# How to find a primitive root of N

Show 40 post(s) from this thread on one page
Page 1 of 2 12 Last
• May 13th 2008, 06:19 PM
fanfan1609
How to find a primitive root of N
Let N have primitive roots .Which the way does find all primitive roots generally ? I think if testing all i ,(i,N)=1, is so long
THanks
• May 13th 2008, 07:14 PM
ThePerfectHacker
Quote:

Originally Posted by fanfan1609
Let N have primitive roots .Which the way does find all primitive roots generally ? I think if testing all i ,(i,N)=1, is so long
THanks

There is no known way to find primitive roots. There are methods which make it faster. Of course, a computer program is the best way to go. See this.
• May 13th 2008, 08:49 PM
fanfan1609
Quote:

Originally Posted by ThePerfectHacker
There is no known way to find primitive roots. There are methods which make it faster. Of course, a computer program is the best way to go. See this.

I also know that way .YOu can help me my problem .
find the primitive root of 11^2 if we know 2,3 is primitive root of 11.
My teacher said that "if p is primitive root of N ,or p or p+N is primitive root of N^2 " but he didn't said the general primitive root of 11^2.
You can help me ?
Thanks
• May 13th 2008, 09:08 PM
ThePerfectHacker
For 11 you can use a simple fact that if $p$ is a odd prime and $(p-1)/2$ is also a prime then $2$ is a primitive root for $p$. I leave that for you to prove.
• May 13th 2008, 10:38 PM
fanfan1609
Quote:

Originally Posted by ThePerfectHacker
For 11 you can use a simple fact that if $p$ is a odd prime and $(p-1)/2$ is also a prime then $2$ is a primitive root for $p$. I leave that for you to prove.

maybe i am not smart enough to know your advice.YOu can tell clearly .
I know if p is primitive root modulo n ,p^phi(n) =1 mod(n) ,
so how do test all i ,i<11 and (i,11)=1?
• May 14th 2008, 06:59 AM
ThePerfectHacker
Quote:

Originally Posted by fanfan1609
maybe i am not smart enough to know your advice.YOu can tell clearly .
I know if p is primitive root modulo n ,p^phi(n) =1 mod(n) ,
so how do test all i ,i<11 and (i,11)=1?

The direct way with to deal with $11$ is to list all $n$ with $\gcd (n,11)=1$ so $n=1,2,...,10$. And now eliminate each of these numbers. For example, take $n=4$. Check whether the order of $4$ mod $11$ is $\phi(11) = 10$. Remember order means the smallest exponent $k$ so that $4^k\equiv 1(\bmod 11)$. But since $4^{10}\equiv 1(\bmod p)$ (Fermat's theorem) it follows by properties of orders that $k|10$ so $k=1,2,5,10$. Since it cannot be $1$ it means you just need to check $k=2,5,10$. A simple calculation will show that $k=5$ is the order, so $4$ is not a primitive root. So that eliminates one number on the list $\{ 1,2,...,10\}$. Now pick another number, say $2$, using the approach as above you will find that $k=10$. So that means $2$ is a primitive root mod $11$. Now you can save yourself a lot of time, since you found one primitive root all other primitive roots are $2^{m}$ where $1\leq 1\leq 10$ and $\gcd(m,\phi(11)) = \gcd(m,10)=1$. In this case $m=1,3,7,9$. Thus, $2, 2^3 \equiv 8, 2^7\equiv 7, 2^9\equiv 6$ are the primitive roots.
• May 14th 2008, 03:17 PM
fanfan1609
Quote:

Originally Posted by ThePerfectHacker
The direct way with to deal with $11$ is to list all $n$ with $\gcd (n,11)=1$ so $n=1,2,...,10$. And now eliminate each of these numbers. For example, take $n=4$. Check whether the order of $4$ mod $11$ is $\phi(11) = 10$. Remember order means the smallest exponent $k$ so that $4^k\equiv 1(\bmod 11)$. But since $4^{10}\equiv 1(\bmod p)$ (Fermat's theorem) it follows by properties of orders that $k|10$ so $k=1,2,5,10$. Since it cannot be $1$ it means you just need to check $k=2,5,10$. A simple calculation will show that $k=5$ is the order, so $4$ is not a primitive root. So that eliminates one number on the list $\{ 1,2,...,10\}$. Now pick another number, say $2$, using the approach as above you will find that $k=10$. So that means $2$ is a primitive root mod $11$. Now you can save yourself a lot of time, since you found one primitive root all other primitive roots are $2^{m}$ where $1\leq 1\leq 10$ and $\gcd(m,\phi(11)) = \gcd(m,10)=1$. In this case $m=1,3,7,9$. Thus, $2, 2^3 \equiv 8, 2^7\equiv 7, 2^9\equiv 6$ are the primitive roots.

oh .i see.Thanks very much?
• May 14th 2008, 03:19 PM
ThePerfectHacker
Quote:

Originally Posted by fanfan1609
oh .i see.Thanks very much?

In my other posts. I say that if $(p-1)/2$ is also a prime then $2$ is a primitive root of $p$. Say $p=11$ then $(11-1)/2=5$ is prime, which means $2$ is primitive root of $11$. Now since you one primitive root you found them all.
• May 14th 2008, 03:27 PM
fanfan1609
Quote:

Originally Posted by ThePerfectHacker
In my other posts. I say that if $(p-1)/2$ is also a prime then $2$ is a primitive root of $p$. Say $p=11$ then $(11-1)/2=5$ is prime, which means $2$ is primitive root of $11$. Now since you one primitive root you found them all.

if i don't misunderstand ,the way you say is if i replace 2 by another number i and $(p-1)/i$ is prime ,then i is primitive root of n .
Do i have mistake ?
• May 14th 2008, 04:08 PM
ThePerfectHacker
Quote:

Originally Posted by fanfan1609
if i don't misunderstand ,the way you say is if i replace 2 by another number i and $(p-1)/i$ is prime ,then i is primitive root of n .
Do i have mistake ?

If $(p-1)/i$ is prime DOES NOT mean $i$ is primitive roots of $p$. That only works for the case $(p-1)/2$. I never said this in generality.
• Jun 14th 2008, 07:42 AM
duggaboy
Can you make a general statment or "rule" to determining if g^k is a primitive root? Is there a generalization?
• Jun 14th 2008, 08:01 AM
fanfan1609
Quote:

Originally Posted by duggaboy
Can you make a general statment or "rule" to determining if g^k is a primitive root? Is there a generalization?

If you want to determine a is a primitive root modulo n ,i think you should do something :
1/you find phi(n)
2/determine d with d|n
3/you test a^d mod n if the result isn't 1 with every d ,a is a primitive root modulo n
if you want find other primitive root ,you find e with e|phi(phi(n)) ,
with each e ,a^e mod n is a primitive root modulo n
• Jun 14th 2008, 08:27 AM
duggaboy
but how would you find d? just choose vales and try?
• Jun 14th 2008, 03:17 PM
fanfan1609
Quote:

Originally Posted by duggaboy
but how would you find d? just choose vales and try?

it is easy to find d.For example,if n is 10 ,then d is 1,2,and 5.If with every d ,which is diffirent n, a^d mod n is not 1 ,then a is a primitive root modulo n
• Jun 14th 2008, 05:46 PM
ThePerfectHacker
Quote:

Originally Posted by duggaboy
Can you make a general statment or "rule" to determining if g^k is a primitive root? Is there a generalization?

If $g$ is primitive root mod $n$ then $g^k$ is primitive root iff $\gcd(k,\phi(n)) = 1$.
Show 40 post(s) from this thread on one page
Page 1 of 2 12 Last