Hello Codeforces How to calculate pow(a,nCr) % p efficiently? Here 1 <= n <= 10^6, 1 <= r <= 10^6 and 1<= a <= 10^6
# | User | Rating |
---|---|---|
1 | tourist | 3845 |
2 | jiangly | 3707 |
3 | Benq | 3630 |
4 | orzdevinwang | 3573 |
5 | Geothermal | 3569 |
5 | cnnfls_csy | 3569 |
7 | jqdai0815 | 3532 |
8 | ecnerwala | 3501 |
9 | gyh20 | 3447 |
10 | Rebelz | 3409 |
# | User | Contrib. |
---|---|---|
1 | awoo | 161 |
2 | maomao90 | 160 |
3 | adamant | 157 |
4 | maroonrk | 154 |
5 | -is-this-fft- | 148 |
5 | SecondThread | 148 |
7 | Petr | 147 |
7 | atcoder_official | 147 |
9 | TheScrasse | 145 |
9 | nor | 145 |
Hello Codeforces How to calculate pow(a,nCr) % p efficiently? Here 1 <= n <= 10^6, 1 <= r <= 10^6 and 1<= a <= 10^6
Name |
---|
Sorry, i got it wrong
Hint: Use Legendre's formula and Fermat's little Theorem.
To find a^b %m when b is too large, we calculate a^(b%(m-1))%m this can be proceed using fermat little theorem.
To calculate nCr%(m-1) you can use chinese remainder theorem if m-1 is non prime and square free.
I prove a^b %m= a^(b%(m-1))%m in this way, but I am not sure whether it is correct or not.
Fermat's little theorem tells that a^(m-1) %m=1, and thus we can write b=x*(m-1)+y. Then, a^b %m=a^(x*(m-1)+y) %m= a^y * (a^(m-1))^x %m= (a^y %m) * (a^(m-1) %m)^x=a^y %m =a ^ (b%(m-1)) %m.
If this is correct, I really have never considered using Fermat's little theorem in this manner... My first reaction is to calculate the inverse of b when we need to compute a/b %m. Thank you so much for sharing such wonderful idea and extending my thought.
Can you elaborate more on how you calculte nCr%(m — 1) using chinese remainder theoreom ? Suppose that (m — 1) is non-prime.