Average Error: 16.0 → 16.0
Time: 2.3s
Precision: binary64
\[\frac{\sqrt{n \cdot s2 - s1 \cdot s1}}{n}\]
↓
\[\frac{\sqrt{n \cdot s2 - s1 \cdot s1}}{n}\]
\frac{\sqrt{n \cdot s2 - s1 \cdot s1}}{n}
↓
\frac{\sqrt{n \cdot s2 - s1 \cdot s1}}{n}
double code(double n, double s2, double s1) {
	return ((double) (((double) sqrt(((double) (((double) (n * s2)) - ((double) (s1 * s1)))))) / n));
}
↓
double code(double n, double s2, double s1) {
	return ((double) (((double) sqrt(((double) (((double) (n * s2)) - ((double) (s1 * s1)))))) / n));
}

Error

Bits error versus n

Bits error versus s2

Bits error versus s1

Try it out

Your Program's Arguments

Results

Enter valid numbers for all inputs

Derivation

  1. Initial program 16.0

    \[\frac{\sqrt{n \cdot s2 - s1 \cdot s1}}{n}\]

  2. Final simplification16.0

    \[\leadsto \frac{\sqrt{n \cdot s2 - s1 \cdot s1}}{n}\]

Reproduce

herbie shell --seed 2020153 
(FPCore (n s2 s1)
  :name "(/ (sqrt (- (* n s2) (* s1 s1))) n)"
  :precision binary64
  (/ (sqrt (- (* n s2) (* s1 s1))) n))