pub fn root<F: Float>(n: F, y: F) -> F
The real nth root of y, computed without rounding 1/n first. A negative y has one only for an odd whole n. n = 0, an infinite n and NaN give NaN.