How to calculate (a power b) mod p correctly if b is exceeding long long integer limit in c++.
# | User | Rating |
---|---|---|
1 | tourist | 3993 |
2 | jiangly | 3743 |
3 | orzdevinwang | 3707 |
4 | Radewoosh | 3627 |
5 | jqdai0815 | 3620 |
6 | Benq | 3564 |
7 | Kevin114514 | 3443 |
8 | ksun48 | 3434 |
9 | Rewinding | 3397 |
10 | Um_nik | 3396 |
# | User | Contrib. |
---|---|---|
1 | cry | 167 |
2 | Um_nik | 163 |
3 | maomao90 | 162 |
3 | atcoder_official | 162 |
5 | adamant | 159 |
6 | -is-this-fft- | 158 |
7 | awoo | 156 |
8 | TheScrasse | 154 |
9 | Dominater069 | 153 |
9 | nor | 153 |
How to calculate (a power b) mod p correctly if b is exceeding long long integer limit in c++.
Name |
---|
O(log2(b))?
yup!
Let b = b1b2.....bn(concatenated) By induction, Let x = a ^ (b1b2...bk) (mod p)
a ^ (b1b2....b(k+1)) = ( a^(b1b2....bk) ) ^ 10 * (a ^ b(k+1)) = x^10 * a^b(k+1)
now you can calculate modulo without overflow
got it! Thanks
first i need to store b. But b is exceeding the integer limit! How can i store b so that i correctly calculate (a power b) mod p. p is prime.
If b is in the form of a string, then we can calculate with the following code:
Now assuming that a is nonzero.
EDIT: This is exactly what DongwonShin wrote above. If b is being calculated via a DP table, then you can just take all values mod p - 1 and there won't be overflow.
got it. thanks!
but why p-1?
Fermat's theorem tells us that if (a, p) = 1. So we can take the power by the modulo.
Thanks man!
In general if you have two co prime integers m and n, then m^p modulo n = m^(p%f(n)) modulo n, where f(n) is Euler totient function This.
For any prime n, f(n)=n-1.So, the Fermat's theorem follows from this.
I am using precomputed factorial and inverse factorial array to calculate b so should i use n-1 as the mod there also?
What if b is being calculated through nCr using precomputed factorial and inverse modulo arrays?Then also we've to use same mod p-1 ?
As long as your base(b) and p are co prime, I don't see why it can't be done. So, yes use same mod (p-1).
I tried but then for example 11*(inv[11)) didn't turn out to be 1 when i used p-1 as mod but while using p as mod it did show 1. here inv[11] = power(11,mod-2,mod) where power is fast exponential function.
What if b is calculated as a-c and c requires division operation?Then while using inverse modulo for division i should use which mod?I am confused.Please help.
Refer https://www.geeksforgeeks.org/find-abm-where-b-is-very-large/