lelio and Norman for not...
David Broadhurst
djbroadhurst
Offline Send Email
Jul 3, 2009
9:57 am

... Note that 73 and 137 are the smallest prime factors of 10^(8n-4) +1. They are also prime factors of 10^8n -1, where n is a positive integer. Useful? ...
Patrick Capelle
conjectureprime
Offline Send Email
Jul 4, 2009
7:42 am
 First  |  |  Next > Last 
Advanced

Copyright © 2009 Yahoo! Inc. All rights reserved.
Privacy Policy - Terms of Service - Guidelines - Help