\frac{e^{x} - 1}{x}\begin{array}{l}
\mathbf{if}\;x \le -1.4843984217390087 \cdot 10^{-4}:\\
\;\;\;\;\frac{\frac{{\left(e^{2}\right)}^{x} - 1 \cdot 1}{e^{x} + 1}}{x}\\
\mathbf{else}:\\
\;\;\;\;\frac{1}{6} \cdot {x}^{2} + \left(\frac{1}{2} \cdot x + 1\right)\\
\end{array}double code(double x) {
return ((exp(x) - 1.0) / x);
}
double code(double x) {
double VAR;
if ((x <= -0.00014843984217390087)) {
VAR = (((pow(exp(2.0), x) - (1.0 * 1.0)) / (exp(x) + 1.0)) / x);
} else {
VAR = ((0.16666666666666666 * pow(x, 2.0)) + ((0.5 * x) + 1.0));
}
return VAR;
}




Bits error versus x
Results
| Original | 40.2 |
|---|---|
| Target | 40.7 |
| Herbie | 0.3 |
if x < -0.00014843984217390087Initial program 0.0
rmApplied flip--0.1
rmApplied *-un-lft-identity0.1
Applied exp-prod0.1
Applied *-un-lft-identity0.1
Applied exp-prod0.1
Applied pow-prod-down0.0
Simplified0.0
if -0.00014843984217390087 < x Initial program 60.1
Taylor expanded around 0 0.5
Final simplification0.3
herbie shell --seed 2020091
(FPCore (x)
:name "Kahan's exp quotient"
:precision binary64
:herbie-target
(if (and (< x 1) (> x -1)) (/ (- (exp x) 1) (log (exp x))) (/ (- (exp x) 1) x))
(/ (- (exp x) 1) x))