# Efficient method to compute $\mathrm{gcd}(2^n-1,n!)$

How can we compute the value of $\mathrm{gcd}(2^n-1,n!)$ efficiently where $n$ is very large? I couldn't think of any fast and efficient method.

-
How large is very large? n=1000? n=1000000? – TonyK Jan 31 '11 at 18:54
TonyK : ~10^5 . . – Prasoon Saurav Feb 1 '11 at 3:10

One crude way uses Fermat's little theorem. $2^{p-1} \equiv 1 \mod p$ so that $p$ divides $2^p - 1$ only when $p-1$ divides $n$. For general integers $m$, Fermat's little theorem says $2^{\phi(m)} - 1 \equiv 0 \mod m$ where $\phi(\cdot)$ is the Euler phi function. So maybe the problem boils down to finding $m$ such that $\phi(m)$ divides $n$.

-
Your second sentence is true only if $2$ is a primitive root mod $p$. (I'm assuming that $2^p$ is a typo for $2^n$.) For instance, $7$ divides $2^3-1$, but $6$ doesn't divide $3$. – TonyK Feb 1 '11 at 12:57
@TonyK: I'm not sure I understand your counterexample. 2 is not a primitive root modulo 7, as you have shown, yet $7 \mid 2^6 - 1$. The reason this doesn't hold modulo 6 is because $2 \not\in \mathbb{Z}_6^*$. – Zach Langley Mar 28 '11 at 10:45
Perhaps you meant to say this only holds when $\gcd(2, m) = 1$? – Zach Langley Mar 28 '11 at 11:02

Here's what I would do:

Initialise a large integer $N = 2^n - 1$. (This requires 1 or 2 kilobytes.)

For each prime $p \le n$, compute $2^n$ mod $p$ = $2^{n \mod (p-1)}$ mod $p$, using a square-and-multiply algorithm (see this Wikipedia article for details). If the result is $1$, then $N$ is divisible by $p$, and we want the highest power of $p$ that divides both $N$ and $n!$

The highest power of $p$ that divides $n!$ is just $a = \sum_{r=1}^K \lfloor \frac{n}{p^r}\rfloor$, where $K$ is any integer such that $p^K>n$ (see e.g. this article). So try to divide $N$ by $p$ up to $a$ times, stopping if the remainder is non-zero. The number of times that this succeeds is the power that we want.

Now just multiply together all these prime powers.

Note that $N$ gets smaller and smaller as the computation proceeds, limiting the amount of arithmetic required. It might be faster to process the primes in reverse order, to accelerate this trend; experimentation will tell.

-
I'm out of votes for today; otherwise I'd vote for this one. – Carl Brannen Mar 28 '11 at 9:04