Average Error: 40.9 → 0.3
Time: 3.9s
Precision: binary64
\[\]
↓
\[\]
↓
double code(double x) {
	return ((double) sqrt(((double) (((double) (((double) exp(((double) (2.0 * x)))) - 1.0)) / ((double) (((double) exp(x)) - 1.0))))));
}

↓
double code(double x) {
	double VAR;
	if ((x <= -3.597097828746943e-08)) {
		VAR = ((double) sqrt(((double) (((double) (((double) pow(((double) exp(x)), 2.0)) - 1.0)) / ((double) (((double) (((double) pow(((double) exp(x)), 2.0)) - ((double) (1.0 * 1.0)))) / ((double) (((double) exp(x)) + 1.0))))))));
	} else {
		VAR = ((double) sqrt(((double) (x + ((double) (2.0 + ((double) (x * ((double) (x * 0.5))))))))));
	}
	return VAR;
}

Error
Bits error versus x
Try it out
Your Program's Arguments
Results
Enter valid numbers for all inputs
Derivation
  1. Split input into 2 regimes
  2. if  x  < -3.5970978287469428e-8
    1. Initial program 0.2
      \[\]
    2. Simplified0.2
      \[\leadsto \]
    3. Using strategy rm
    4. Applied flip--0.0
      \[\leadsto \]
    5. Simplified0.0
      \[\leadsto \]
    if -3.5970978287469428e-8 <  x 
    1. Initial program 62.0
      \[\]
    2. Simplified61.6
      \[\leadsto \]
    3. Taylor expanded around 0 0.5
      \[\leadsto \]
    4. Simplified0.5
      \[\leadsto \]
  3. Recombined 2 regimes into one program.
  4. Final simplification0.3
    \[\leadsto \]
Reproduce
herbie shell --seed 2020192 
(FPCore (x)
  :name "sqrtexp (problem 3.4.4)"
  :precision binary64
  (sqrt (/ (- (exp (* 2.0 x)) 1.0) (- (exp x) 1.0))))