Radioactive exchange between two surfaces

Percentage Accurate: 85.7% → 99.8%
Time: 6.0s
Alternatives: 6
Speedup: 7.4×

Specification

?
\[\begin{array}{l} \\ {x}^{4} - {y}^{4} \end{array} \]
(FPCore (x y) :precision binary64 (- (pow x 4.0) (pow y 4.0)))
double code(double x, double y) {
	return pow(x, 4.0) - pow(y, 4.0);
}
real(8) function code(x, y)
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    code = (x ** 4.0d0) - (y ** 4.0d0)
end function
public static double code(double x, double y) {
	return Math.pow(x, 4.0) - Math.pow(y, 4.0);
}
def code(x, y):
	return math.pow(x, 4.0) - math.pow(y, 4.0)
function code(x, y)
	return Float64((x ^ 4.0) - (y ^ 4.0))
end
function tmp = code(x, y)
	tmp = (x ^ 4.0) - (y ^ 4.0);
end
code[x_, y_] := N[(N[Power[x, 4.0], $MachinePrecision] - N[Power[y, 4.0], $MachinePrecision]), $MachinePrecision]
\begin{array}{l}

\\
{x}^{4} - {y}^{4}
\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 6 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: 85.7% accurate, 1.0× speedup?

\[\begin{array}{l} \\ {x}^{4} - {y}^{4} \end{array} \]
(FPCore (x y) :precision binary64 (- (pow x 4.0) (pow y 4.0)))
double code(double x, double y) {
	return pow(x, 4.0) - pow(y, 4.0);
}
real(8) function code(x, y)
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    code = (x ** 4.0d0) - (y ** 4.0d0)
end function
public static double code(double x, double y) {
	return Math.pow(x, 4.0) - Math.pow(y, 4.0);
}
def code(x, y):
	return math.pow(x, 4.0) - math.pow(y, 4.0)
function code(x, y)
	return Float64((x ^ 4.0) - (y ^ 4.0))
end
function tmp = code(x, y)
	tmp = (x ^ 4.0) - (y ^ 4.0);
end
code[x_, y_] := N[(N[Power[x, 4.0], $MachinePrecision] - N[Power[y, 4.0], $MachinePrecision]), $MachinePrecision]
\begin{array}{l}

\\
{x}^{4} - {y}^{4}
\end{array}

Alternative 1: 99.8% accurate, 7.4× speedup?

\[\begin{array}{l} \\ \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \left(x - y\right)\right) \end{array} \]
(FPCore (x y) :precision binary64 (* (fma y y (* x x)) (* (+ x y) (- x y))))
double code(double x, double y) {
	return fma(y, y, (x * x)) * ((x + y) * (x - y));
}
function code(x, y)
	return Float64(fma(y, y, Float64(x * x)) * Float64(Float64(x + y) * Float64(x - y)))
end
code[x_, y_] := N[(N[(y * y + N[(x * x), $MachinePrecision]), $MachinePrecision] * N[(N[(x + y), $MachinePrecision] * N[(x - y), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]
\begin{array}{l}

\\
\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \left(x - y\right)\right)
\end{array}
Derivation
  1. Initial program 86.7%

    \[{x}^{4} - {y}^{4} \]
  2. Add Preprocessing
  3. Step-by-step derivation
    1. lift--.f64N/A

      \[\leadsto \color{blue}{{x}^{4} - {y}^{4}} \]
    2. lift-pow.f64N/A

      \[\leadsto \color{blue}{{x}^{4}} - {y}^{4} \]
    3. sqr-powN/A

      \[\leadsto \color{blue}{{x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)}} - {y}^{4} \]
    4. lift-pow.f64N/A

      \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{4}} \]
    5. sqr-powN/A

      \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{\left(\frac{4}{2}\right)} \cdot {y}^{\left(\frac{4}{2}\right)}} \]
    6. difference-of-squaresN/A

      \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
    7. lower-*.f64N/A

      \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
    8. +-commutativeN/A

      \[\leadsto \color{blue}{\left({y}^{\left(\frac{4}{2}\right)} + {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    9. metadata-evalN/A

      \[\leadsto \left({y}^{\color{blue}{2}} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    10. unpow2N/A

      \[\leadsto \left(\color{blue}{y \cdot y} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    11. lower-fma.f64N/A

      \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    12. metadata-evalN/A

      \[\leadsto \mathsf{fma}\left(y, y, {x}^{\color{blue}{2}}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    13. unpow2N/A

      \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    14. lower-*.f64N/A

      \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    15. metadata-evalN/A

      \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{\color{blue}{2}} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    16. unpow2N/A

      \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{x \cdot x} - {y}^{\left(\frac{4}{2}\right)}\right) \]
    17. metadata-evalN/A

      \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - {y}^{\color{blue}{2}}\right) \]
    18. unpow2N/A

      \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - \color{blue}{y \cdot y}\right) \]
    19. difference-of-squaresN/A

      \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
    20. lower-*.f64N/A

      \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
    21. lower-+.f64N/A

      \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{\left(x + y\right)} \cdot \left(x - y\right)\right) \]
    22. lower--.f6499.8

      \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \color{blue}{\left(x - y\right)}\right) \]
  4. Applied rewrites99.8%

    \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
  5. Add Preprocessing

Alternative 2: 92.3% accurate, 0.9× speedup?

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;{x}^{4} - {y}^{4} \leq -5 \cdot 10^{-323}:\\ \;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x\right)\\ \end{array} \end{array} \]
(FPCore (x y)
 :precision binary64
 (if (<= (- (pow x 4.0) (pow y 4.0)) -5e-323)
   (* (* y y) (* (- y) y))
   (* (fma y y (* x x)) (* x x))))
double code(double x, double y) {
	double tmp;
	if ((pow(x, 4.0) - pow(y, 4.0)) <= -5e-323) {
		tmp = (y * y) * (-y * y);
	} else {
		tmp = fma(y, y, (x * x)) * (x * x);
	}
	return tmp;
}
function code(x, y)
	tmp = 0.0
	if (Float64((x ^ 4.0) - (y ^ 4.0)) <= -5e-323)
		tmp = Float64(Float64(y * y) * Float64(Float64(-y) * y));
	else
		tmp = Float64(fma(y, y, Float64(x * x)) * Float64(x * x));
	end
	return tmp
end
code[x_, y_] := If[LessEqual[N[(N[Power[x, 4.0], $MachinePrecision] - N[Power[y, 4.0], $MachinePrecision]), $MachinePrecision], -5e-323], N[(N[(y * y), $MachinePrecision] * N[((-y) * y), $MachinePrecision]), $MachinePrecision], N[(N[(y * y + N[(x * x), $MachinePrecision]), $MachinePrecision] * N[(x * x), $MachinePrecision]), $MachinePrecision]]
\begin{array}{l}

\\
\begin{array}{l}
\mathbf{if}\;{x}^{4} - {y}^{4} \leq -5 \cdot 10^{-323}:\\
\;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\

\mathbf{else}:\\
\;\;\;\;\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x\right)\\


\end{array}
\end{array}
Derivation
  1. Split input into 2 regimes
  2. if (-.f64 (pow.f64 x #s(literal 4 binary64)) (pow.f64 y #s(literal 4 binary64))) < -4.94066e-323

    1. Initial program 100.0%

      \[{x}^{4} - {y}^{4} \]
    2. Add Preprocessing
    3. Taylor expanded in x around 0

      \[\leadsto \color{blue}{-1 \cdot {y}^{4}} \]
    4. Step-by-step derivation
      1. mul-1-negN/A

        \[\leadsto \color{blue}{\mathsf{neg}\left({y}^{4}\right)} \]
      2. lower-neg.f64N/A

        \[\leadsto \color{blue}{-{y}^{4}} \]
      3. lower-pow.f64100.0

        \[\leadsto -\color{blue}{{y}^{4}} \]
    5. Applied rewrites100.0%

      \[\leadsto \color{blue}{-{y}^{4}} \]
    6. Step-by-step derivation
      1. Applied rewrites99.7%

        \[\leadsto \left(\left(-y\right) \cdot y\right) \cdot \color{blue}{\left(y \cdot y\right)} \]

      if -4.94066e-323 < (-.f64 (pow.f64 x #s(literal 4 binary64)) (pow.f64 y #s(literal 4 binary64)))

      1. Initial program 79.1%

        \[{x}^{4} - {y}^{4} \]
      2. Add Preprocessing
      3. Step-by-step derivation
        1. lift--.f64N/A

          \[\leadsto \color{blue}{{x}^{4} - {y}^{4}} \]
        2. lift-pow.f64N/A

          \[\leadsto \color{blue}{{x}^{4}} - {y}^{4} \]
        3. sqr-powN/A

          \[\leadsto \color{blue}{{x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)}} - {y}^{4} \]
        4. lift-pow.f64N/A

          \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{4}} \]
        5. sqr-powN/A

          \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{\left(\frac{4}{2}\right)} \cdot {y}^{\left(\frac{4}{2}\right)}} \]
        6. difference-of-squaresN/A

          \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
        7. lower-*.f64N/A

          \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
        8. +-commutativeN/A

          \[\leadsto \color{blue}{\left({y}^{\left(\frac{4}{2}\right)} + {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        9. metadata-evalN/A

          \[\leadsto \left({y}^{\color{blue}{2}} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        10. unpow2N/A

          \[\leadsto \left(\color{blue}{y \cdot y} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        11. lower-fma.f64N/A

          \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        12. metadata-evalN/A

          \[\leadsto \mathsf{fma}\left(y, y, {x}^{\color{blue}{2}}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        13. unpow2N/A

          \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        14. lower-*.f64N/A

          \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        15. metadata-evalN/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{\color{blue}{2}} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        16. unpow2N/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{x \cdot x} - {y}^{\left(\frac{4}{2}\right)}\right) \]
        17. metadata-evalN/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - {y}^{\color{blue}{2}}\right) \]
        18. unpow2N/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - \color{blue}{y \cdot y}\right) \]
        19. difference-of-squaresN/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
        20. lower-*.f64N/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
        21. lower-+.f64N/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{\left(x + y\right)} \cdot \left(x - y\right)\right) \]
        22. lower--.f6499.8

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \color{blue}{\left(x - y\right)}\right) \]
      4. Applied rewrites99.8%

        \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
      5. Taylor expanded in x around inf

        \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left({x}^{2} \cdot \left(1 + \left(-1 \cdot \frac{y}{x} + \frac{y}{x}\right)\right)\right)} \]
      6. Step-by-step derivation
        1. distribute-lft1-inN/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{\left(-1 + 1\right) \cdot \frac{y}{x}}\right)\right) \]
        2. metadata-evalN/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{0} \cdot \frac{y}{x}\right)\right) \]
        3. mul0-lftN/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{0}\right)\right) \]
        4. metadata-evalN/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \color{blue}{1}\right) \]
        5. *-rgt-identityN/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{{x}^{2}} \]
        6. unpow2N/A

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
        7. lower-*.f6490.6

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
      7. Applied rewrites90.6%

        \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
    7. Recombined 2 regimes into one program.
    8. Final simplification93.9%

      \[\leadsto \begin{array}{l} \mathbf{if}\;{x}^{4} - {y}^{4} \leq -5 \cdot 10^{-323}:\\ \;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x\right)\\ \end{array} \]
    9. Add Preprocessing

    Alternative 3: 92.2% accurate, 0.9× speedup?

    \[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;{x}^{4} - {y}^{4} \leq -5 \cdot 10^{-323}:\\ \;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\ \mathbf{else}:\\ \;\;\;\;\left(y - x\right) \cdot \left(\left(\left(y - x\right) \cdot x\right) \cdot x\right)\\ \end{array} \end{array} \]
    (FPCore (x y)
     :precision binary64
     (if (<= (- (pow x 4.0) (pow y 4.0)) -5e-323)
       (* (* y y) (* (- y) y))
       (* (- y x) (* (* (- y x) x) x))))
    double code(double x, double y) {
    	double tmp;
    	if ((pow(x, 4.0) - pow(y, 4.0)) <= -5e-323) {
    		tmp = (y * y) * (-y * y);
    	} else {
    		tmp = (y - x) * (((y - x) * x) * x);
    	}
    	return tmp;
    }
    
    real(8) function code(x, y)
        real(8), intent (in) :: x
        real(8), intent (in) :: y
        real(8) :: tmp
        if (((x ** 4.0d0) - (y ** 4.0d0)) <= (-5d-323)) then
            tmp = (y * y) * (-y * y)
        else
            tmp = (y - x) * (((y - x) * x) * x)
        end if
        code = tmp
    end function
    
    public static double code(double x, double y) {
    	double tmp;
    	if ((Math.pow(x, 4.0) - Math.pow(y, 4.0)) <= -5e-323) {
    		tmp = (y * y) * (-y * y);
    	} else {
    		tmp = (y - x) * (((y - x) * x) * x);
    	}
    	return tmp;
    }
    
    def code(x, y):
    	tmp = 0
    	if (math.pow(x, 4.0) - math.pow(y, 4.0)) <= -5e-323:
    		tmp = (y * y) * (-y * y)
    	else:
    		tmp = (y - x) * (((y - x) * x) * x)
    	return tmp
    
    function code(x, y)
    	tmp = 0.0
    	if (Float64((x ^ 4.0) - (y ^ 4.0)) <= -5e-323)
    		tmp = Float64(Float64(y * y) * Float64(Float64(-y) * y));
    	else
    		tmp = Float64(Float64(y - x) * Float64(Float64(Float64(y - x) * x) * x));
    	end
    	return tmp
    end
    
    function tmp_2 = code(x, y)
    	tmp = 0.0;
    	if (((x ^ 4.0) - (y ^ 4.0)) <= -5e-323)
    		tmp = (y * y) * (-y * y);
    	else
    		tmp = (y - x) * (((y - x) * x) * x);
    	end
    	tmp_2 = tmp;
    end
    
    code[x_, y_] := If[LessEqual[N[(N[Power[x, 4.0], $MachinePrecision] - N[Power[y, 4.0], $MachinePrecision]), $MachinePrecision], -5e-323], N[(N[(y * y), $MachinePrecision] * N[((-y) * y), $MachinePrecision]), $MachinePrecision], N[(N[(y - x), $MachinePrecision] * N[(N[(N[(y - x), $MachinePrecision] * x), $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision]]
    
    \begin{array}{l}
    
    \\
    \begin{array}{l}
    \mathbf{if}\;{x}^{4} - {y}^{4} \leq -5 \cdot 10^{-323}:\\
    \;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\
    
    \mathbf{else}:\\
    \;\;\;\;\left(y - x\right) \cdot \left(\left(\left(y - x\right) \cdot x\right) \cdot x\right)\\
    
    
    \end{array}
    \end{array}
    
    Derivation
    1. Split input into 2 regimes
    2. if (-.f64 (pow.f64 x #s(literal 4 binary64)) (pow.f64 y #s(literal 4 binary64))) < -4.94066e-323

      1. Initial program 100.0%

        \[{x}^{4} - {y}^{4} \]
      2. Add Preprocessing
      3. Taylor expanded in x around 0

        \[\leadsto \color{blue}{-1 \cdot {y}^{4}} \]
      4. Step-by-step derivation
        1. mul-1-negN/A

          \[\leadsto \color{blue}{\mathsf{neg}\left({y}^{4}\right)} \]
        2. lower-neg.f64N/A

          \[\leadsto \color{blue}{-{y}^{4}} \]
        3. lower-pow.f64100.0

          \[\leadsto -\color{blue}{{y}^{4}} \]
      5. Applied rewrites100.0%

        \[\leadsto \color{blue}{-{y}^{4}} \]
      6. Step-by-step derivation
        1. Applied rewrites99.7%

          \[\leadsto \left(\left(-y\right) \cdot y\right) \cdot \color{blue}{\left(y \cdot y\right)} \]

        if -4.94066e-323 < (-.f64 (pow.f64 x #s(literal 4 binary64)) (pow.f64 y #s(literal 4 binary64)))

        1. Initial program 79.1%

          \[{x}^{4} - {y}^{4} \]
        2. Add Preprocessing
        3. Step-by-step derivation
          1. lift--.f64N/A

            \[\leadsto \color{blue}{{x}^{4} - {y}^{4}} \]
          2. lift-pow.f64N/A

            \[\leadsto \color{blue}{{x}^{4}} - {y}^{4} \]
          3. sqr-powN/A

            \[\leadsto \color{blue}{{x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)}} - {y}^{4} \]
          4. lift-pow.f64N/A

            \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{4}} \]
          5. sqr-powN/A

            \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{\left(\frac{4}{2}\right)} \cdot {y}^{\left(\frac{4}{2}\right)}} \]
          6. difference-of-squaresN/A

            \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
          7. lower-*.f64N/A

            \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
          8. +-commutativeN/A

            \[\leadsto \color{blue}{\left({y}^{\left(\frac{4}{2}\right)} + {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          9. metadata-evalN/A

            \[\leadsto \left({y}^{\color{blue}{2}} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          10. unpow2N/A

            \[\leadsto \left(\color{blue}{y \cdot y} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          11. lower-fma.f64N/A

            \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          12. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, {x}^{\color{blue}{2}}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          13. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          14. lower-*.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          15. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{\color{blue}{2}} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          16. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{x \cdot x} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          17. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - {y}^{\color{blue}{2}}\right) \]
          18. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - \color{blue}{y \cdot y}\right) \]
          19. difference-of-squaresN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
          20. lower-*.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
          21. lower-+.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{\left(x + y\right)} \cdot \left(x - y\right)\right) \]
          22. lower--.f6499.8

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \color{blue}{\left(x - y\right)}\right) \]
        4. Applied rewrites99.8%

          \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
        5. Applied rewrites90.5%

          \[\leadsto \color{blue}{\left(y - x\right) \cdot \left(\left(y - x\right) \cdot \mathsf{fma}\left(y, y, x \cdot x\right)\right)} \]
        6. Taylor expanded in y around 0

          \[\leadsto \left(y - x\right) \cdot \color{blue}{\left(-1 \cdot {x}^{3} + {x}^{2} \cdot y\right)} \]
        7. Step-by-step derivation
          1. +-commutativeN/A

            \[\leadsto \left(y - x\right) \cdot \color{blue}{\left({x}^{2} \cdot y + -1 \cdot {x}^{3}\right)} \]
          2. *-commutativeN/A

            \[\leadsto \left(y - x\right) \cdot \left(\color{blue}{y \cdot {x}^{2}} + -1 \cdot {x}^{3}\right) \]
          3. unpow2N/A

            \[\leadsto \left(y - x\right) \cdot \left(y \cdot \color{blue}{\left(x \cdot x\right)} + -1 \cdot {x}^{3}\right) \]
          4. associate-*r*N/A

            \[\leadsto \left(y - x\right) \cdot \left(\color{blue}{\left(y \cdot x\right) \cdot x} + -1 \cdot {x}^{3}\right) \]
          5. *-commutativeN/A

            \[\leadsto \left(y - x\right) \cdot \left(\color{blue}{\left(x \cdot y\right)} \cdot x + -1 \cdot {x}^{3}\right) \]
          6. unpow3N/A

            \[\leadsto \left(y - x\right) \cdot \left(\left(x \cdot y\right) \cdot x + -1 \cdot \color{blue}{\left(\left(x \cdot x\right) \cdot x\right)}\right) \]
          7. unpow2N/A

            \[\leadsto \left(y - x\right) \cdot \left(\left(x \cdot y\right) \cdot x + -1 \cdot \left(\color{blue}{{x}^{2}} \cdot x\right)\right) \]
          8. associate-*r*N/A

            \[\leadsto \left(y - x\right) \cdot \left(\left(x \cdot y\right) \cdot x + \color{blue}{\left(-1 \cdot {x}^{2}\right) \cdot x}\right) \]
          9. distribute-rgt-inN/A

            \[\leadsto \left(y - x\right) \cdot \color{blue}{\left(x \cdot \left(x \cdot y + -1 \cdot {x}^{2}\right)\right)} \]
          10. unpow2N/A

            \[\leadsto \left(y - x\right) \cdot \left(x \cdot \left(x \cdot y + -1 \cdot \color{blue}{\left(x \cdot x\right)}\right)\right) \]
          11. associate-*r*N/A

            \[\leadsto \left(y - x\right) \cdot \left(x \cdot \left(x \cdot y + \color{blue}{\left(-1 \cdot x\right) \cdot x}\right)\right) \]
          12. *-commutativeN/A

            \[\leadsto \left(y - x\right) \cdot \left(x \cdot \left(x \cdot y + \color{blue}{x \cdot \left(-1 \cdot x\right)}\right)\right) \]
          13. distribute-lft-inN/A

            \[\leadsto \left(y - x\right) \cdot \left(x \cdot \color{blue}{\left(x \cdot \left(y + -1 \cdot x\right)\right)}\right) \]
          14. *-commutativeN/A

            \[\leadsto \left(y - x\right) \cdot \color{blue}{\left(\left(x \cdot \left(y + -1 \cdot x\right)\right) \cdot x\right)} \]
          15. lower-*.f64N/A

            \[\leadsto \left(y - x\right) \cdot \color{blue}{\left(\left(x \cdot \left(y + -1 \cdot x\right)\right) \cdot x\right)} \]
          16. *-commutativeN/A

            \[\leadsto \left(y - x\right) \cdot \left(\color{blue}{\left(\left(y + -1 \cdot x\right) \cdot x\right)} \cdot x\right) \]
          17. lower-*.f64N/A

            \[\leadsto \left(y - x\right) \cdot \left(\color{blue}{\left(\left(y + -1 \cdot x\right) \cdot x\right)} \cdot x\right) \]
          18. fp-cancel-sign-sub-invN/A

            \[\leadsto \left(y - x\right) \cdot \left(\left(\color{blue}{\left(y - \left(\mathsf{neg}\left(-1\right)\right) \cdot x\right)} \cdot x\right) \cdot x\right) \]
          19. metadata-evalN/A

            \[\leadsto \left(y - x\right) \cdot \left(\left(\left(y - \color{blue}{1} \cdot x\right) \cdot x\right) \cdot x\right) \]
          20. *-lft-identityN/A

            \[\leadsto \left(y - x\right) \cdot \left(\left(\left(y - \color{blue}{x}\right) \cdot x\right) \cdot x\right) \]
          21. lower--.f6490.5

            \[\leadsto \left(y - x\right) \cdot \left(\left(\color{blue}{\left(y - x\right)} \cdot x\right) \cdot x\right) \]
        8. Applied rewrites90.5%

          \[\leadsto \left(y - x\right) \cdot \color{blue}{\left(\left(\left(y - x\right) \cdot x\right) \cdot x\right)} \]
      7. Recombined 2 regimes into one program.
      8. Final simplification93.8%

        \[\leadsto \begin{array}{l} \mathbf{if}\;{x}^{4} - {y}^{4} \leq -5 \cdot 10^{-323}:\\ \;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\ \mathbf{else}:\\ \;\;\;\;\left(y - x\right) \cdot \left(\left(\left(y - x\right) \cdot x\right) \cdot x\right)\\ \end{array} \]
      9. Add Preprocessing

      Alternative 4: 70.2% accurate, 6.9× speedup?

      \[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;x \leq -6 \cdot 10^{+151} \lor \neg \left(x \leq 2.2 \cdot 10^{+143}\right):\\ \;\;\;\;\left(y \cdot y\right) \cdot \left(x \cdot x\right)\\ \mathbf{else}:\\ \;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\ \end{array} \end{array} \]
      (FPCore (x y)
       :precision binary64
       (if (or (<= x -6e+151) (not (<= x 2.2e+143)))
         (* (* y y) (* x x))
         (* (* y y) (* (- y) y))))
      double code(double x, double y) {
      	double tmp;
      	if ((x <= -6e+151) || !(x <= 2.2e+143)) {
      		tmp = (y * y) * (x * x);
      	} else {
      		tmp = (y * y) * (-y * y);
      	}
      	return tmp;
      }
      
      real(8) function code(x, y)
          real(8), intent (in) :: x
          real(8), intent (in) :: y
          real(8) :: tmp
          if ((x <= (-6d+151)) .or. (.not. (x <= 2.2d+143))) then
              tmp = (y * y) * (x * x)
          else
              tmp = (y * y) * (-y * y)
          end if
          code = tmp
      end function
      
      public static double code(double x, double y) {
      	double tmp;
      	if ((x <= -6e+151) || !(x <= 2.2e+143)) {
      		tmp = (y * y) * (x * x);
      	} else {
      		tmp = (y * y) * (-y * y);
      	}
      	return tmp;
      }
      
      def code(x, y):
      	tmp = 0
      	if (x <= -6e+151) or not (x <= 2.2e+143):
      		tmp = (y * y) * (x * x)
      	else:
      		tmp = (y * y) * (-y * y)
      	return tmp
      
      function code(x, y)
      	tmp = 0.0
      	if ((x <= -6e+151) || !(x <= 2.2e+143))
      		tmp = Float64(Float64(y * y) * Float64(x * x));
      	else
      		tmp = Float64(Float64(y * y) * Float64(Float64(-y) * y));
      	end
      	return tmp
      end
      
      function tmp_2 = code(x, y)
      	tmp = 0.0;
      	if ((x <= -6e+151) || ~((x <= 2.2e+143)))
      		tmp = (y * y) * (x * x);
      	else
      		tmp = (y * y) * (-y * y);
      	end
      	tmp_2 = tmp;
      end
      
      code[x_, y_] := If[Or[LessEqual[x, -6e+151], N[Not[LessEqual[x, 2.2e+143]], $MachinePrecision]], N[(N[(y * y), $MachinePrecision] * N[(x * x), $MachinePrecision]), $MachinePrecision], N[(N[(y * y), $MachinePrecision] * N[((-y) * y), $MachinePrecision]), $MachinePrecision]]
      
      \begin{array}{l}
      
      \\
      \begin{array}{l}
      \mathbf{if}\;x \leq -6 \cdot 10^{+151} \lor \neg \left(x \leq 2.2 \cdot 10^{+143}\right):\\
      \;\;\;\;\left(y \cdot y\right) \cdot \left(x \cdot x\right)\\
      
      \mathbf{else}:\\
      \;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\
      
      
      \end{array}
      \end{array}
      
      Derivation
      1. Split input into 2 regimes
      2. if x < -5.9999999999999998e151 or 2.20000000000000014e143 < x

        1. Initial program 67.1%

          \[{x}^{4} - {y}^{4} \]
        2. Add Preprocessing
        3. Step-by-step derivation
          1. lift--.f64N/A

            \[\leadsto \color{blue}{{x}^{4} - {y}^{4}} \]
          2. lift-pow.f64N/A

            \[\leadsto \color{blue}{{x}^{4}} - {y}^{4} \]
          3. sqr-powN/A

            \[\leadsto \color{blue}{{x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)}} - {y}^{4} \]
          4. lift-pow.f64N/A

            \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{4}} \]
          5. sqr-powN/A

            \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{\left(\frac{4}{2}\right)} \cdot {y}^{\left(\frac{4}{2}\right)}} \]
          6. difference-of-squaresN/A

            \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
          7. lower-*.f64N/A

            \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
          8. +-commutativeN/A

            \[\leadsto \color{blue}{\left({y}^{\left(\frac{4}{2}\right)} + {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          9. metadata-evalN/A

            \[\leadsto \left({y}^{\color{blue}{2}} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          10. unpow2N/A

            \[\leadsto \left(\color{blue}{y \cdot y} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          11. lower-fma.f64N/A

            \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          12. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, {x}^{\color{blue}{2}}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          13. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          14. lower-*.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          15. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{\color{blue}{2}} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          16. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{x \cdot x} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          17. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - {y}^{\color{blue}{2}}\right) \]
          18. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - \color{blue}{y \cdot y}\right) \]
          19. difference-of-squaresN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
          20. lower-*.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
          21. lower-+.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{\left(x + y\right)} \cdot \left(x - y\right)\right) \]
          22. lower--.f64100.0

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \color{blue}{\left(x - y\right)}\right) \]
        4. Applied rewrites100.0%

          \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
        5. Taylor expanded in x around inf

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left({x}^{2} \cdot \left(1 + \left(-1 \cdot \frac{y}{x} + \frac{y}{x}\right)\right)\right)} \]
        6. Step-by-step derivation
          1. distribute-lft1-inN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{\left(-1 + 1\right) \cdot \frac{y}{x}}\right)\right) \]
          2. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{0} \cdot \frac{y}{x}\right)\right) \]
          3. mul0-lftN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{0}\right)\right) \]
          4. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \color{blue}{1}\right) \]
          5. *-rgt-identityN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{{x}^{2}} \]
          6. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
          7. lower-*.f6491.8

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
        7. Applied rewrites91.8%

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
        8. Taylor expanded in x around 0

          \[\leadsto \color{blue}{{y}^{2}} \cdot \left(x \cdot x\right) \]
        9. Step-by-step derivation
          1. unpow2N/A

            \[\leadsto \color{blue}{\left(y \cdot y\right)} \cdot \left(x \cdot x\right) \]
          2. lower-*.f6465.9

            \[\leadsto \color{blue}{\left(y \cdot y\right)} \cdot \left(x \cdot x\right) \]
        10. Applied rewrites65.9%

          \[\leadsto \color{blue}{\left(y \cdot y\right)} \cdot \left(x \cdot x\right) \]

        if -5.9999999999999998e151 < x < 2.20000000000000014e143

        1. Initial program 94.5%

          \[{x}^{4} - {y}^{4} \]
        2. Add Preprocessing
        3. Taylor expanded in x around 0

          \[\leadsto \color{blue}{-1 \cdot {y}^{4}} \]
        4. Step-by-step derivation
          1. mul-1-negN/A

            \[\leadsto \color{blue}{\mathsf{neg}\left({y}^{4}\right)} \]
          2. lower-neg.f64N/A

            \[\leadsto \color{blue}{-{y}^{4}} \]
          3. lower-pow.f6472.0

            \[\leadsto -\color{blue}{{y}^{4}} \]
        5. Applied rewrites72.0%

          \[\leadsto \color{blue}{-{y}^{4}} \]
        6. Step-by-step derivation
          1. Applied rewrites71.8%

            \[\leadsto \left(\left(-y\right) \cdot y\right) \cdot \color{blue}{\left(y \cdot y\right)} \]
        7. Recombined 2 regimes into one program.
        8. Final simplification70.1%

          \[\leadsto \begin{array}{l} \mathbf{if}\;x \leq -6 \cdot 10^{+151} \lor \neg \left(x \leq 2.2 \cdot 10^{+143}\right):\\ \;\;\;\;\left(y \cdot y\right) \cdot \left(x \cdot x\right)\\ \mathbf{else}:\\ \;\;\;\;\left(y \cdot y\right) \cdot \left(\left(-y\right) \cdot y\right)\\ \end{array} \]
        9. Add Preprocessing

        Alternative 5: 32.5% accurate, 12.9× speedup?

        \[\begin{array}{l} \\ \left(y \cdot y\right) \cdot \left(x \cdot x\right) \end{array} \]
        (FPCore (x y) :precision binary64 (* (* y y) (* x x)))
        double code(double x, double y) {
        	return (y * y) * (x * x);
        }
        
        real(8) function code(x, y)
            real(8), intent (in) :: x
            real(8), intent (in) :: y
            code = (y * y) * (x * x)
        end function
        
        public static double code(double x, double y) {
        	return (y * y) * (x * x);
        }
        
        def code(x, y):
        	return (y * y) * (x * x)
        
        function code(x, y)
        	return Float64(Float64(y * y) * Float64(x * x))
        end
        
        function tmp = code(x, y)
        	tmp = (y * y) * (x * x);
        end
        
        code[x_, y_] := N[(N[(y * y), $MachinePrecision] * N[(x * x), $MachinePrecision]), $MachinePrecision]
        
        \begin{array}{l}
        
        \\
        \left(y \cdot y\right) \cdot \left(x \cdot x\right)
        \end{array}
        
        Derivation
        1. Initial program 86.7%

          \[{x}^{4} - {y}^{4} \]
        2. Add Preprocessing
        3. Step-by-step derivation
          1. lift--.f64N/A

            \[\leadsto \color{blue}{{x}^{4} - {y}^{4}} \]
          2. lift-pow.f64N/A

            \[\leadsto \color{blue}{{x}^{4}} - {y}^{4} \]
          3. sqr-powN/A

            \[\leadsto \color{blue}{{x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)}} - {y}^{4} \]
          4. lift-pow.f64N/A

            \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{4}} \]
          5. sqr-powN/A

            \[\leadsto {x}^{\left(\frac{4}{2}\right)} \cdot {x}^{\left(\frac{4}{2}\right)} - \color{blue}{{y}^{\left(\frac{4}{2}\right)} \cdot {y}^{\left(\frac{4}{2}\right)}} \]
          6. difference-of-squaresN/A

            \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
          7. lower-*.f64N/A

            \[\leadsto \color{blue}{\left({x}^{\left(\frac{4}{2}\right)} + {y}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right)} \]
          8. +-commutativeN/A

            \[\leadsto \color{blue}{\left({y}^{\left(\frac{4}{2}\right)} + {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          9. metadata-evalN/A

            \[\leadsto \left({y}^{\color{blue}{2}} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          10. unpow2N/A

            \[\leadsto \left(\color{blue}{y \cdot y} + {x}^{\left(\frac{4}{2}\right)}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          11. lower-fma.f64N/A

            \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, {x}^{\left(\frac{4}{2}\right)}\right)} \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          12. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, {x}^{\color{blue}{2}}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          13. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          14. lower-*.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, \color{blue}{x \cdot x}\right) \cdot \left({x}^{\left(\frac{4}{2}\right)} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          15. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{\color{blue}{2}} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          16. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{x \cdot x} - {y}^{\left(\frac{4}{2}\right)}\right) \]
          17. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - {y}^{\color{blue}{2}}\right) \]
          18. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(x \cdot x - \color{blue}{y \cdot y}\right) \]
          19. difference-of-squaresN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
          20. lower-*.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
          21. lower-+.f64N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\color{blue}{\left(x + y\right)} \cdot \left(x - y\right)\right) \]
          22. lower--.f6499.8

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \color{blue}{\left(x - y\right)}\right) \]
        4. Applied rewrites99.8%

          \[\leadsto \color{blue}{\mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left(\left(x + y\right) \cdot \left(x - y\right)\right)} \]
        5. Taylor expanded in x around inf

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left({x}^{2} \cdot \left(1 + \left(-1 \cdot \frac{y}{x} + \frac{y}{x}\right)\right)\right)} \]
        6. Step-by-step derivation
          1. distribute-lft1-inN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{\left(-1 + 1\right) \cdot \frac{y}{x}}\right)\right) \]
          2. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{0} \cdot \frac{y}{x}\right)\right) \]
          3. mul0-lftN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \left(1 + \color{blue}{0}\right)\right) \]
          4. metadata-evalN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \left({x}^{2} \cdot \color{blue}{1}\right) \]
          5. *-rgt-identityN/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{{x}^{2}} \]
          6. unpow2N/A

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
          7. lower-*.f6458.5

            \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
        7. Applied rewrites58.5%

          \[\leadsto \mathsf{fma}\left(y, y, x \cdot x\right) \cdot \color{blue}{\left(x \cdot x\right)} \]
        8. Taylor expanded in x around 0

          \[\leadsto \color{blue}{{y}^{2}} \cdot \left(x \cdot x\right) \]
        9. Step-by-step derivation
          1. unpow2N/A

            \[\leadsto \color{blue}{\left(y \cdot y\right)} \cdot \left(x \cdot x\right) \]
          2. lower-*.f6432.6

            \[\leadsto \color{blue}{\left(y \cdot y\right)} \cdot \left(x \cdot x\right) \]
        10. Applied rewrites32.6%

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

        Alternative 6: 21.9% accurate, 12.9× speedup?

        \[\begin{array}{l} \\ \left(y \cdot y\right) \cdot \left(y \cdot y\right) \end{array} \]
        (FPCore (x y) :precision binary64 (* (* y y) (* y y)))
        double code(double x, double y) {
        	return (y * y) * (y * y);
        }
        
        real(8) function code(x, y)
            real(8), intent (in) :: x
            real(8), intent (in) :: y
            code = (y * y) * (y * y)
        end function
        
        public static double code(double x, double y) {
        	return (y * y) * (y * y);
        }
        
        def code(x, y):
        	return (y * y) * (y * y)
        
        function code(x, y)
        	return Float64(Float64(y * y) * Float64(y * y))
        end
        
        function tmp = code(x, y)
        	tmp = (y * y) * (y * y);
        end
        
        code[x_, y_] := N[(N[(y * y), $MachinePrecision] * N[(y * y), $MachinePrecision]), $MachinePrecision]
        
        \begin{array}{l}
        
        \\
        \left(y \cdot y\right) \cdot \left(y \cdot y\right)
        \end{array}
        
        Derivation
        1. Initial program 86.7%

          \[{x}^{4} - {y}^{4} \]
        2. Add Preprocessing
        3. Taylor expanded in x around 0

          \[\leadsto \color{blue}{-1 \cdot {y}^{4}} \]
        4. Step-by-step derivation
          1. mul-1-negN/A

            \[\leadsto \color{blue}{\mathsf{neg}\left({y}^{4}\right)} \]
          2. lower-neg.f64N/A

            \[\leadsto \color{blue}{-{y}^{4}} \]
          3. lower-pow.f6454.0

            \[\leadsto -\color{blue}{{y}^{4}} \]
        5. Applied rewrites54.0%

          \[\leadsto \color{blue}{-{y}^{4}} \]
        6. Step-by-step derivation
          1. Applied rewrites20.4%

            \[\leadsto \left(y \cdot y\right) \cdot \color{blue}{\left(y \cdot y\right)} \]
          2. Add Preprocessing

          Reproduce

          ?
          herbie shell --seed 2024340 
          (FPCore (x y)
            :name "Radioactive exchange between two surfaces"
            :precision binary64
            (- (pow x 4.0) (pow y 4.0)))