Rust f64::asinh

Percentage Accurate: 30.9% → 99.9%

Time: 2.4s

Alternatives: 6

Speedup: 5.4×

Specification

Math

\[\begin{array}{l} \\ \mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \end{array} \]

FPCore

(FPCore (x)
 :precision binary64
 (copysign (log (+ (fabs x) (sqrt (+ (* x x) 1.0)))) x))

C

double code(double x) {
	return copysign(log((fabs(x) + sqrt(((x * x) + 1.0)))), x);
}

Java

public static double code(double x) {
	return Math.copySign(Math.log((Math.abs(x) + Math.sqrt(((x * x) + 1.0)))), x);
}

Python

def code(x):
	return math.copysign(math.log((math.fabs(x) + math.sqrt(((x * x) + 1.0)))), x)

Julia

function code(x)
	return copysign(log(Float64(abs(x) + sqrt(Float64(Float64(x * x) + 1.0)))), x)
end

MATLAB

function tmp = code(x)
	tmp = sign(x) * abs(log((abs(x) + sqrt(((x * x) + 1.0)))));
end

Wolfram

code[x_] := N[With[{TMP1 = Abs[N[Log[N[(N[Abs[x], $MachinePrecision] + N[Sqrt[N[(N[(x * x), $MachinePrecision] + 1.0), $MachinePrecision]], $MachinePrecision]), $MachinePrecision]], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]

Initial Program: 30.9% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \end{array} \]

FPCore

(FPCore (x)
 :precision binary64
 (copysign (log (+ (fabs x) (sqrt (+ (* x x) 1.0)))) x))

C

double code(double x) {
	return copysign(log((fabs(x) + sqrt(((x * x) + 1.0)))), x);
}

Java

public static double code(double x) {
	return Math.copySign(Math.log((Math.abs(x) + Math.sqrt(((x * x) + 1.0)))), x);
}

Python

def code(x):
	return math.copysign(math.log((math.fabs(x) + math.sqrt(((x * x) + 1.0)))), x)

Julia

function code(x)
	return copysign(log(Float64(abs(x) + sqrt(Float64(Float64(x * x) + 1.0)))), x)
end

MATLAB

function tmp = code(x)
	tmp = sign(x) * abs(log((abs(x) + sqrt(((x * x) + 1.0)))));
end

Wolfram

code[x_] := N[With[{TMP1 = Abs[N[Log[N[(N[Abs[x], $MachinePrecision] + N[Sqrt[N[(N[(x * x), $MachinePrecision] + 1.0), $MachinePrecision]], $MachinePrecision]), $MachinePrecision]], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]

Alternative 1: 99.9% accurate, 1.6× speedup

Math

\[\begin{array}{l} \\ \mathsf{copysign}\left(\sinh^{-1} x, x\right) \end{array} \]

FPCore

(FPCore (x) :precision binary64 (copysign (asinh x) x))

C

double code(double x) {
	return copysign(asinh(x), x);
}

Python

def code(x):
	return math.copysign(math.asinh(x), x)

Julia

function code(x)
	return copysign(asinh(x), x)
end

MATLAB

function tmp = code(x)
	tmp = sign(x) * abs(asinh(x));
end

Wolfram

code[x_] := N[With[{TMP1 = Abs[N[ArcSinh[x], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]

Derivation

1. Initial program 30.9%

   \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]
2. Step-by-step derivation

3. Applied rewrites 99.9%

   \[\leadsto \color{blue}{\mathsf{copysign}\left(\sinh^{-1} x, x\right)} \]
4. Add Preprocessing

Alternative 2: 82.3% accurate, 1.1× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;x \leq -3.2:\\ \;\;\;\;\mathsf{copysign}\left(\log \left(-x\right), x\right)\\ \mathbf{elif}\;x \leq 1.25:\\ \;\;\;\;\mathsf{copysign}\left(x, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{copysign}\left(\log \left(x + x\right), x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x)
 :precision binary64
 (if (<= x -3.2)
   (copysign (log (- x)) x)
   (if (<= x 1.25) (copysign x x) (copysign (log (+ x x)) x))))

C

double code(double x) {
	double tmp;
	if (x <= -3.2) {
		tmp = copysign(log(-x), x);
	} else if (x <= 1.25) {
		tmp = copysign(x, x);
	} else {
		tmp = copysign(log((x + x)), x);
	}
	return tmp;
}

Java

public static double code(double x) {
	double tmp;
	if (x <= -3.2) {
		tmp = Math.copySign(Math.log(-x), x);
	} else if (x <= 1.25) {
		tmp = Math.copySign(x, x);
	} else {
		tmp = Math.copySign(Math.log((x + x)), x);
	}
	return tmp;
}

Python

def code(x):
	tmp = 0
	if x <= -3.2:
		tmp = math.copysign(math.log(-x), x)
	elif x <= 1.25:
		tmp = math.copysign(x, x)
	else:
		tmp = math.copysign(math.log((x + x)), x)
	return tmp

Julia

function code(x)
	tmp = 0.0
	if (x <= -3.2)
		tmp = copysign(log(Float64(-x)), x);
	elseif (x <= 1.25)
		tmp = copysign(x, x);
	else
		tmp = copysign(log(Float64(x + x)), x);
	end
	return tmp
end

MATLAB

function tmp_2 = code(x)
	tmp = 0.0;
	if (x <= -3.2)
		tmp = sign(x) * abs(log(-x));
	elseif (x <= 1.25)
		tmp = sign(x) * abs(x);
	else
		tmp = sign(x) * abs(log((x + x)));
	end
	tmp_2 = tmp;
end

Wolfram

code[x_] := If[LessEqual[x, -3.2], N[With[{TMP1 = Abs[N[Log[(-x)], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision], If[LessEqual[x, 1.25], N[With[{TMP1 = Abs[x], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision], N[With[{TMP1 = Abs[N[Log[N[(x + x), $MachinePrecision]], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]]]

Derivation

1. Split input into 3 regimes

2. if x < -3.2000000000000002

   1. Initial program 52.4%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]

   2. Taylor expanded in x around -inf

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(-1 \cdot x\right)}, x\right) \]
   3. Step-by-step derivation

      1. mul-1-neg N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(\mathsf{neg}\left(x\right)\right), x\right) \]

      2. lower-neg.f64 31.4

         \[\leadsto \mathsf{copysign}\left(\log \left(-x\right), x\right) \]

   4. Applied rewrites 31.4%

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(-x\right)}, x\right) \]

   if -3.2000000000000002 < x < 1.25

   1. Initial program 8.6%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]
   2. Step-by-step derivation

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\mathsf{copysign}\left(\sinh^{-1} x, x\right)} \]

   4. Taylor expanded in x around 0

      \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]
   5. Step-by-step derivation

      1. unpow1 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      2. metadata-eval 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      3. sqrt-pow1 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      4. pow2 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      5. rem-sqrt-square-rev 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      6. asinh-def-rev 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      7. sqr-abs-rev 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      8. pow2 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      9. +-commutative 99.0

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      10. +-commutative 99.0

          \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      11. pow2 99.0

          \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   6. Applied rewrites 99.0%

      \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]

   if 1.25 < x

   1. Initial program 53.9%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]

   2. Taylor expanded in x around inf

      \[\leadsto \mathsf{copysign}\left(\log \left(\left|x\right| + \color{blue}{x}\right), x\right) \]
   3. Step-by-step derivation

   4. Applied rewrites 99.1%

      \[\leadsto \mathsf{copysign}\left(\log \left(\left|x\right| + \color{blue}{x}\right), x\right) \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(\left|x\right| + x\right)}, x\right) \]

      2. lift-fabs.f64 N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(\color{blue}{\left|x\right|} + x\right), x\right) \]

      3. +-commutative N/A

         \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(x + \left|x\right|\right)}, x\right) \]

      4. rem-sqrt-square-rev N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + \color{blue}{\sqrt{x \cdot x}}\right), x\right) \]

      5. pow2 N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + \sqrt{\color{blue}{{x}^{2}}}\right), x\right) \]

      6. sqrt-pow1 N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + \color{blue}{{x}^{\left(\frac{2}{2}\right)}}\right), x\right) \]

      7. metadata-eval N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + {x}^{\color{blue}{1}}\right), x\right) \]

      8. unpow1 N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + \color{blue}{x}\right), x\right) \]

      9. lower-+.f64 99.1

         \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(x + x\right)}, x\right) \]

   6. Applied rewrites 99.1%

      \[\leadsto \color{blue}{\mathsf{copysign}\left(\log \left(x + x\right), x\right)} \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 3: 65.1% accurate, 1.1× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;x \leq -3.2:\\ \;\;\;\;\mathsf{copysign}\left(\log \left(-x\right), x\right)\\ \mathbf{elif}\;x \leq 1.6:\\ \;\;\;\;\mathsf{copysign}\left(x, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{copysign}\left(\log \left(x - -1\right), x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x)
 :precision binary64
 (if (<= x -3.2)
   (copysign (log (- x)) x)
   (if (<= x 1.6) (copysign x x) (copysign (log (- x -1.0)) x))))

C

double code(double x) {
	double tmp;
	if (x <= -3.2) {
		tmp = copysign(log(-x), x);
	} else if (x <= 1.6) {
		tmp = copysign(x, x);
	} else {
		tmp = copysign(log((x - -1.0)), x);
	}
	return tmp;
}

Java

public static double code(double x) {
	double tmp;
	if (x <= -3.2) {
		tmp = Math.copySign(Math.log(-x), x);
	} else if (x <= 1.6) {
		tmp = Math.copySign(x, x);
	} else {
		tmp = Math.copySign(Math.log((x - -1.0)), x);
	}
	return tmp;
}

Python

def code(x):
	tmp = 0
	if x <= -3.2:
		tmp = math.copysign(math.log(-x), x)
	elif x <= 1.6:
		tmp = math.copysign(x, x)
	else:
		tmp = math.copysign(math.log((x - -1.0)), x)
	return tmp

Julia

function code(x)
	tmp = 0.0
	if (x <= -3.2)
		tmp = copysign(log(Float64(-x)), x);
	elseif (x <= 1.6)
		tmp = copysign(x, x);
	else
		tmp = copysign(log(Float64(x - -1.0)), x);
	end
	return tmp
end

MATLAB

function tmp_2 = code(x)
	tmp = 0.0;
	if (x <= -3.2)
		tmp = sign(x) * abs(log(-x));
	elseif (x <= 1.6)
		tmp = sign(x) * abs(x);
	else
		tmp = sign(x) * abs(log((x - -1.0)));
	end
	tmp_2 = tmp;
end

Wolfram

code[x_] := If[LessEqual[x, -3.2], N[With[{TMP1 = Abs[N[Log[(-x)], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision], If[LessEqual[x, 1.6], N[With[{TMP1 = Abs[x], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision], N[With[{TMP1 = Abs[N[Log[N[(x - -1.0), $MachinePrecision]], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]]]

Derivation

1. Split input into 3 regimes

2. if x < -3.2000000000000002

   1. Initial program 52.4%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]

   2. Taylor expanded in x around -inf

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(-1 \cdot x\right)}, x\right) \]
   3. Step-by-step derivation

      1. mul-1-neg N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(\mathsf{neg}\left(x\right)\right), x\right) \]

      2. lower-neg.f64 31.4

         \[\leadsto \mathsf{copysign}\left(\log \left(-x\right), x\right) \]

   4. Applied rewrites 31.4%

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(-x\right)}, x\right) \]

   if -3.2000000000000002 < x < 1.6000000000000001

   1. Initial program 8.6%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]
   2. Step-by-step derivation

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\mathsf{copysign}\left(\sinh^{-1} x, x\right)} \]

   4. Taylor expanded in x around 0

      \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]
   5. Step-by-step derivation

      1. unpow1 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      2. metadata-eval 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      3. sqrt-pow1 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      4. pow2 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      5. rem-sqrt-square-rev 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      6. asinh-def-rev 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      7. sqr-abs-rev 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      8. pow2 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      9. +-commutative 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      10. +-commutative 98.9

          \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      11. pow2 98.9

          \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   6. Applied rewrites 98.9%

      \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]

   if 1.6000000000000001 < x

   1. Initial program 53.8%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]

   2. Taylor expanded in x around inf

      \[\leadsto \mathsf{copysign}\left(\log \left(\left|x\right| + \color{blue}{x}\right), x\right) \]
   3. Step-by-step derivation

   4. Applied rewrites 99.1%

      \[\leadsto \mathsf{copysign}\left(\log \left(\left|x\right| + \color{blue}{x}\right), x\right) \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(\left|x\right| + x\right)}, x\right) \]

      2. lift-fabs.f64 N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(\color{blue}{\left|x\right|} + x\right), x\right) \]

      3. +-commutative N/A

         \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(x + \left|x\right|\right)}, x\right) \]

      4. rem-sqrt-square-rev N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + \color{blue}{\sqrt{x \cdot x}}\right), x\right) \]

      5. pow2 N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + \sqrt{\color{blue}{{x}^{2}}}\right), x\right) \]

      6. sqrt-pow1 N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + \color{blue}{{x}^{\left(\frac{2}{2}\right)}}\right), x\right) \]

      7. metadata-eval N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + {x}^{\color{blue}{1}}\right), x\right) \]

      8. unpow1 N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(x + \color{blue}{x}\right), x\right) \]

      9. lower-+.f64 99.1

         \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(x + x\right)}, x\right) \]

   6. Applied rewrites 99.1%

      \[\leadsto \color{blue}{\mathsf{copysign}\left(\log \left(x + x\right), x\right)} \]

   7. Taylor expanded in x around -inf

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(\frac{\frac{-1}{2}}{x}\right)}, x\right) \]
   8. Step-by-step derivation

      1. lower-/.f64 0.0

         \[\leadsto \mathsf{copysign}\left(\log \left(\frac{-0.5}{\color{blue}{x}}\right), x\right) \]

   9. Applied rewrites 0.0%

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(\frac{-0.5}{x}\right)}, x\right) \]

   10. Taylor expanded in x around 0

       \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(1 + x\right)}, x\right) \]
   11. Step-by-step derivation

       1. +-commutative N/A

          \[\leadsto \mathsf{copysign}\left(\log \left(x + \color{blue}{1}\right), x\right) \]

       2. metadata-eval N/A

          \[\leadsto \mathsf{copysign}\left(\log \left(x + 1 \cdot \color{blue}{1}\right), x\right) \]

       3. fp-cancel-sign-sub-inv N/A

          \[\leadsto \mathsf{copysign}\left(\log \left(x - \color{blue}{\left(\mathsf{neg}\left(1\right)\right) \cdot 1}\right), x\right) \]

       4. metadata-eval N/A

          \[\leadsto \mathsf{copysign}\left(\log \left(x - -1 \cdot 1\right), x\right) \]

       5. metadata-eval N/A

          \[\leadsto \mathsf{copysign}\left(\log \left(x - -1\right), x\right) \]

       6. lower--.f64 31.3

          \[\leadsto \mathsf{copysign}\left(\log \left(x - \color{blue}{-1}\right), x\right) \]

   12. Applied rewrites 31.3%

       \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(x - -1\right)}, x\right) \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 4: 65.1% accurate, 1.2× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;x \leq -3.2:\\ \;\;\;\;\mathsf{copysign}\left(\log \left(-x\right), x\right)\\ \mathbf{elif}\;x \leq 3.2:\\ \;\;\;\;\mathsf{copysign}\left(x, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{copysign}\left(\log x, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x)
 :precision binary64
 (if (<= x -3.2)
   (copysign (log (- x)) x)
   (if (<= x 3.2) (copysign x x) (copysign (log x) x))))

C

double code(double x) {
	double tmp;
	if (x <= -3.2) {
		tmp = copysign(log(-x), x);
	} else if (x <= 3.2) {
		tmp = copysign(x, x);
	} else {
		tmp = copysign(log(x), x);
	}
	return tmp;
}

Java

public static double code(double x) {
	double tmp;
	if (x <= -3.2) {
		tmp = Math.copySign(Math.log(-x), x);
	} else if (x <= 3.2) {
		tmp = Math.copySign(x, x);
	} else {
		tmp = Math.copySign(Math.log(x), x);
	}
	return tmp;
}

Python

def code(x):
	tmp = 0
	if x <= -3.2:
		tmp = math.copysign(math.log(-x), x)
	elif x <= 3.2:
		tmp = math.copysign(x, x)
	else:
		tmp = math.copysign(math.log(x), x)
	return tmp

Julia

function code(x)
	tmp = 0.0
	if (x <= -3.2)
		tmp = copysign(log(Float64(-x)), x);
	elseif (x <= 3.2)
		tmp = copysign(x, x);
	else
		tmp = copysign(log(x), x);
	end
	return tmp
end

MATLAB

function tmp_2 = code(x)
	tmp = 0.0;
	if (x <= -3.2)
		tmp = sign(x) * abs(log(-x));
	elseif (x <= 3.2)
		tmp = sign(x) * abs(x);
	else
		tmp = sign(x) * abs(log(x));
	end
	tmp_2 = tmp;
end

Wolfram

code[x_] := If[LessEqual[x, -3.2], N[With[{TMP1 = Abs[N[Log[(-x)], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision], If[LessEqual[x, 3.2], N[With[{TMP1 = Abs[x], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision], N[With[{TMP1 = Abs[N[Log[x], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]]]

Derivation

1. Split input into 3 regimes

2. if x < -3.2000000000000002

   1. Initial program 52.4%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]

   2. Taylor expanded in x around -inf

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(-1 \cdot x\right)}, x\right) \]
   3. Step-by-step derivation

      1. mul-1-neg N/A

         \[\leadsto \mathsf{copysign}\left(\log \left(\mathsf{neg}\left(x\right)\right), x\right) \]

      2. lower-neg.f64 31.4

         \[\leadsto \mathsf{copysign}\left(\log \left(-x\right), x\right) \]

   4. Applied rewrites 31.4%

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{\left(-x\right)}, x\right) \]

   if -3.2000000000000002 < x < 3.2000000000000002

   1. Initial program 8.6%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]
   2. Step-by-step derivation

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\mathsf{copysign}\left(\sinh^{-1} x, x\right)} \]

   4. Taylor expanded in x around 0

      \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]
   5. Step-by-step derivation

      1. unpow1 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      2. metadata-eval 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      3. sqrt-pow1 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      4. pow2 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      5. rem-sqrt-square-rev 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      6. asinh-def-rev 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      7. sqr-abs-rev 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      8. pow2 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      9. +-commutative 98.9

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      10. +-commutative 98.9

          \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      11. pow2 98.9

          \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   6. Applied rewrites 98.9%

      \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]

   if 3.2000000000000002 < x

   1. Initial program 53.8%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]

   2. Taylor expanded in x around inf

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{x}, x\right) \]
   3. Step-by-step derivation

   4. Applied rewrites 31.3%

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{x}, x\right) \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 5: 58.8% accurate, 1.6× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;x \leq 3.2:\\ \;\;\;\;\mathsf{copysign}\left(x, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{copysign}\left(\log x, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x)
 :precision binary64
 (if (<= x 3.2) (copysign x x) (copysign (log x) x)))

C

double code(double x) {
	double tmp;
	if (x <= 3.2) {
		tmp = copysign(x, x);
	} else {
		tmp = copysign(log(x), x);
	}
	return tmp;
}

Java

public static double code(double x) {
	double tmp;
	if (x <= 3.2) {
		tmp = Math.copySign(x, x);
	} else {
		tmp = Math.copySign(Math.log(x), x);
	}
	return tmp;
}

Python

def code(x):
	tmp = 0
	if x <= 3.2:
		tmp = math.copysign(x, x)
	else:
		tmp = math.copysign(math.log(x), x)
	return tmp

Julia

function code(x)
	tmp = 0.0
	if (x <= 3.2)
		tmp = copysign(x, x);
	else
		tmp = copysign(log(x), x);
	end
	return tmp
end

MATLAB

function tmp_2 = code(x)
	tmp = 0.0;
	if (x <= 3.2)
		tmp = sign(x) * abs(x);
	else
		tmp = sign(x) * abs(log(x));
	end
	tmp_2 = tmp;
end

Wolfram

code[x_] := If[LessEqual[x, 3.2], N[With[{TMP1 = Abs[x], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision], N[With[{TMP1 = Abs[N[Log[x], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if x < 3.2000000000000002

   1. Initial program 23.1%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]
   2. Step-by-step derivation

   3. Applied rewrites 99.9%

      \[\leadsto \color{blue}{\mathsf{copysign}\left(\sinh^{-1} x, x\right)} \]

   4. Taylor expanded in x around 0

      \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]
   5. Step-by-step derivation

      1. unpow1 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      2. metadata-eval 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      3. sqrt-pow1 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      4. pow2 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      5. rem-sqrt-square-rev 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      6. asinh-def-rev 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      7. sqr-abs-rev 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      8. pow2 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      9. +-commutative 68.1

         \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      10. +-commutative 68.1

          \[\leadsto \mathsf{copysign}\left(x, x\right) \]

      11. pow2 68.1

          \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   6. Applied rewrites 68.1%

      \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]

   if 3.2000000000000002 < x

   1. Initial program 53.8%

      \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]

   2. Taylor expanded in x around inf

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{x}, x\right) \]
   3. Step-by-step derivation

   4. Applied rewrites 31.3%

      \[\leadsto \mathsf{copysign}\left(\log \color{blue}{x}, x\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 6: 52.2% accurate, 5.4× speedup

Math

\[\begin{array}{l} \\ \mathsf{copysign}\left(x, x\right) \end{array} \]

FPCore

(FPCore (x) :precision binary64 (copysign x x))

C

double code(double x) {
	return copysign(x, x);
}

Java

public static double code(double x) {
	return Math.copySign(x, x);
}

Python

def code(x):
	return math.copysign(x, x)

Julia

function code(x)
	return copysign(x, x)
end

MATLAB

function tmp = code(x)
	tmp = sign(x) * abs(x);
end

Wolfram

code[x_] := N[With[{TMP1 = Abs[x], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]

Derivation

1. Initial program 30.9%

   \[\mathsf{copysign}\left(\log \left(\left|x\right| + \sqrt{x \cdot x + 1}\right), x\right) \]
2. Step-by-step derivation

3. Applied rewrites 99.9%

   \[\leadsto \color{blue}{\mathsf{copysign}\left(\sinh^{-1} x, x\right)} \]

4. Taylor expanded in x around 0

   \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]
5. Step-by-step derivation

   1. unpow1 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   2. metadata-eval 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   3. sqrt-pow1 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   4. pow2 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   5. rem-sqrt-square-rev 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   6. asinh-def-rev 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   7. sqr-abs-rev 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   8. pow2 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   9. +-commutative 52.2

      \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   10. +-commutative 52.2

       \[\leadsto \mathsf{copysign}\left(x, x\right) \]

   11. pow2 52.2

       \[\leadsto \mathsf{copysign}\left(x, x\right) \]

6. Applied rewrites 52.2%

   \[\leadsto \mathsf{copysign}\left(\color{blue}{x}, x\right) \]
7. Add Preprocessing

Developer Target 1: 100.0% accurate, 0.4× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_0 := \frac{1}{\left|x\right|}\\ \mathsf{copysign}\left(\mathsf{log1p}\left(\left|x\right| + \frac{\left|x\right|}{\mathsf{hypot}\left(1, t\_0\right) + t\_0}\right), x\right) \end{array} \end{array} \]

FPCore

(FPCore (x)
 :precision binary64
 (let* ((t_0 (/ 1.0 (fabs x))))
   (copysign (log1p (+ (fabs x) (/ (fabs x) (+ (hypot 1.0 t_0) t_0)))) x)))

C

double code(double x) {
	double t_0 = 1.0 / fabs(x);
	return copysign(log1p((fabs(x) + (fabs(x) / (hypot(1.0, t_0) + t_0)))), x);
}

Java

public static double code(double x) {
	double t_0 = 1.0 / Math.abs(x);
	return Math.copySign(Math.log1p((Math.abs(x) + (Math.abs(x) / (Math.hypot(1.0, t_0) + t_0)))), x);
}

Python

def code(x):
	t_0 = 1.0 / math.fabs(x)
	return math.copysign(math.log1p((math.fabs(x) + (math.fabs(x) / (math.hypot(1.0, t_0) + t_0)))), x)

Julia

function code(x)
	t_0 = Float64(1.0 / abs(x))
	return copysign(log1p(Float64(abs(x) + Float64(abs(x) / Float64(hypot(1.0, t_0) + t_0)))), x)
end

Wolfram

code[x_] := Block[{t$95$0 = N[(1.0 / N[Abs[x], $MachinePrecision]), $MachinePrecision]}, N[With[{TMP1 = Abs[N[Log[1 + N[(N[Abs[x], $MachinePrecision] + N[(N[Abs[x], $MachinePrecision] / N[(N[Sqrt[1.0 ^ 2 + t$95$0 ^ 2], $MachinePrecision] + t$95$0), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]], $MachinePrecision]], TMP2 = Sign[x]}, TMP1 * If[TMP2 == 0, 1, TMP2]], $MachinePrecision]]

Reproduce

herbie shell --seed 2025112 
(FPCore (x)
  :name "Rust f64::asinh"
  :precision binary64

  :alt
  (! :herbie-platform c (let* ((ax (fabs x)) (ix (/ 1 ax))) (copysign (log1p (+ ax (/ ax (+ (hypot 1 ix) ix)))) x)))

  (copysign (log (+ (fabs x) (sqrt (+ (* x x) 1.0)))) x))