In the RSA algorithm , the process of" evaluate the n m of X "is often used, generally, the O (log (N) modulo repeats the square algorithm to improve efficiency.
in fact, this is the content in information security mathematical basics, which was specially written in Code is used for the test in the blog space -......
Likewise, the recursive algorithms of O (log (N) are easy to understand:
/* C */INT square (int x) {return x * X;}/* Can I use a macro here? */INT fast_mod (INT X, int N, int m) {If (n = 0) return 1; else if (N % 2 = 0) return square (fast_mod (x, n/2, m) % m; else return x * fast_mod (x, n-1, m) % m;} # pythonfast_mod = Lambda X, n, m: 1 if n = 0 else fast_mod (X, n/2, m) ** 2% M if n % 2 = 0 else x * fast_mod (x, n-1, m) % m; Scheme (define (even? N) (= (remainder N 2) 0) (define (mod-fast x n m) (define (square X) (* X )) (cond (= N 0) 1) (even? N) (remainder (square (mod-fast X (/N 2) M) (else (remainder (* X (mod-fast X (-N 1) m) (mod-fast 79 24 33 );
However, an iterative version is required. I remember that I could write N as binary (for example, n = 13,110 1) and then push it one by one.
I tried to push it from the high position to the low position, which is quite simple. Remember that B [I] is the value of the I bit, calculate a [0] from high to low using a [I] = A [I + 1] ^ 2 * x ^ B [I], and the result is displayed. But the problem is that, in order to calculate from high to low, another recursion does not seem to meet the requirements.
In turn, we had to push from low to high. This process is actually quite simple, for example:
When n = 13, that is, the binary 1101 = 1*2 ^ 3 + 1*2 ^ 2 + 0*2 ^ 1 + 1*2 ^ 0, the final result is
Ans = x ^ n % m
= X ^ (1*2 ^ 3 + 1*2 ^ 2 + 0*2 ^ 1 + 1*2 ^ 0) % m
= [X ^ (1*2 ^ 3)] * [x ^ (1*2 ^ 2)] * [(x ^ (0*2 ^ 1)] * [x ^ (1*2 ^ 0)] % m
In other words, from low to high, x ^ (bit [I] * 2 ^ I) % m is multiplied to the result in the I-th bit. Here, we can change it slightly: When bit [I] = 1, we can multiply x ^ (2 ^ I) % m. Therefore, an auxiliary variable B can be used to save x ^ (2 ^ I) % m. During each iteration, B =
B ^ 2% M.
So the implementation is easy:
/* C */INT fast_mod_iter (int x, int N, int m) {int A = 1, B = X; // when I = 0, B = x ^ (2 ^ 0) = x while (n) {If (N % 2 = 1) A = a * B % m; B = B * B % m; N/= 2;} return a ;}# pythondef fast_mod (x, n, m): a = 1 B = x while true: if n = 0: return a if n % 2 = 1: A = a * B % m B = B * B % m n/= 2; scheme (define (even? N) (= (remainder N 2) 0) (define (mod-fast-iter x n m) (define (iter a B N) (cond (= N 0) a) (even? N) (iter a (remainder (* B) M) (/N 2) (else (ITER (remainder (* a B) M) (* B) (/(-N 1) 2) (ITER 1 x n) (mod-fast-iter 79 24 33);
transferred from Daniel http://www.felix021.com/blog/read.php? 2112