bug500, discussion (missed optimization)

Percentage Accurate: 53.4% → 97.9%
Time: 11.6s
Alternatives: 5
Speedup: 40.6×

Specification

?
\[\begin{array}{l} \\ \log \left(\frac{\sinh x}{x}\right) \end{array} \]
(FPCore (x) :precision binary64 (log (/ (sinh x) x)))
double code(double x) {
	return log((sinh(x) / x));
}
real(8) function code(x)
    real(8), intent (in) :: x
    code = log((sinh(x) / x))
end function
public static double code(double x) {
	return Math.log((Math.sinh(x) / x));
}
def code(x):
	return math.log((math.sinh(x) / x))
function code(x)
	return log(Float64(sinh(x) / x))
end
function tmp = code(x)
	tmp = log((sinh(x) / x));
end
code[x_] := N[Log[N[(N[Sinh[x], $MachinePrecision] / x), $MachinePrecision]], $MachinePrecision]
\begin{array}{l}

\\
\log \left(\frac{\sinh x}{x}\right)
\end{array}

Sampling outcomes in binary64 precision:

Local Percentage Accuracy vs ?

The average percentage accuracy by input value. Horizontal axis shows value of an input variable; the variable is choosen in the title. Vertical axis is accuracy; higher is better. Red represent the original program, while blue represents Herbie's suggestion. These can be toggled with buttons below the plot. The line is an average while dots represent individual samples.

Accuracy vs Speed?

Herbie found 5 alternatives:

AlternativeAccuracySpeedup
The accuracy (vertical axis) and speed (horizontal axis) of each alternatives. Up and to the right is better. The red square shows the initial program, and each blue circle shows an alternative.The line shows the best available speed-accuracy tradeoffs.

Initial Program: 53.4% accurate, 1.0× speedup?

\[\begin{array}{l} \\ \log \left(\frac{\sinh x}{x}\right) \end{array} \]
(FPCore (x) :precision binary64 (log (/ (sinh x) x)))
double code(double x) {
	return log((sinh(x) / x));
}
real(8) function code(x)
    real(8), intent (in) :: x
    code = log((sinh(x) / x))
end function
public static double code(double x) {
	return Math.log((Math.sinh(x) / x));
}
def code(x):
	return math.log((math.sinh(x) / x))
function code(x)
	return log(Float64(sinh(x) / x))
end
function tmp = code(x)
	tmp = log((sinh(x) / x));
end
code[x_] := N[Log[N[(N[Sinh[x], $MachinePrecision] / x), $MachinePrecision]], $MachinePrecision]
\begin{array}{l}

\\
\log \left(\frac{\sinh x}{x}\right)
\end{array}

Alternative 1: 97.9% accurate, 0.5× speedup?

\[\begin{array}{l} x_m = \left|x\right| \\ \begin{array}{l} \mathbf{if}\;\frac{\sinh x\_m}{x\_m} \leq 1.01:\\ \;\;\;\;\left(x\_m \cdot x\_m\right) \cdot \left(0.16666666666666666 + \left(x\_m \cdot x\_m\right) \cdot \left(\left(x\_m \cdot x\_m\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right)\\ \mathbf{else}:\\ \;\;\;\;\log \sinh x\_m - \log x\_m\\ \end{array} \end{array} \]
x_m = (fabs.f64 x)
(FPCore (x_m)
 :precision binary64
 (if (<= (/ (sinh x_m) x_m) 1.01)
   (*
    (* x_m x_m)
    (+
     0.16666666666666666
     (*
      (* x_m x_m)
      (- (* (* x_m x_m) 0.0003527336860670194) 0.005555555555555556))))
   (- (log (sinh x_m)) (log x_m))))
x_m = fabs(x);
double code(double x_m) {
	double tmp;
	if ((sinh(x_m) / x_m) <= 1.01) {
		tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
	} else {
		tmp = log(sinh(x_m)) - log(x_m);
	}
	return tmp;
}
x_m = abs(x)
real(8) function code(x_m)
    real(8), intent (in) :: x_m
    real(8) :: tmp
    if ((sinh(x_m) / x_m) <= 1.01d0) then
        tmp = (x_m * x_m) * (0.16666666666666666d0 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194d0) - 0.005555555555555556d0)))
    else
        tmp = log(sinh(x_m)) - log(x_m)
    end if
    code = tmp
end function
x_m = Math.abs(x);
public static double code(double x_m) {
	double tmp;
	if ((Math.sinh(x_m) / x_m) <= 1.01) {
		tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
	} else {
		tmp = Math.log(Math.sinh(x_m)) - Math.log(x_m);
	}
	return tmp;
}
x_m = math.fabs(x)
def code(x_m):
	tmp = 0
	if (math.sinh(x_m) / x_m) <= 1.01:
		tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)))
	else:
		tmp = math.log(math.sinh(x_m)) - math.log(x_m)
	return tmp
x_m = abs(x)
function code(x_m)
	tmp = 0.0
	if (Float64(sinh(x_m) / x_m) <= 1.01)
		tmp = Float64(Float64(x_m * x_m) * Float64(0.16666666666666666 + Float64(Float64(x_m * x_m) * Float64(Float64(Float64(x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556))));
	else
		tmp = Float64(log(sinh(x_m)) - log(x_m));
	end
	return tmp
end
x_m = abs(x);
function tmp_2 = code(x_m)
	tmp = 0.0;
	if ((sinh(x_m) / x_m) <= 1.01)
		tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
	else
		tmp = log(sinh(x_m)) - log(x_m);
	end
	tmp_2 = tmp;
end
x_m = N[Abs[x], $MachinePrecision]
code[x$95$m_] := If[LessEqual[N[(N[Sinh[x$95$m], $MachinePrecision] / x$95$m), $MachinePrecision], 1.01], N[(N[(x$95$m * x$95$m), $MachinePrecision] * N[(0.16666666666666666 + N[(N[(x$95$m * x$95$m), $MachinePrecision] * N[(N[(N[(x$95$m * x$95$m), $MachinePrecision] * 0.0003527336860670194), $MachinePrecision] - 0.005555555555555556), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], N[(N[Log[N[Sinh[x$95$m], $MachinePrecision]], $MachinePrecision] - N[Log[x$95$m], $MachinePrecision]), $MachinePrecision]]
\begin{array}{l}
x_m = \left|x\right|

\\
\begin{array}{l}
\mathbf{if}\;\frac{\sinh x\_m}{x\_m} \leq 1.01:\\
\;\;\;\;\left(x\_m \cdot x\_m\right) \cdot \left(0.16666666666666666 + \left(x\_m \cdot x\_m\right) \cdot \left(\left(x\_m \cdot x\_m\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right)\\

\mathbf{else}:\\
\;\;\;\;\log \sinh x\_m - \log x\_m\\


\end{array}
\end{array}
Derivation
  1. Split input into 2 regimes
  2. if (/.f64 (sinh.f64 x) x) < 1.01000000000000001

    1. Initial program 48.6%

      \[\log \left(\frac{\sinh x}{x}\right) \]
    2. Add Preprocessing
    3. Taylor expanded in x around 0 99.7%

      \[\leadsto \color{blue}{{x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot {x}^{2} - 0.005555555555555556\right)\right)} \]
    4. Step-by-step derivation
      1. unpow299.7%

        \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
    5. Applied egg-rr99.7%

      \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
    6. Step-by-step derivation
      1. unpow299.7%

        \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
    7. Applied egg-rr99.7%

      \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + \color{blue}{\left(x \cdot x\right)} \cdot \left(0.0003527336860670194 \cdot \left(x \cdot x\right) - 0.005555555555555556\right)\right) \]
    8. Step-by-step derivation
      1. unpow299.7%

        \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
    9. Applied egg-rr99.7%

      \[\leadsto \color{blue}{\left(x \cdot x\right)} \cdot \left(0.16666666666666666 + \left(x \cdot x\right) \cdot \left(0.0003527336860670194 \cdot \left(x \cdot x\right) - 0.005555555555555556\right)\right) \]

    if 1.01000000000000001 < (/.f64 (sinh.f64 x) x)

    1. Initial program 63.3%

      \[\log \left(\frac{\sinh x}{x}\right) \]
    2. Add Preprocessing
    3. Step-by-step derivation
      1. log-div26.9%

        \[\leadsto \color{blue}{\log \sinh x - \log x} \]
      2. sub-neg26.9%

        \[\leadsto \color{blue}{\log \sinh x + \left(-\log x\right)} \]
    4. Applied egg-rr26.9%

      \[\leadsto \color{blue}{\log \sinh x + \left(-\log x\right)} \]
    5. Step-by-step derivation
      1. sub-neg26.9%

        \[\leadsto \color{blue}{\log \sinh x - \log x} \]
    6. Simplified26.9%

      \[\leadsto \color{blue}{\log \sinh x - \log x} \]
  3. Recombined 2 regimes into one program.
  4. Final simplification96.6%

    \[\leadsto \begin{array}{l} \mathbf{if}\;\frac{\sinh x}{x} \leq 1.01:\\ \;\;\;\;\left(x \cdot x\right) \cdot \left(0.16666666666666666 + \left(x \cdot x\right) \cdot \left(\left(x \cdot x\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right)\\ \mathbf{else}:\\ \;\;\;\;\log \sinh x - \log x\\ \end{array} \]
  5. Add Preprocessing

Alternative 2: 97.9% accurate, 0.7× speedup?

\[\begin{array}{l} x_m = \left|x\right| \\ \begin{array}{l} t_0 := \frac{\sinh x\_m}{x\_m}\\ \mathbf{if}\;t\_0 \leq 1.01:\\ \;\;\;\;\left(x\_m \cdot x\_m\right) \cdot \left(0.16666666666666666 + \left(x\_m \cdot x\_m\right) \cdot \left(\left(x\_m \cdot x\_m\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right)\\ \mathbf{else}:\\ \;\;\;\;\log t\_0\\ \end{array} \end{array} \]
x_m = (fabs.f64 x)
(FPCore (x_m)
 :precision binary64
 (let* ((t_0 (/ (sinh x_m) x_m)))
   (if (<= t_0 1.01)
     (*
      (* x_m x_m)
      (+
       0.16666666666666666
       (*
        (* x_m x_m)
        (- (* (* x_m x_m) 0.0003527336860670194) 0.005555555555555556))))
     (log t_0))))
x_m = fabs(x);
double code(double x_m) {
	double t_0 = sinh(x_m) / x_m;
	double tmp;
	if (t_0 <= 1.01) {
		tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
	} else {
		tmp = log(t_0);
	}
	return tmp;
}
x_m = abs(x)
real(8) function code(x_m)
    real(8), intent (in) :: x_m
    real(8) :: t_0
    real(8) :: tmp
    t_0 = sinh(x_m) / x_m
    if (t_0 <= 1.01d0) then
        tmp = (x_m * x_m) * (0.16666666666666666d0 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194d0) - 0.005555555555555556d0)))
    else
        tmp = log(t_0)
    end if
    code = tmp
end function
x_m = Math.abs(x);
public static double code(double x_m) {
	double t_0 = Math.sinh(x_m) / x_m;
	double tmp;
	if (t_0 <= 1.01) {
		tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
	} else {
		tmp = Math.log(t_0);
	}
	return tmp;
}
x_m = math.fabs(x)
def code(x_m):
	t_0 = math.sinh(x_m) / x_m
	tmp = 0
	if t_0 <= 1.01:
		tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)))
	else:
		tmp = math.log(t_0)
	return tmp
x_m = abs(x)
function code(x_m)
	t_0 = Float64(sinh(x_m) / x_m)
	tmp = 0.0
	if (t_0 <= 1.01)
		tmp = Float64(Float64(x_m * x_m) * Float64(0.16666666666666666 + Float64(Float64(x_m * x_m) * Float64(Float64(Float64(x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556))));
	else
		tmp = log(t_0);
	end
	return tmp
end
x_m = abs(x);
function tmp_2 = code(x_m)
	t_0 = sinh(x_m) / x_m;
	tmp = 0.0;
	if (t_0 <= 1.01)
		tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
	else
		tmp = log(t_0);
	end
	tmp_2 = tmp;
end
x_m = N[Abs[x], $MachinePrecision]
code[x$95$m_] := Block[{t$95$0 = N[(N[Sinh[x$95$m], $MachinePrecision] / x$95$m), $MachinePrecision]}, If[LessEqual[t$95$0, 1.01], N[(N[(x$95$m * x$95$m), $MachinePrecision] * N[(0.16666666666666666 + N[(N[(x$95$m * x$95$m), $MachinePrecision] * N[(N[(N[(x$95$m * x$95$m), $MachinePrecision] * 0.0003527336860670194), $MachinePrecision] - 0.005555555555555556), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], N[Log[t$95$0], $MachinePrecision]]]
\begin{array}{l}
x_m = \left|x\right|

\\
\begin{array}{l}
t_0 := \frac{\sinh x\_m}{x\_m}\\
\mathbf{if}\;t\_0 \leq 1.01:\\
\;\;\;\;\left(x\_m \cdot x\_m\right) \cdot \left(0.16666666666666666 + \left(x\_m \cdot x\_m\right) \cdot \left(\left(x\_m \cdot x\_m\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right)\\

\mathbf{else}:\\
\;\;\;\;\log t\_0\\


\end{array}
\end{array}
Derivation
  1. Split input into 2 regimes
  2. if (/.f64 (sinh.f64 x) x) < 1.01000000000000001

    1. Initial program 48.6%

      \[\log \left(\frac{\sinh x}{x}\right) \]
    2. Add Preprocessing
    3. Taylor expanded in x around 0 99.7%

      \[\leadsto \color{blue}{{x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot {x}^{2} - 0.005555555555555556\right)\right)} \]
    4. Step-by-step derivation
      1. unpow299.7%

        \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
    5. Applied egg-rr99.7%

      \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
    6. Step-by-step derivation
      1. unpow299.7%

        \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
    7. Applied egg-rr99.7%

      \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + \color{blue}{\left(x \cdot x\right)} \cdot \left(0.0003527336860670194 \cdot \left(x \cdot x\right) - 0.005555555555555556\right)\right) \]
    8. Step-by-step derivation
      1. unpow299.7%

        \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
    9. Applied egg-rr99.7%

      \[\leadsto \color{blue}{\left(x \cdot x\right)} \cdot \left(0.16666666666666666 + \left(x \cdot x\right) \cdot \left(0.0003527336860670194 \cdot \left(x \cdot x\right) - 0.005555555555555556\right)\right) \]

    if 1.01000000000000001 < (/.f64 (sinh.f64 x) x)

    1. Initial program 63.3%

      \[\log \left(\frac{\sinh x}{x}\right) \]
    2. Add Preprocessing
  3. Recombined 2 regimes into one program.
  4. Final simplification98.1%

    \[\leadsto \begin{array}{l} \mathbf{if}\;\frac{\sinh x}{x} \leq 1.01:\\ \;\;\;\;\left(x \cdot x\right) \cdot \left(0.16666666666666666 + \left(x \cdot x\right) \cdot \left(\left(x \cdot x\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right)\\ \mathbf{else}:\\ \;\;\;\;\log \left(\frac{\sinh x}{x}\right)\\ \end{array} \]
  5. Add Preprocessing

Alternative 3: 97.2% accurate, 11.9× speedup?

\[\begin{array}{l} x_m = \left|x\right| \\ \left(x\_m \cdot x\_m\right) \cdot \left(0.16666666666666666 + \left(x\_m \cdot x\_m\right) \cdot \left(\left(x\_m \cdot x\_m\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right) \end{array} \]
x_m = (fabs.f64 x)
(FPCore (x_m)
 :precision binary64
 (*
  (* x_m x_m)
  (+
   0.16666666666666666
   (*
    (* x_m x_m)
    (- (* (* x_m x_m) 0.0003527336860670194) 0.005555555555555556)))))
x_m = fabs(x);
double code(double x_m) {
	return (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
}
x_m = abs(x)
real(8) function code(x_m)
    real(8), intent (in) :: x_m
    code = (x_m * x_m) * (0.16666666666666666d0 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194d0) - 0.005555555555555556d0)))
end function
x_m = Math.abs(x);
public static double code(double x_m) {
	return (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
}
x_m = math.fabs(x)
def code(x_m):
	return (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)))
x_m = abs(x)
function code(x_m)
	return Float64(Float64(x_m * x_m) * Float64(0.16666666666666666 + Float64(Float64(x_m * x_m) * Float64(Float64(Float64(x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556))))
end
x_m = abs(x);
function tmp = code(x_m)
	tmp = (x_m * x_m) * (0.16666666666666666 + ((x_m * x_m) * (((x_m * x_m) * 0.0003527336860670194) - 0.005555555555555556)));
end
x_m = N[Abs[x], $MachinePrecision]
code[x$95$m_] := N[(N[(x$95$m * x$95$m), $MachinePrecision] * N[(0.16666666666666666 + N[(N[(x$95$m * x$95$m), $MachinePrecision] * N[(N[(N[(x$95$m * x$95$m), $MachinePrecision] * 0.0003527336860670194), $MachinePrecision] - 0.005555555555555556), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]
\begin{array}{l}
x_m = \left|x\right|

\\
\left(x\_m \cdot x\_m\right) \cdot \left(0.16666666666666666 + \left(x\_m \cdot x\_m\right) \cdot \left(\left(x\_m \cdot x\_m\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right)
\end{array}
Derivation
  1. Initial program 49.2%

    \[\log \left(\frac{\sinh x}{x}\right) \]
  2. Add Preprocessing
  3. Taylor expanded in x around 0 96.3%

    \[\leadsto \color{blue}{{x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot {x}^{2} - 0.005555555555555556\right)\right)} \]
  4. Step-by-step derivation
    1. unpow296.3%

      \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
  5. Applied egg-rr96.3%

    \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
  6. Step-by-step derivation
    1. unpow296.3%

      \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
  7. Applied egg-rr96.3%

    \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + \color{blue}{\left(x \cdot x\right)} \cdot \left(0.0003527336860670194 \cdot \left(x \cdot x\right) - 0.005555555555555556\right)\right) \]
  8. Step-by-step derivation
    1. unpow296.3%

      \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
  9. Applied egg-rr96.3%

    \[\leadsto \color{blue}{\left(x \cdot x\right)} \cdot \left(0.16666666666666666 + \left(x \cdot x\right) \cdot \left(0.0003527336860670194 \cdot \left(x \cdot x\right) - 0.005555555555555556\right)\right) \]
  10. Final simplification96.3%

    \[\leadsto \left(x \cdot x\right) \cdot \left(0.16666666666666666 + \left(x \cdot x\right) \cdot \left(\left(x \cdot x\right) \cdot 0.0003527336860670194 - 0.005555555555555556\right)\right) \]
  11. Add Preprocessing

Alternative 4: 96.6% accurate, 40.6× speedup?

\[\begin{array}{l} x_m = \left|x\right| \\ \left(x\_m \cdot x\_m\right) \cdot 0.16666666666666666 \end{array} \]
x_m = (fabs.f64 x)
(FPCore (x_m) :precision binary64 (* (* x_m x_m) 0.16666666666666666))
x_m = fabs(x);
double code(double x_m) {
	return (x_m * x_m) * 0.16666666666666666;
}
x_m = abs(x)
real(8) function code(x_m)
    real(8), intent (in) :: x_m
    code = (x_m * x_m) * 0.16666666666666666d0
end function
x_m = Math.abs(x);
public static double code(double x_m) {
	return (x_m * x_m) * 0.16666666666666666;
}
x_m = math.fabs(x)
def code(x_m):
	return (x_m * x_m) * 0.16666666666666666
x_m = abs(x)
function code(x_m)
	return Float64(Float64(x_m * x_m) * 0.16666666666666666)
end
x_m = abs(x);
function tmp = code(x_m)
	tmp = (x_m * x_m) * 0.16666666666666666;
end
x_m = N[Abs[x], $MachinePrecision]
code[x$95$m_] := N[(N[(x$95$m * x$95$m), $MachinePrecision] * 0.16666666666666666), $MachinePrecision]
\begin{array}{l}
x_m = \left|x\right|

\\
\left(x\_m \cdot x\_m\right) \cdot 0.16666666666666666
\end{array}
Derivation
  1. Initial program 49.2%

    \[\log \left(\frac{\sinh x}{x}\right) \]
  2. Add Preprocessing
  3. Taylor expanded in x around 0 47.0%

    \[\leadsto \log \color{blue}{\left(1 + 0.16666666666666666 \cdot {x}^{2}\right)} \]
  4. Step-by-step derivation
    1. +-commutative47.0%

      \[\leadsto \log \color{blue}{\left(0.16666666666666666 \cdot {x}^{2} + 1\right)} \]
  5. Simplified47.0%

    \[\leadsto \log \color{blue}{\left(0.16666666666666666 \cdot {x}^{2} + 1\right)} \]
  6. Taylor expanded in x around 0 95.4%

    \[\leadsto \color{blue}{0.16666666666666666 \cdot {x}^{2}} \]
  7. Step-by-step derivation
    1. *-commutative95.4%

      \[\leadsto \color{blue}{{x}^{2} \cdot 0.16666666666666666} \]
  8. Simplified95.4%

    \[\leadsto \color{blue}{{x}^{2} \cdot 0.16666666666666666} \]
  9. Step-by-step derivation
    1. unpow296.3%

      \[\leadsto {x}^{2} \cdot \left(0.16666666666666666 + {x}^{2} \cdot \left(0.0003527336860670194 \cdot \color{blue}{\left(x \cdot x\right)} - 0.005555555555555556\right)\right) \]
  10. Applied egg-rr95.4%

    \[\leadsto \color{blue}{\left(x \cdot x\right)} \cdot 0.16666666666666666 \]
  11. Add Preprocessing

Alternative 5: 51.2% accurate, 203.0× speedup?

\[\begin{array}{l} x_m = \left|x\right| \\ 0 \end{array} \]
x_m = (fabs.f64 x)
(FPCore (x_m) :precision binary64 0.0)
x_m = fabs(x);
double code(double x_m) {
	return 0.0;
}
x_m = abs(x)
real(8) function code(x_m)
    real(8), intent (in) :: x_m
    code = 0.0d0
end function
x_m = Math.abs(x);
public static double code(double x_m) {
	return 0.0;
}
x_m = math.fabs(x)
def code(x_m):
	return 0.0
x_m = abs(x)
function code(x_m)
	return 0.0
end
x_m = abs(x);
function tmp = code(x_m)
	tmp = 0.0;
end
x_m = N[Abs[x], $MachinePrecision]
code[x$95$m_] := 0.0
\begin{array}{l}
x_m = \left|x\right|

\\
0
\end{array}
Derivation
  1. Initial program 49.2%

    \[\log \left(\frac{\sinh x}{x}\right) \]
  2. Add Preprocessing
  3. Taylor expanded in x around 0 45.4%

    \[\leadsto \log \color{blue}{1} \]
  4. Step-by-step derivation
    1. metadata-eval45.4%

      \[\leadsto \color{blue}{0} \]
  5. Applied egg-rr45.4%

    \[\leadsto \color{blue}{0} \]
  6. Add Preprocessing

Developer Target 1: 98.0% accurate, 0.5× speedup?

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;\left|x\right| < 0.085:\\ \;\;\;\;\left(x \cdot x\right) \cdot \mathsf{fma}\left(\mathsf{fma}\left(\mathsf{fma}\left(-2.6455026455026456 \cdot 10^{-5}, x \cdot x, 0.0003527336860670194\right), x \cdot x, -0.005555555555555556\right), x \cdot x, 0.16666666666666666\right)\\ \mathbf{else}:\\ \;\;\;\;\log \left(\frac{\sinh x}{x}\right)\\ \end{array} \end{array} \]
(FPCore (x)
 :precision binary64
 (if (< (fabs x) 0.085)
   (*
    (* x x)
    (fma
     (fma
      (fma -2.6455026455026456e-5 (* x x) 0.0003527336860670194)
      (* x x)
      -0.005555555555555556)
     (* x x)
     0.16666666666666666))
   (log (/ (sinh x) x))))
double code(double x) {
	double tmp;
	if (fabs(x) < 0.085) {
		tmp = (x * x) * fma(fma(fma(-2.6455026455026456e-5, (x * x), 0.0003527336860670194), (x * x), -0.005555555555555556), (x * x), 0.16666666666666666);
	} else {
		tmp = log((sinh(x) / x));
	}
	return tmp;
}
function code(x)
	tmp = 0.0
	if (abs(x) < 0.085)
		tmp = Float64(Float64(x * x) * fma(fma(fma(-2.6455026455026456e-5, Float64(x * x), 0.0003527336860670194), Float64(x * x), -0.005555555555555556), Float64(x * x), 0.16666666666666666));
	else
		tmp = log(Float64(sinh(x) / x));
	end
	return tmp
end
code[x_] := If[Less[N[Abs[x], $MachinePrecision], 0.085], N[(N[(x * x), $MachinePrecision] * N[(N[(N[(-2.6455026455026456e-5 * N[(x * x), $MachinePrecision] + 0.0003527336860670194), $MachinePrecision] * N[(x * x), $MachinePrecision] + -0.005555555555555556), $MachinePrecision] * N[(x * x), $MachinePrecision] + 0.16666666666666666), $MachinePrecision]), $MachinePrecision], N[Log[N[(N[Sinh[x], $MachinePrecision] / x), $MachinePrecision]], $MachinePrecision]]
\begin{array}{l}

\\
\begin{array}{l}
\mathbf{if}\;\left|x\right| < 0.085:\\
\;\;\;\;\left(x \cdot x\right) \cdot \mathsf{fma}\left(\mathsf{fma}\left(\mathsf{fma}\left(-2.6455026455026456 \cdot 10^{-5}, x \cdot x, 0.0003527336860670194\right), x \cdot x, -0.005555555555555556\right), x \cdot x, 0.16666666666666666\right)\\

\mathbf{else}:\\
\;\;\;\;\log \left(\frac{\sinh x}{x}\right)\\


\end{array}
\end{array}

Reproduce

?
herbie shell --seed 2024172 
(FPCore (x)
  :name "bug500, discussion (missed optimization)"
  :precision binary64

  :alt
  (! :herbie-platform default (if (< (fabs x) 17/200) (let ((x2 (* x x))) (* x2 (fma (fma (fma -1/37800 x2 1/2835) x2 -1/180) x2 1/6))) (log (/ (sinh x) x))))

  (log (/ (sinh x) x)))