Efficient implementation of power function with argument constraints

Question

I\'m looking for efficient implementation of function pow(a, b) where a is limited to (0,1) interval and b is >=