Numeric.SpecFunctions:logGamma from math-functions-0.1.5.2, B

Percentage Accurate: 69.1% → 99.4%

Time: 3.1s

Alternatives: 11

Speedup: 5.0×

Specification

Math

\[\begin{array}{l} \\ x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (+
  x
  (/
   (*
    y
    (+
     (* (+ (* z 0.0692910599291889) 0.4917317610505968) z)
     0.279195317918525))
   (+ (* (+ z 6.012459259764103) z) 3.350343815022304))))

C

double code(double x, double y, double z) {
	return x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304));
}

Fortran

module fmin_fmax_functions
    implicit none
    private
    public fmax
    public fmin

    interface fmax
        module procedure fmax88
        module procedure fmax44
        module procedure fmax84
        module procedure fmax48
    end interface
    interface fmin
        module procedure fmin88
        module procedure fmin44
        module procedure fmin84
        module procedure fmin48
    end interface
contains
    real(8) function fmax88(x, y) result (res)
        real(8), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(x, max(x, y), y /= y), x /= x)
    end function
    real(4) function fmax44(x, y) result (res)
        real(4), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(y, merge(x, max(x, y), y /= y), x /= x)
    end function
    real(8) function fmax84(x, y) result(res)
        real(8), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
    end function
    real(8) function fmax48(x, y) result(res)
        real(4), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
    end function
    real(8) function fmin88(x, y) result (res)
        real(8), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(x, min(x, y), y /= y), x /= x)
    end function
    real(4) function fmin44(x, y) result (res)
        real(4), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(y, merge(x, min(x, y), y /= y), x /= x)
    end function
    real(8) function fmin84(x, y) result(res)
        real(8), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
    end function
    real(8) function fmin48(x, y) result(res)
        real(4), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
    end function
end module

real(8) function code(x, y, z)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    code = x + ((y * ((((z * 0.0692910599291889d0) + 0.4917317610505968d0) * z) + 0.279195317918525d0)) / (((z + 6.012459259764103d0) * z) + 3.350343815022304d0))
end function

Java

public static double code(double x, double y, double z) {
	return x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304));
}

Python

def code(x, y, z):
	return x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304))

Julia

function code(x, y, z)
	return Float64(x + Float64(Float64(y * Float64(Float64(Float64(Float64(z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / Float64(Float64(Float64(z + 6.012459259764103) * z) + 3.350343815022304)))
end

MATLAB

function tmp = code(x, y, z)
	tmp = x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304));
end

Wolfram

code[x_, y_, z_] := N[(x + N[(N[(y * N[(N[(N[(N[(z * 0.0692910599291889), $MachinePrecision] + 0.4917317610505968), $MachinePrecision] * z), $MachinePrecision] + 0.279195317918525), $MachinePrecision]), $MachinePrecision] / N[(N[(N[(z + 6.012459259764103), $MachinePrecision] * z), $MachinePrecision] + 3.350343815022304), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]

Initial Program: 69.1% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (+
  x
  (/
   (*
    y
    (+
     (* (+ (* z 0.0692910599291889) 0.4917317610505968) z)
     0.279195317918525))
   (+ (* (+ z 6.012459259764103) z) 3.350343815022304))))

C

double code(double x, double y, double z) {
	return x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304));
}

Fortran

module fmin_fmax_functions
    implicit none
    private
    public fmax
    public fmin

    interface fmax
        module procedure fmax88
        module procedure fmax44
        module procedure fmax84
        module procedure fmax48
    end interface
    interface fmin
        module procedure fmin88
        module procedure fmin44
        module procedure fmin84
        module procedure fmin48
    end interface
contains
    real(8) function fmax88(x, y) result (res)
        real(8), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(x, max(x, y), y /= y), x /= x)
    end function
    real(4) function fmax44(x, y) result (res)
        real(4), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(y, merge(x, max(x, y), y /= y), x /= x)
    end function
    real(8) function fmax84(x, y) result(res)
        real(8), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
    end function
    real(8) function fmax48(x, y) result(res)
        real(4), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
    end function
    real(8) function fmin88(x, y) result (res)
        real(8), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(x, min(x, y), y /= y), x /= x)
    end function
    real(4) function fmin44(x, y) result (res)
        real(4), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(y, merge(x, min(x, y), y /= y), x /= x)
    end function
    real(8) function fmin84(x, y) result(res)
        real(8), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
    end function
    real(8) function fmin48(x, y) result(res)
        real(4), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
    end function
end module

real(8) function code(x, y, z)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    code = x + ((y * ((((z * 0.0692910599291889d0) + 0.4917317610505968d0) * z) + 0.279195317918525d0)) / (((z + 6.012459259764103d0) * z) + 3.350343815022304d0))
end function

Java

public static double code(double x, double y, double z) {
	return x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304));
}

Python

def code(x, y, z):
	return x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304))

Julia

function code(x, y, z)
	return Float64(x + Float64(Float64(y * Float64(Float64(Float64(Float64(z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / Float64(Float64(Float64(z + 6.012459259764103) * z) + 3.350343815022304)))
end

MATLAB

function tmp = code(x, y, z)
	tmp = x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304));
end

Wolfram

code[x_, y_, z_] := N[(x + N[(N[(y * N[(N[(N[(N[(z * 0.0692910599291889), $MachinePrecision] + 0.4917317610505968), $MachinePrecision] * z), $MachinePrecision] + 0.279195317918525), $MachinePrecision]), $MachinePrecision] / N[(N[(N[(z + 6.012459259764103), $MachinePrecision] * z), $MachinePrecision] + 3.350343815022304), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]

Alternative 1: 99.4% accurate, 0.5× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \leq 5 \cdot 10^{+258}:\\ \;\;\;\;\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<=
      (+
       x
       (/
        (*
         y
         (+
          (* (+ (* z 0.0692910599291889) 0.4917317610505968) z)
          0.279195317918525))
        (+ (* (+ z 6.012459259764103) z) 3.350343815022304)))
      5e+258)
   (fma
    (/
     (fma (fma 0.0692910599291889 z 0.4917317610505968) z 0.279195317918525)
     (fma (- z -6.012459259764103) z 3.350343815022304))
    y
    x)
   (fma 0.0692910599291889 y x)))

C

double code(double x, double y, double z) {
	double tmp;
	if ((x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304))) <= 5e+258) {
		tmp = fma((fma(fma(0.0692910599291889, z, 0.4917317610505968), z, 0.279195317918525) / fma((z - -6.012459259764103), z, 3.350343815022304)), y, x);
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (Float64(x + Float64(Float64(y * Float64(Float64(Float64(Float64(z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / Float64(Float64(Float64(z + 6.012459259764103) * z) + 3.350343815022304))) <= 5e+258)
		tmp = fma(Float64(fma(fma(0.0692910599291889, z, 0.4917317610505968), z, 0.279195317918525) / fma(Float64(z - -6.012459259764103), z, 3.350343815022304)), y, x);
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[N[(x + N[(N[(y * N[(N[(N[(N[(z * 0.0692910599291889), $MachinePrecision] + 0.4917317610505968), $MachinePrecision] * z), $MachinePrecision] + 0.279195317918525), $MachinePrecision]), $MachinePrecision] / N[(N[(N[(z + 6.012459259764103), $MachinePrecision] * z), $MachinePrecision] + 3.350343815022304), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], 5e+258], N[(N[(N[(N[(0.0692910599291889 * z + 0.4917317610505968), $MachinePrecision] * z + 0.279195317918525), $MachinePrecision] / N[(N[(z - -6.012459259764103), $MachinePrecision] * z + 3.350343815022304), $MachinePrecision]), $MachinePrecision] * y + x), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if (+.f64 x (/.f64 (*.f64 y (+.f64 (*.f64 (+.f64 (*.f64 z #s(literal 692910599291889/10000000000000000 binary64)) #s(literal 307332350656623/625000000000000 binary64)) z) #s(literal 11167812716741/40000000000000 binary64))) (+.f64 (*.f64 (+.f64 z #s(literal 6012459259764103/1000000000000000 binary64)) z) #s(literal 104698244219447/31250000000000 binary64)))) < 5e258

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   if 5e258 < (+.f64 x (/.f64 (*.f64 y (+.f64 (*.f64 (+.f64 (*.f64 z #s(literal 692910599291889/10000000000000000 binary64)) #s(literal 307332350656623/625000000000000 binary64)) z) #s(literal 11167812716741/40000000000000 binary64))) (+.f64 (*.f64 (+.f64 z #s(literal 6012459259764103/1000000000000000 binary64)) z) #s(literal 104698244219447/31250000000000 binary64))))

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 2: 99.3% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;z \leq -5.5:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889 + -1 \cdot \frac{0.4046220386999212 \cdot \frac{1}{z} - 0.07512208616047561}{z}, y, x\right)\\ \mathbf{elif}\;z \leq 2.7:\\ \;\;\;\;\mathsf{fma}\left(0.08333333333333323 + z \cdot \left(z \cdot \left(0.0007936505811533442 + -0.0005951669793454025 \cdot z\right) - 0.00277777777751721\right), y, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<= z -5.5)
   (fma
    (+
     0.0692910599291889
     (* -1.0 (/ (- (* 0.4046220386999212 (/ 1.0 z)) 0.07512208616047561) z)))
    y
    x)
   (if (<= z 2.7)
     (fma
      (+
       0.08333333333333323
       (*
        z
        (-
         (* z (+ 0.0007936505811533442 (* -0.0005951669793454025 z)))
         0.00277777777751721)))
      y
      x)
     (fma 0.0692910599291889 y x))))

C

double code(double x, double y, double z) {
	double tmp;
	if (z <= -5.5) {
		tmp = fma((0.0692910599291889 + (-1.0 * (((0.4046220386999212 * (1.0 / z)) - 0.07512208616047561) / z))), y, x);
	} else if (z <= 2.7) {
		tmp = fma((0.08333333333333323 + (z * ((z * (0.0007936505811533442 + (-0.0005951669793454025 * z))) - 0.00277777777751721))), y, x);
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (z <= -5.5)
		tmp = fma(Float64(0.0692910599291889 + Float64(-1.0 * Float64(Float64(Float64(0.4046220386999212 * Float64(1.0 / z)) - 0.07512208616047561) / z))), y, x);
	elseif (z <= 2.7)
		tmp = fma(Float64(0.08333333333333323 + Float64(z * Float64(Float64(z * Float64(0.0007936505811533442 + Float64(-0.0005951669793454025 * z))) - 0.00277777777751721))), y, x);
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[z, -5.5], N[(N[(0.0692910599291889 + N[(-1.0 * N[(N[(N[(0.4046220386999212 * N[(1.0 / z), $MachinePrecision]), $MachinePrecision] - 0.07512208616047561), $MachinePrecision] / z), $MachinePrecision]), $MachinePrecision]), $MachinePrecision] * y + x), $MachinePrecision], If[LessEqual[z, 2.7], N[(N[(0.08333333333333323 + N[(z * N[(N[(z * N[(0.0007936505811533442 + N[(-0.0005951669793454025 * z), $MachinePrecision]), $MachinePrecision]), $MachinePrecision] - 0.00277777777751721), $MachinePrecision]), $MachinePrecision]), $MachinePrecision] * y + x), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]]

Derivation

1. Split input into 3 regimes

2. if z < -5.5

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around -inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000} + -1 \cdot \frac{\frac{4046220386999211718548694042263781576003973599}{10000000000000000000000000000000000000000000000} \cdot \frac{1}{z} - \frac{751220861604756070699018739433}{10000000000000000000000000000000}}{z}}, y, x\right) \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \color{blue}{-1 \cdot \frac{\frac{4046220386999211718548694042263781576003973599}{10000000000000000000000000000000000000000000000} \cdot \frac{1}{z} - \frac{751220861604756070699018739433}{10000000000000000000000000000000}}{z}}, y, x\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + -1 \cdot \color{blue}{\frac{\frac{4046220386999211718548694042263781576003973599}{10000000000000000000000000000000000000000000000} \cdot \frac{1}{z} - \frac{751220861604756070699018739433}{10000000000000000000000000000000}}{z}}, y, x\right) \]

      3. lower-/.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + -1 \cdot \frac{\frac{4046220386999211718548694042263781576003973599}{10000000000000000000000000000000000000000000000} \cdot \frac{1}{z} - \frac{751220861604756070699018739433}{10000000000000000000000000000000}}{\color{blue}{z}}, y, x\right) \]

      4. lower--.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + -1 \cdot \frac{\frac{4046220386999211718548694042263781576003973599}{10000000000000000000000000000000000000000000000} \cdot \frac{1}{z} - \frac{751220861604756070699018739433}{10000000000000000000000000000000}}{z}, y, x\right) \]

      5. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + -1 \cdot \frac{\frac{4046220386999211718548694042263781576003973599}{10000000000000000000000000000000000000000000000} \cdot \frac{1}{z} - \frac{751220861604756070699018739433}{10000000000000000000000000000000}}{z}, y, x\right) \]

      6. lower-/.f64 57.5

         \[\leadsto \mathsf{fma}\left(0.0692910599291889 + -1 \cdot \frac{0.4046220386999212 \cdot \frac{1}{z} - 0.07512208616047561}{z}, y, x\right) \]

   6. Applied rewrites 57.5%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889 + -1 \cdot \frac{0.4046220386999212 \cdot \frac{1}{z} - 0.07512208616047561}{z}}, y, x\right) \]

   if -5.5 < z < 2.7000000000000002

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around 0

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{279195317918525}{3350343815022304} + z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + \color{blue}{z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \color{blue}{\left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]

      3. lower--.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \color{blue}{\frac{155900051080628738716045985239}{56124018394291031809500087342080}}\right), y, x\right) \]

      4. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right), y, x\right) \]

      5. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right), y, x\right) \]

      6. lower-*.f64 56.3

         \[\leadsto \mathsf{fma}\left(0.08333333333333323 + z \cdot \left(z \cdot \left(0.0007936505811533442 + -0.0005951669793454025 \cdot z\right) - 0.00277777777751721\right), y, x\right) \]

   6. Applied rewrites 56.3%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.08333333333333323 + z \cdot \left(z \cdot \left(0.0007936505811533442 + -0.0005951669793454025 \cdot z\right) - 0.00277777777751721\right)}, y, x\right) \]

   if 2.7000000000000002 < z

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 3: 99.3% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;z \leq -5.5:\\ \;\;\;\;\mathsf{fma}\left(\frac{0.07512208616047561}{z} - -0.0692910599291889, y, x\right)\\ \mathbf{elif}\;z \leq 2.7:\\ \;\;\;\;\mathsf{fma}\left(0.08333333333333323 + z \cdot \left(z \cdot \left(0.0007936505811533442 + -0.0005951669793454025 \cdot z\right) - 0.00277777777751721\right), y, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<= z -5.5)
   (fma (- (/ 0.07512208616047561 z) -0.0692910599291889) y x)
   (if (<= z 2.7)
     (fma
      (+
       0.08333333333333323
       (*
        z
        (-
         (* z (+ 0.0007936505811533442 (* -0.0005951669793454025 z)))
         0.00277777777751721)))
      y
      x)
     (fma 0.0692910599291889 y x))))

C

double code(double x, double y, double z) {
	double tmp;
	if (z <= -5.5) {
		tmp = fma(((0.07512208616047561 / z) - -0.0692910599291889), y, x);
	} else if (z <= 2.7) {
		tmp = fma((0.08333333333333323 + (z * ((z * (0.0007936505811533442 + (-0.0005951669793454025 * z))) - 0.00277777777751721))), y, x);
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (z <= -5.5)
		tmp = fma(Float64(Float64(0.07512208616047561 / z) - -0.0692910599291889), y, x);
	elseif (z <= 2.7)
		tmp = fma(Float64(0.08333333333333323 + Float64(z * Float64(Float64(z * Float64(0.0007936505811533442 + Float64(-0.0005951669793454025 * z))) - 0.00277777777751721))), y, x);
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[z, -5.5], N[(N[(N[(0.07512208616047561 / z), $MachinePrecision] - -0.0692910599291889), $MachinePrecision] * y + x), $MachinePrecision], If[LessEqual[z, 2.7], N[(N[(0.08333333333333323 + N[(z * N[(N[(z * N[(0.0007936505811533442 + N[(-0.0005951669793454025 * z), $MachinePrecision]), $MachinePrecision]), $MachinePrecision] - 0.00277777777751721), $MachinePrecision]), $MachinePrecision]), $MachinePrecision] * y + x), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]]

Derivation

1. Split input into 3 regimes

2. if z < -5.5

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000} + \frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \color{blue}{\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \color{blue}{\frac{1}{z}}, y, x\right) \]

      3. lower-/.f64 64.8

         \[\leadsto \mathsf{fma}\left(0.0692910599291889 + 0.07512208616047561 \cdot \frac{1}{\color{blue}{z}}, y, x\right) \]

   6. Applied rewrites 64.8%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889 + 0.07512208616047561 \cdot \frac{1}{z}}, y, x\right) \]
   7. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \color{blue}{\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]

      2. +-commutative N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} + \color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]

      3. add-flip N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \color{blue}{\left(\mathsf{neg}\left(\frac{692910599291889}{10000000000000000}\right)\right)}, y, x\right) \]

      4. metadata-eval N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      5. lower--.f64 64.8

         \[\leadsto \mathsf{fma}\left(0.07512208616047561 \cdot \frac{1}{z} - \color{blue}{-0.0692910599291889}, y, x\right) \]

      6. lift-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      7. lift-/.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      8. mult-flip-rev N/A

         \[\leadsto \mathsf{fma}\left(\frac{\frac{751220861604756070699018739433}{10000000000000000000000000000000}}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      9. lower-/.f64 64.8

         \[\leadsto \mathsf{fma}\left(\frac{0.07512208616047561}{z} - -0.0692910599291889, y, x\right) \]

   8. Applied rewrites 64.8%

      \[\leadsto \mathsf{fma}\left(\frac{0.07512208616047561}{z} - \color{blue}{-0.0692910599291889}, y, x\right) \]

   if -5.5 < z < 2.7000000000000002

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around 0

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{279195317918525}{3350343815022304} + z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + \color{blue}{z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \color{blue}{\left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]

      3. lower--.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \color{blue}{\frac{155900051080628738716045985239}{56124018394291031809500087342080}}\right), y, x\right) \]

      4. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right), y, x\right) \]

      5. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \left(z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} + \frac{-374943941275717765274452559944207169728571246668095556552487}{629981088144543617699065742275429975113587435159029727787745280} \cdot z\right) - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right), y, x\right) \]

      6. lower-*.f64 56.3

         \[\leadsto \mathsf{fma}\left(0.08333333333333323 + z \cdot \left(z \cdot \left(0.0007936505811533442 + -0.0005951669793454025 \cdot z\right) - 0.00277777777751721\right), y, x\right) \]

   6. Applied rewrites 56.3%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.08333333333333323 + z \cdot \left(z \cdot \left(0.0007936505811533442 + -0.0005951669793454025 \cdot z\right) - 0.00277777777751721\right)}, y, x\right) \]

   if 2.7000000000000002 < z

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 4: 99.3% accurate, 1.2× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;z \leq -5.5:\\ \;\;\;\;\mathsf{fma}\left(\frac{0.07512208616047561}{z} - -0.0692910599291889, y, x\right)\\ \mathbf{elif}\;z \leq 4.4:\\ \;\;\;\;\mathsf{fma}\left(0.08333333333333323 + z \cdot \left(0.0007936505811533442 \cdot z - 0.00277777777751721\right), y, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<= z -5.5)
   (fma (- (/ 0.07512208616047561 z) -0.0692910599291889) y x)
   (if (<= z 4.4)
     (fma
      (+
       0.08333333333333323
       (* z (- (* 0.0007936505811533442 z) 0.00277777777751721)))
      y
      x)
     (fma 0.0692910599291889 y x))))

C

double code(double x, double y, double z) {
	double tmp;
	if (z <= -5.5) {
		tmp = fma(((0.07512208616047561 / z) - -0.0692910599291889), y, x);
	} else if (z <= 4.4) {
		tmp = fma((0.08333333333333323 + (z * ((0.0007936505811533442 * z) - 0.00277777777751721))), y, x);
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (z <= -5.5)
		tmp = fma(Float64(Float64(0.07512208616047561 / z) - -0.0692910599291889), y, x);
	elseif (z <= 4.4)
		tmp = fma(Float64(0.08333333333333323 + Float64(z * Float64(Float64(0.0007936505811533442 * z) - 0.00277777777751721))), y, x);
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[z, -5.5], N[(N[(N[(0.07512208616047561 / z), $MachinePrecision] - -0.0692910599291889), $MachinePrecision] * y + x), $MachinePrecision], If[LessEqual[z, 4.4], N[(N[(0.08333333333333323 + N[(z * N[(N[(0.0007936505811533442 * z), $MachinePrecision] - 0.00277777777751721), $MachinePrecision]), $MachinePrecision]), $MachinePrecision] * y + x), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]]

Derivation

1. Split input into 3 regimes

2. if z < -5.5

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000} + \frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \color{blue}{\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \color{blue}{\frac{1}{z}}, y, x\right) \]

      3. lower-/.f64 64.8

         \[\leadsto \mathsf{fma}\left(0.0692910599291889 + 0.07512208616047561 \cdot \frac{1}{\color{blue}{z}}, y, x\right) \]

   6. Applied rewrites 64.8%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889 + 0.07512208616047561 \cdot \frac{1}{z}}, y, x\right) \]
   7. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \color{blue}{\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]

      2. +-commutative N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} + \color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]

      3. add-flip N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \color{blue}{\left(\mathsf{neg}\left(\frac{692910599291889}{10000000000000000}\right)\right)}, y, x\right) \]

      4. metadata-eval N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      5. lower--.f64 64.8

         \[\leadsto \mathsf{fma}\left(0.07512208616047561 \cdot \frac{1}{z} - \color{blue}{-0.0692910599291889}, y, x\right) \]

      6. lift-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      7. lift-/.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      8. mult-flip-rev N/A

         \[\leadsto \mathsf{fma}\left(\frac{\frac{751220861604756070699018739433}{10000000000000000000000000000000}}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      9. lower-/.f64 64.8

         \[\leadsto \mathsf{fma}\left(\frac{0.07512208616047561}{z} - -0.0692910599291889, y, x\right) \]

   8. Applied rewrites 64.8%

      \[\leadsto \mathsf{fma}\left(\frac{0.07512208616047561}{z} - \color{blue}{-0.0692910599291889}, y, x\right) \]

   if -5.5 < z < 4.4000000000000004

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around 0

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{279195317918525}{3350343815022304} + z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} \cdot z - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + \color{blue}{z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} \cdot z - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \color{blue}{\left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} \cdot z - \frac{155900051080628738716045985239}{56124018394291031809500087342080}\right)}, y, x\right) \]

      3. lower--.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{279195317918525}{3350343815022304} + z \cdot \left(\frac{149233894885562575800992648418763933371314529}{188034757901510979839193143041976607183277752320} \cdot z - \color{blue}{\frac{155900051080628738716045985239}{56124018394291031809500087342080}}\right), y, x\right) \]

      4. lower-*.f64 59.4

         \[\leadsto \mathsf{fma}\left(0.08333333333333323 + z \cdot \left(0.0007936505811533442 \cdot z - 0.00277777777751721\right), y, x\right) \]

   6. Applied rewrites 59.4%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.08333333333333323 + z \cdot \left(0.0007936505811533442 \cdot z - 0.00277777777751721\right)}, y, x\right) \]

   if 4.4000000000000004 < z

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 5: 99.1% accurate, 1.6× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;z \leq -5.5:\\ \;\;\;\;\mathsf{fma}\left(\frac{0.07512208616047561}{z} - -0.0692910599291889, y, x\right)\\ \mathbf{elif}\;z \leq 5:\\ \;\;\;\;y \cdot \mathsf{fma}\left(-0.00277777777751721, z, 0.08333333333333323\right) + x\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<= z -5.5)
   (fma (- (/ 0.07512208616047561 z) -0.0692910599291889) y x)
   (if (<= z 5.0)
     (+ (* y (fma -0.00277777777751721 z 0.08333333333333323)) x)
     (fma 0.0692910599291889 y x))))

C

double code(double x, double y, double z) {
	double tmp;
	if (z <= -5.5) {
		tmp = fma(((0.07512208616047561 / z) - -0.0692910599291889), y, x);
	} else if (z <= 5.0) {
		tmp = (y * fma(-0.00277777777751721, z, 0.08333333333333323)) + x;
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (z <= -5.5)
		tmp = fma(Float64(Float64(0.07512208616047561 / z) - -0.0692910599291889), y, x);
	elseif (z <= 5.0)
		tmp = Float64(Float64(y * fma(-0.00277777777751721, z, 0.08333333333333323)) + x);
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[z, -5.5], N[(N[(N[(0.07512208616047561 / z), $MachinePrecision] - -0.0692910599291889), $MachinePrecision] * y + x), $MachinePrecision], If[LessEqual[z, 5.0], N[(N[(y * N[(-0.00277777777751721 * z + 0.08333333333333323), $MachinePrecision]), $MachinePrecision] + x), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]]

Derivation

1. Split input into 3 regimes

2. if z < -5.5

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000} + \frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \color{blue}{\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \color{blue}{\frac{1}{z}}, y, x\right) \]

      3. lower-/.f64 64.8

         \[\leadsto \mathsf{fma}\left(0.0692910599291889 + 0.07512208616047561 \cdot \frac{1}{\color{blue}{z}}, y, x\right) \]

   6. Applied rewrites 64.8%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889 + 0.07512208616047561 \cdot \frac{1}{z}}, y, x\right) \]
   7. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{692910599291889}{10000000000000000} + \color{blue}{\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z}}, y, x\right) \]

      2. +-commutative N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} + \color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]

      3. add-flip N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \color{blue}{\left(\mathsf{neg}\left(\frac{692910599291889}{10000000000000000}\right)\right)}, y, x\right) \]

      4. metadata-eval N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      5. lower--.f64 64.8

         \[\leadsto \mathsf{fma}\left(0.07512208616047561 \cdot \frac{1}{z} - \color{blue}{-0.0692910599291889}, y, x\right) \]

      6. lift-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      7. lift-/.f64 N/A

         \[\leadsto \mathsf{fma}\left(\frac{751220861604756070699018739433}{10000000000000000000000000000000} \cdot \frac{1}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      8. mult-flip-rev N/A

         \[\leadsto \mathsf{fma}\left(\frac{\frac{751220861604756070699018739433}{10000000000000000000000000000000}}{z} - \frac{-692910599291889}{10000000000000000}, y, x\right) \]

      9. lower-/.f64 64.8

         \[\leadsto \mathsf{fma}\left(\frac{0.07512208616047561}{z} - -0.0692910599291889, y, x\right) \]

   8. Applied rewrites 64.8%

      \[\leadsto \mathsf{fma}\left(\frac{0.07512208616047561}{z} - \color{blue}{-0.0692910599291889}, y, x\right) \]

   if -5.5 < z < 5

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   2. Taylor expanded in z around 0

      \[\leadsto x + \color{blue}{\left(\frac{279195317918525}{3350343815022304} \cdot y + z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right)} \]
   3. Step-by-step derivation

      1. lower-fma.f64 N/A

         \[\leadsto x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, \color{blue}{y}, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) \]

      2. lower-*.f64 N/A

         \[\leadsto x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) \]

      3. lower--.f64 N/A

         \[\leadsto x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) \]

      4. lower-*.f64 N/A

         \[\leadsto x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) \]

      5. lower-*.f64 66.3

         \[\leadsto x + \mathsf{fma}\left(0.08333333333333323, y, z \cdot \left(0.14677053705526136 \cdot y - 0.14954831483277858 \cdot y\right)\right) \]

   4. Applied rewrites 66.3%

      \[\leadsto x + \color{blue}{\mathsf{fma}\left(0.08333333333333323, y, z \cdot \left(0.14677053705526136 \cdot y - 0.14954831483277858 \cdot y\right)\right)} \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \color{blue}{x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right)} \]

      2. +-commutative N/A

         \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) + x} \]

      3. lower-+.f64 66.3

         \[\leadsto \color{blue}{\mathsf{fma}\left(0.08333333333333323, y, z \cdot \left(0.14677053705526136 \cdot y - 0.14954831483277858 \cdot y\right)\right) + x} \]

   6. Applied rewrites 66.3%

      \[\leadsto \color{blue}{y \cdot \mathsf{fma}\left(-0.00277777777751721, z, 0.08333333333333323\right) + x} \]

   if 5 < z

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 6: 98.9% accurate, 1.6× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;z \leq -5.5:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \mathbf{elif}\;z \leq 5:\\ \;\;\;\;y \cdot \mathsf{fma}\left(-0.00277777777751721, z, 0.08333333333333323\right) + x\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<= z -5.5)
   (fma 0.0692910599291889 y x)
   (if (<= z 5.0)
     (+ (* y (fma -0.00277777777751721 z 0.08333333333333323)) x)
     (fma 0.0692910599291889 y x))))

C

double code(double x, double y, double z) {
	double tmp;
	if (z <= -5.5) {
		tmp = fma(0.0692910599291889, y, x);
	} else if (z <= 5.0) {
		tmp = (y * fma(-0.00277777777751721, z, 0.08333333333333323)) + x;
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (z <= -5.5)
		tmp = fma(0.0692910599291889, y, x);
	elseif (z <= 5.0)
		tmp = Float64(Float64(y * fma(-0.00277777777751721, z, 0.08333333333333323)) + x);
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[z, -5.5], N[(0.0692910599291889 * y + x), $MachinePrecision], If[LessEqual[z, 5.0], N[(N[(y * N[(-0.00277777777751721 * z + 0.08333333333333323), $MachinePrecision]), $MachinePrecision] + x), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]]

Derivation

1. Split input into 2 regimes

2. if z < -5.5 or 5 < z

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]

   if -5.5 < z < 5

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   2. Taylor expanded in z around 0

      \[\leadsto x + \color{blue}{\left(\frac{279195317918525}{3350343815022304} \cdot y + z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right)} \]
   3. Step-by-step derivation

      1. lower-fma.f64 N/A

         \[\leadsto x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, \color{blue}{y}, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) \]

      2. lower-*.f64 N/A

         \[\leadsto x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) \]

      3. lower--.f64 N/A

         \[\leadsto x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) \]

      4. lower-*.f64 N/A

         \[\leadsto x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) \]

      5. lower-*.f64 66.3

         \[\leadsto x + \mathsf{fma}\left(0.08333333333333323, y, z \cdot \left(0.14677053705526136 \cdot y - 0.14954831483277858 \cdot y\right)\right) \]

   4. Applied rewrites 66.3%

      \[\leadsto x + \color{blue}{\mathsf{fma}\left(0.08333333333333323, y, z \cdot \left(0.14677053705526136 \cdot y - 0.14954831483277858 \cdot y\right)\right)} \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \color{blue}{x + \mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right)} \]

      2. +-commutative N/A

         \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{279195317918525}{3350343815022304}, y, z \cdot \left(\frac{307332350656623}{2093964884388940} \cdot y - \frac{1678650474502018223880473708075}{11224803678858206361900017468416} \cdot y\right)\right) + x} \]

      3. lower-+.f64 66.3

         \[\leadsto \color{blue}{\mathsf{fma}\left(0.08333333333333323, y, z \cdot \left(0.14677053705526136 \cdot y - 0.14954831483277858 \cdot y\right)\right) + x} \]

   6. Applied rewrites 66.3%

      \[\leadsto \color{blue}{y \cdot \mathsf{fma}\left(-0.00277777777751721, z, 0.08333333333333323\right) + x} \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 7: 98.9% accurate, 2.1× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;z \leq -5.5:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \mathbf{elif}\;z \leq 5.9:\\ \;\;\;\;x - -0.08333333333333323 \cdot y\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<= z -5.5)
   (fma 0.0692910599291889 y x)
   (if (<= z 5.9)
     (- x (* -0.08333333333333323 y))
     (fma 0.0692910599291889 y x))))

C

double code(double x, double y, double z) {
	double tmp;
	if (z <= -5.5) {
		tmp = fma(0.0692910599291889, y, x);
	} else if (z <= 5.9) {
		tmp = x - (-0.08333333333333323 * y);
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (z <= -5.5)
		tmp = fma(0.0692910599291889, y, x);
	elseif (z <= 5.9)
		tmp = Float64(x - Float64(-0.08333333333333323 * y));
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[z, -5.5], N[(0.0692910599291889 * y + x), $MachinePrecision], If[LessEqual[z, 5.9], N[(x - N[(-0.08333333333333323 * y), $MachinePrecision]), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]]

Derivation

1. Split input into 2 regimes

2. if z < -5.5 or 5.9000000000000004 < z

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]

   if -5.5 < z < 5.9000000000000004

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   2. Taylor expanded in z around 0

      \[\leadsto x + \frac{\color{blue}{\frac{11167812716741}{40000000000000} \cdot y + z \cdot \left(\frac{692910599291889}{10000000000000000} \cdot \left(y \cdot z\right) + \frac{307332350656623}{625000000000000} \cdot y\right)}}{\left(z + \frac{6012459259764103}{1000000000000000}\right) \cdot z + \frac{104698244219447}{31250000000000}} \]
   3. Step-by-step derivation

      1. lower-fma.f64 N/A

         \[\leadsto x + \frac{\mathsf{fma}\left(\frac{11167812716741}{40000000000000}, \color{blue}{y}, z \cdot \left(\frac{692910599291889}{10000000000000000} \cdot \left(y \cdot z\right) + \frac{307332350656623}{625000000000000} \cdot y\right)\right)}{\left(z + \frac{6012459259764103}{1000000000000000}\right) \cdot z + \frac{104698244219447}{31250000000000}} \]

      2. lower-*.f64 N/A

         \[\leadsto x + \frac{\mathsf{fma}\left(\frac{11167812716741}{40000000000000}, y, z \cdot \left(\frac{692910599291889}{10000000000000000} \cdot \left(y \cdot z\right) + \frac{307332350656623}{625000000000000} \cdot y\right)\right)}{\left(z + \frac{6012459259764103}{1000000000000000}\right) \cdot z + \frac{104698244219447}{31250000000000}} \]

      3. lower-fma.f64 N/A

         \[\leadsto x + \frac{\mathsf{fma}\left(\frac{11167812716741}{40000000000000}, y, z \cdot \mathsf{fma}\left(\frac{692910599291889}{10000000000000000}, y \cdot z, \frac{307332350656623}{625000000000000} \cdot y\right)\right)}{\left(z + \frac{6012459259764103}{1000000000000000}\right) \cdot z + \frac{104698244219447}{31250000000000}} \]

      4. lower-*.f64 N/A

         \[\leadsto x + \frac{\mathsf{fma}\left(\frac{11167812716741}{40000000000000}, y, z \cdot \mathsf{fma}\left(\frac{692910599291889}{10000000000000000}, y \cdot z, \frac{307332350656623}{625000000000000} \cdot y\right)\right)}{\left(z + \frac{6012459259764103}{1000000000000000}\right) \cdot z + \frac{104698244219447}{31250000000000}} \]

      5. lower-*.f64 74.3

         \[\leadsto x + \frac{\mathsf{fma}\left(0.279195317918525, y, z \cdot \mathsf{fma}\left(0.0692910599291889, y \cdot z, 0.4917317610505968 \cdot y\right)\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   4. Applied rewrites 74.3%

      \[\leadsto x + \frac{\color{blue}{\mathsf{fma}\left(0.279195317918525, y, z \cdot \mathsf{fma}\left(0.0692910599291889, y \cdot z, 0.4917317610505968 \cdot y\right)\right)}}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   5. Taylor expanded in z around inf

      \[\leadsto x + \frac{\mathsf{fma}\left(\frac{11167812716741}{40000000000000}, y, z \cdot \left(\frac{692910599291889}{10000000000000000} \cdot \left(y \cdot z\right)\right)\right)}{\left(z + \frac{6012459259764103}{1000000000000000}\right) \cdot z + \frac{104698244219447}{31250000000000}} \]
   6. Step-by-step derivation

      1. lower-*.f64 N/A

         \[\leadsto x + \frac{\mathsf{fma}\left(\frac{11167812716741}{40000000000000}, y, z \cdot \left(\frac{692910599291889}{10000000000000000} \cdot \left(y \cdot z\right)\right)\right)}{\left(z + \frac{6012459259764103}{1000000000000000}\right) \cdot z + \frac{104698244219447}{31250000000000}} \]

      2. lower-*.f64 73.4

         \[\leadsto x + \frac{\mathsf{fma}\left(0.279195317918525, y, z \cdot \left(0.0692910599291889 \cdot \left(y \cdot z\right)\right)\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   7. Applied rewrites 73.4%

      \[\leadsto x + \frac{\mathsf{fma}\left(0.279195317918525, y, z \cdot \left(0.0692910599291889 \cdot \left(y \cdot z\right)\right)\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   9. Applied rewrites 73.4%

      \[\leadsto \color{blue}{x - \frac{\mathsf{fma}\left(\left(z \cdot y\right) \cdot 0.0692910599291889, z, 0.279195317918525 \cdot y\right)}{\mathsf{fma}\left(-6.012459259764103 - z, z, -3.350343815022304\right)}} \]

   10. Taylor expanded in z around 0

       \[\leadsto x - \color{blue}{\frac{-279195317918525}{3350343815022304} \cdot y} \]
   11. Step-by-step derivation

       1. lower-*.f64 79.9

          \[\leadsto x - -0.08333333333333323 \cdot \color{blue}{y} \]

   12. Applied rewrites 79.9%

       \[\leadsto x - \color{blue}{-0.08333333333333323 \cdot y} \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 8: 98.1% accurate, 2.2× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;z \leq -5.5:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \mathbf{elif}\;z \leq 5.9:\\ \;\;\;\;\mathsf{fma}\left(0.08333333333333323, y, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<= z -5.5)
   (fma 0.0692910599291889 y x)
   (if (<= z 5.9) (fma 0.08333333333333323 y x) (fma 0.0692910599291889 y x))))

C

double code(double x, double y, double z) {
	double tmp;
	if (z <= -5.5) {
		tmp = fma(0.0692910599291889, y, x);
	} else if (z <= 5.9) {
		tmp = fma(0.08333333333333323, y, x);
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (z <= -5.5)
		tmp = fma(0.0692910599291889, y, x);
	elseif (z <= 5.9)
		tmp = fma(0.08333333333333323, y, x);
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[z, -5.5], N[(0.0692910599291889 * y + x), $MachinePrecision], If[LessEqual[z, 5.9], N[(0.08333333333333323 * y + x), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]]

Derivation

1. Split input into 2 regimes

2. if z < -5.5 or 5.9000000000000004 < z

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]

   if -5.5 < z < 5.9000000000000004

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around 0

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{279195317918525}{3350343815022304}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.9%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.08333333333333323}, y, x\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 9: 97.7% accurate, 0.5× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \leq \infty:\\ \;\;\;\;\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889 \cdot z, z, 0.279195317918525\right), \frac{y}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(0.0692910599291889, y, x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z)
 :precision binary64
 (if (<=
      (+
       x
       (/
        (*
         y
         (+
          (* (+ (* z 0.0692910599291889) 0.4917317610505968) z)
          0.279195317918525))
        (+ (* (+ z 6.012459259764103) z) 3.350343815022304)))
      INFINITY)
   (fma
    (fma (* 0.0692910599291889 z) z 0.279195317918525)
    (/ y (fma (- z -6.012459259764103) z 3.350343815022304))
    x)
   (fma 0.0692910599291889 y x)))

C

double code(double x, double y, double z) {
	double tmp;
	if ((x + ((y * ((((z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / (((z + 6.012459259764103) * z) + 3.350343815022304))) <= ((double) INFINITY)) {
		tmp = fma(fma((0.0692910599291889 * z), z, 0.279195317918525), (y / fma((z - -6.012459259764103), z, 3.350343815022304)), x);
	} else {
		tmp = fma(0.0692910599291889, y, x);
	}
	return tmp;
}

Julia

function code(x, y, z)
	tmp = 0.0
	if (Float64(x + Float64(Float64(y * Float64(Float64(Float64(Float64(z * 0.0692910599291889) + 0.4917317610505968) * z) + 0.279195317918525)) / Float64(Float64(Float64(z + 6.012459259764103) * z) + 3.350343815022304))) <= Inf)
		tmp = fma(fma(Float64(0.0692910599291889 * z), z, 0.279195317918525), Float64(y / fma(Float64(z - -6.012459259764103), z, 3.350343815022304)), x);
	else
		tmp = fma(0.0692910599291889, y, x);
	end
	return tmp
end

Wolfram

code[x_, y_, z_] := If[LessEqual[N[(x + N[(N[(y * N[(N[(N[(N[(z * 0.0692910599291889), $MachinePrecision] + 0.4917317610505968), $MachinePrecision] * z), $MachinePrecision] + 0.279195317918525), $MachinePrecision]), $MachinePrecision] / N[(N[(N[(z + 6.012459259764103), $MachinePrecision] * z), $MachinePrecision] + 3.350343815022304), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], Infinity], N[(N[(N[(0.0692910599291889 * z), $MachinePrecision] * z + 0.279195317918525), $MachinePrecision] * N[(y / N[(N[(z - -6.012459259764103), $MachinePrecision] * z + 3.350343815022304), $MachinePrecision]), $MachinePrecision] + x), $MachinePrecision], N[(0.0692910599291889 * y + x), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if (+.f64 x (/.f64 (*.f64 y (+.f64 (*.f64 (+.f64 (*.f64 z #s(literal 692910599291889/10000000000000000 binary64)) #s(literal 307332350656623/625000000000000 binary64)) z) #s(literal 11167812716741/40000000000000 binary64))) (+.f64 (*.f64 (+.f64 z #s(literal 6012459259764103/1000000000000000 binary64)) z) #s(literal 104698244219447/31250000000000 binary64)))) < +inf.0

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 73.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right), \frac{y}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000} \cdot z}, z, \frac{11167812716741}{40000000000000}\right), \frac{y}{\mathsf{fma}\left(z - \frac{-6012459259764103}{1000000000000000}, z, \frac{104698244219447}{31250000000000}\right)}, x\right) \]
   5. Step-by-step derivation

      1. lower-*.f64 72.5

         \[\leadsto \mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889 \cdot \color{blue}{z}, z, 0.279195317918525\right), \frac{y}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, x\right) \]

   6. Applied rewrites 72.5%

      \[\leadsto \mathsf{fma}\left(\mathsf{fma}\left(\color{blue}{0.0692910599291889 \cdot z}, z, 0.279195317918525\right), \frac{y}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, x\right) \]

   if +inf.0 < (+.f64 x (/.f64 (*.f64 y (+.f64 (*.f64 (+.f64 (*.f64 z #s(literal 692910599291889/10000000000000000 binary64)) #s(literal 307332350656623/625000000000000 binary64)) z) #s(literal 11167812716741/40000000000000 binary64))) (+.f64 (*.f64 (+.f64 z #s(literal 6012459259764103/1000000000000000 binary64)) z) #s(literal 104698244219447/31250000000000 binary64))))

   1. Initial program 69.1%

      \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

   3. Applied rewrites 74.5%

      \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

   4. Taylor expanded in z around inf

      \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
   5. Step-by-step derivation

   6. Applied rewrites 79.1%

      \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 10: 79.1% accurate, 5.0× speedup

Math

\[\begin{array}{l} \\ \mathsf{fma}\left(0.0692910599291889, y, x\right) \end{array} \]

FPCore

(FPCore (x y z) :precision binary64 (fma 0.0692910599291889 y x))

C

double code(double x, double y, double z) {
	return fma(0.0692910599291889, y, x);
}

Julia

function code(x, y, z)
	return fma(0.0692910599291889, y, x)
end

Wolfram

code[x_, y_, z_] := N[(0.0692910599291889 * y + x), $MachinePrecision]

Derivation

1. Initial program 69.1%

   \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

3. Applied rewrites 74.5%

   \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{fma}\left(\mathsf{fma}\left(0.0692910599291889, z, 0.4917317610505968\right), z, 0.279195317918525\right)}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, y, x\right)} \]

4. Taylor expanded in z around inf

   \[\leadsto \mathsf{fma}\left(\color{blue}{\frac{692910599291889}{10000000000000000}}, y, x\right) \]
5. Step-by-step derivation

6. Applied rewrites 79.1%

   \[\leadsto \mathsf{fma}\left(\color{blue}{0.0692910599291889}, y, x\right) \]
7. Add Preprocessing

Alternative 11: 30.6% accurate, 7.5× speedup

Math

\[\begin{array}{l} \\ 0.0692910599291889 \cdot y \end{array} \]

FPCore

(FPCore (x y z) :precision binary64 (* 0.0692910599291889 y))

C

double code(double x, double y, double z) {
	return 0.0692910599291889 * y;
}

Fortran

module fmin_fmax_functions
    implicit none
    private
    public fmax
    public fmin

    interface fmax
        module procedure fmax88
        module procedure fmax44
        module procedure fmax84
        module procedure fmax48
    end interface
    interface fmin
        module procedure fmin88
        module procedure fmin44
        module procedure fmin84
        module procedure fmin48
    end interface
contains
    real(8) function fmax88(x, y) result (res)
        real(8), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(x, max(x, y), y /= y), x /= x)
    end function
    real(4) function fmax44(x, y) result (res)
        real(4), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(y, merge(x, max(x, y), y /= y), x /= x)
    end function
    real(8) function fmax84(x, y) result(res)
        real(8), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
    end function
    real(8) function fmax48(x, y) result(res)
        real(4), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
    end function
    real(8) function fmin88(x, y) result (res)
        real(8), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(x, min(x, y), y /= y), x /= x)
    end function
    real(4) function fmin44(x, y) result (res)
        real(4), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(y, merge(x, min(x, y), y /= y), x /= x)
    end function
    real(8) function fmin84(x, y) result(res)
        real(8), intent (in) :: x
        real(4), intent (in) :: y
        res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
    end function
    real(8) function fmin48(x, y) result(res)
        real(4), intent (in) :: x
        real(8), intent (in) :: y
        res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
    end function
end module

real(8) function code(x, y, z)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    code = 0.0692910599291889d0 * y
end function

Java

public static double code(double x, double y, double z) {
	return 0.0692910599291889 * y;
}

Python

def code(x, y, z):
	return 0.0692910599291889 * y

Julia

function code(x, y, z)
	return Float64(0.0692910599291889 * y)
end

MATLAB

function tmp = code(x, y, z)
	tmp = 0.0692910599291889 * y;
end

Wolfram

code[x_, y_, z_] := N[(0.0692910599291889 * y), $MachinePrecision]

Derivation

1. Initial program 69.1%

   \[x + \frac{y \cdot \left(\left(z \cdot 0.0692910599291889 + 0.4917317610505968\right) \cdot z + 0.279195317918525\right)}{\left(z + 6.012459259764103\right) \cdot z + 3.350343815022304} \]

3. Applied rewrites 82.4%

   \[\leadsto \color{blue}{\mathsf{fma}\left(y, \frac{0.279195317918525}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}, x - z \cdot \frac{\mathsf{fma}\left(-0.0692910599291889, z, -0.4917317610505968\right) \cdot y}{\mathsf{fma}\left(z - -6.012459259764103, z, 3.350343815022304\right)}\right)} \]

4. Taylor expanded in z around inf

   \[\leadsto \color{blue}{x - \frac{-692910599291889}{10000000000000000} \cdot y} \]
5. Step-by-step derivation

   1. lower--.f64 N/A

      \[\leadsto x - \color{blue}{\frac{-692910599291889}{10000000000000000} \cdot y} \]

   2. lower-*.f64 79.1

      \[\leadsto x - -0.0692910599291889 \cdot \color{blue}{y} \]

6. Applied rewrites 79.1%

   \[\leadsto \color{blue}{x - -0.0692910599291889 \cdot y} \]

7. Taylor expanded in x around 0

   \[\leadsto \frac{692910599291889}{10000000000000000} \cdot \color{blue}{y} \]
8. Step-by-step derivation

   1. lower-*.f64 30.6

      \[\leadsto 0.0692910599291889 \cdot y \]

9. Applied rewrites 30.6%

   \[\leadsto 0.0692910599291889 \cdot \color{blue}{y} \]
10. Add Preprocessing

Reproduce

herbie shell --seed 2025156 
(FPCore (x y z)
  :name "Numeric.SpecFunctions:logGamma from math-functions-0.1.5.2, B"
  :precision binary64
  (+ x (/ (* y (+ (* (+ (* z 0.0692910599291889) 0.4917317610505968) z) 0.279195317918525)) (+ (* (+ z 6.012459259764103) z) 3.350343815022304))))