First, pow(x, y) on GNU C++ 4.7 compiler calculates something like with help of Taylor series. It is not 100% accurate, for example, pow(5, 2) may return something like 25.0000000000001 or 24.9999999999999. So you can write your own function that calculates x^y for integer x and y, or you can use round function. Also some C++ compilers provide overload pow(double, long long), which will work as intended, because it is binary exponentiation and double type stores integers less than 2⁵³ without errors.

Second, you declare d as int, but 10⁴·81⁵ + 10⁴ > 10⁹, so overflow will happen.

And third, you don't check, whether d < 10⁹.

AC: 7545312