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

Percentage Accurate: 99.8% → 99.8%

Time: 7.0s

Alternatives: 22

Speedup: 1.0×

Specification

Math

\[\begin{array}{l} \\ \left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (+ (+ (+ (+ (+ (* x (log y)) z) t) a) (* (- b 0.5) (log c))) (* y i)))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    code = (((((x * log(y)) + z) + t) + a) + ((b - 0.5d0) * log(c))) + (y * i)
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (((((x * Math.log(y)) + z) + t) + a) + ((b - 0.5) * Math.log(c))) + (y * i);
}

Python

def code(x, y, z, t, a, b, c, i):
	return (((((x * math.log(y)) + z) + t) + a) + ((b - 0.5) * math.log(c))) + (y * i)

Julia

function code(x, y, z, t, a, b, c, i)
	return Float64(Float64(Float64(Float64(Float64(Float64(x * log(y)) + z) + t) + a) + Float64(Float64(b - 0.5) * log(c))) + Float64(y * i))
end

MATLAB

function tmp = code(x, y, z, t, a, b, c, i)
	tmp = (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := N[(N[(N[(N[(N[(N[(x * N[Log[y], $MachinePrecision]), $MachinePrecision] + z), $MachinePrecision] + t), $MachinePrecision] + a), $MachinePrecision] + N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]

Initial Program: 99.8% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (+ (+ (+ (+ (+ (* x (log y)) z) t) a) (* (- b 0.5) (log c))) (* y i)))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    code = (((((x * log(y)) + z) + t) + a) + ((b - 0.5d0) * log(c))) + (y * i)
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (((((x * Math.log(y)) + z) + t) + a) + ((b - 0.5) * Math.log(c))) + (y * i);
}

Python

def code(x, y, z, t, a, b, c, i):
	return (((((x * math.log(y)) + z) + t) + a) + ((b - 0.5) * math.log(c))) + (y * i)

Julia

function code(x, y, z, t, a, b, c, i)
	return Float64(Float64(Float64(Float64(Float64(Float64(x * log(y)) + z) + t) + a) + Float64(Float64(b - 0.5) * log(c))) + Float64(y * i))
end

MATLAB

function tmp = code(x, y, z, t, a, b, c, i)
	tmp = (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := N[(N[(N[(N[(N[(N[(x * N[Log[y], $MachinePrecision]), $MachinePrecision] + z), $MachinePrecision] + t), $MachinePrecision] + a), $MachinePrecision] + N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]

Alternative 1: 99.8% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (+ (+ (+ (+ (+ (* x (log y)) z) t) a) (* (- b 0.5) (log c))) (* y i)))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    code = (((((x * log(y)) + z) + t) + a) + ((b - 0.5d0) * log(c))) + (y * i)
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (((((x * Math.log(y)) + z) + t) + a) + ((b - 0.5) * Math.log(c))) + (y * i);
}

Python

def code(x, y, z, t, a, b, c, i):
	return (((((x * math.log(y)) + z) + t) + a) + ((b - 0.5) * math.log(c))) + (y * i)

Julia

function code(x, y, z, t, a, b, c, i)
	return Float64(Float64(Float64(Float64(Float64(Float64(x * log(y)) + z) + t) + a) + Float64(Float64(b - 0.5) * log(c))) + Float64(y * i))
end

MATLAB

function tmp = code(x, y, z, t, a, b, c, i)
	tmp = (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := N[(N[(N[(N[(N[(N[(x * N[Log[y], $MachinePrecision]), $MachinePrecision] + z), $MachinePrecision] + t), $MachinePrecision] + a), $MachinePrecision] + N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]

Derivation

1. Initial program 99.8%

   \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]
2. Add Preprocessing

Alternative 2: 22.2% accurate, 0.3× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i\\ \mathbf{if}\;t\_1 \leq -1 \cdot 10^{+307}:\\ \;\;\;\;i \cdot y\\ \mathbf{elif}\;t\_1 \leq -200:\\ \;\;\;\;z\\ \mathbf{elif}\;t\_1 \leq \infty:\\ \;\;\;\;a\\ \mathbf{else}:\\ \;\;\;\;i \cdot y\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1
         (+
          (+ (+ (+ (+ (* x (log y)) z) t) a) (* (- b 0.5) (log c)))
          (* y i))))
   (if (<= t_1 -1e+307)
     (* i y)
     (if (<= t_1 -200.0) z (if (<= t_1 INFINITY) a (* i y))))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
	double tmp;
	if (t_1 <= -1e+307) {
		tmp = i * y;
	} else if (t_1 <= -200.0) {
		tmp = z;
	} else if (t_1 <= ((double) INFINITY)) {
		tmp = a;
	} else {
		tmp = i * y;
	}
	return tmp;
}

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = (((((x * Math.log(y)) + z) + t) + a) + ((b - 0.5) * Math.log(c))) + (y * i);
	double tmp;
	if (t_1 <= -1e+307) {
		tmp = i * y;
	} else if (t_1 <= -200.0) {
		tmp = z;
	} else if (t_1 <= Double.POSITIVE_INFINITY) {
		tmp = a;
	} else {
		tmp = i * y;
	}
	return tmp;
}

Python

def code(x, y, z, t, a, b, c, i):
	t_1 = (((((x * math.log(y)) + z) + t) + a) + ((b - 0.5) * math.log(c))) + (y * i)
	tmp = 0
	if t_1 <= -1e+307:
		tmp = i * y
	elif t_1 <= -200.0:
		tmp = z
	elif t_1 <= math.inf:
		tmp = a
	else:
		tmp = i * y
	return tmp

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(Float64(Float64(Float64(Float64(Float64(x * log(y)) + z) + t) + a) + Float64(Float64(b - 0.5) * log(c))) + Float64(y * i))
	tmp = 0.0
	if (t_1 <= -1e+307)
		tmp = Float64(i * y);
	elseif (t_1 <= -200.0)
		tmp = z;
	elseif (t_1 <= Inf)
		tmp = a;
	else
		tmp = Float64(i * y);
	end
	return tmp
end

MATLAB

function tmp_2 = code(x, y, z, t, a, b, c, i)
	t_1 = (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
	tmp = 0.0;
	if (t_1 <= -1e+307)
		tmp = i * y;
	elseif (t_1 <= -200.0)
		tmp = z;
	elseif (t_1 <= Inf)
		tmp = a;
	else
		tmp = i * y;
	end
	tmp_2 = tmp;
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[(N[(N[(N[(N[(x * N[Log[y], $MachinePrecision]), $MachinePrecision] + z), $MachinePrecision] + t), $MachinePrecision] + a), $MachinePrecision] + N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$1, -1e+307], N[(i * y), $MachinePrecision], If[LessEqual[t$95$1, -200.0], z, If[LessEqual[t$95$1, Infinity], a, N[(i * y), $MachinePrecision]]]]]

Derivation

1. Split input into 3 regimes

2. if (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i)) < -9.99999999999999986e306 or +inf.0 < (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i))

   1. Initial program 98.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in y around inf

      \[\leadsto \color{blue}{i \cdot y} \]
   3. Step-by-step derivation

      1. lower-*.f64 89.8

         \[\leadsto i \cdot \color{blue}{y} \]

   4. Applied rewrites 89.8%

      \[\leadsto \color{blue}{i \cdot y} \]

   if -9.99999999999999986e306 < (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i)) < -200

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in z around inf

      \[\leadsto \color{blue}{z} \]
   3. Step-by-step derivation

   4. Applied rewrites 17.9%

      \[\leadsto \color{blue}{z} \]

   if -200 < (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i)) < +inf.0

   1. Initial program 99.9%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in a around inf

      \[\leadsto \color{blue}{a} \]
   3. Step-by-step derivation

   4. Applied rewrites 15.7%

      \[\leadsto \color{blue}{a} \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 3: 27.0% accurate, 0.5× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i\\ \mathbf{if}\;t\_1 \leq -200:\\ \;\;\;\;\mathsf{fma}\left(y, i, z\right)\\ \mathbf{elif}\;t\_1 \leq \infty:\\ \;\;\;\;a\\ \mathbf{else}:\\ \;\;\;\;i \cdot y\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1
         (+
          (+ (+ (+ (+ (* x (log y)) z) t) a) (* (- b 0.5) (log c)))
          (* y i))))
   (if (<= t_1 -200.0) (fma y i z) (if (<= t_1 INFINITY) a (* i y)))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = (((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i);
	double tmp;
	if (t_1 <= -200.0) {
		tmp = fma(y, i, z);
	} else if (t_1 <= ((double) INFINITY)) {
		tmp = a;
	} else {
		tmp = i * y;
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(Float64(Float64(Float64(Float64(Float64(x * log(y)) + z) + t) + a) + Float64(Float64(b - 0.5) * log(c))) + Float64(y * i))
	tmp = 0.0
	if (t_1 <= -200.0)
		tmp = fma(y, i, z);
	elseif (t_1 <= Inf)
		tmp = a;
	else
		tmp = Float64(i * y);
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[(N[(N[(N[(N[(x * N[Log[y], $MachinePrecision]), $MachinePrecision] + z), $MachinePrecision] + t), $MachinePrecision] + a), $MachinePrecision] + N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$1, -200.0], N[(y * i + z), $MachinePrecision], If[LessEqual[t$95$1, Infinity], a, N[(i * y), $MachinePrecision]]]]

Derivation

1. Split input into 3 regimes

2. if (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i)) < -200

   1. Initial program 99.9%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in z around inf

      \[\leadsto \color{blue}{z} + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 38.5%

      \[\leadsto \color{blue}{z} + y \cdot i \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \color{blue}{z + y \cdot i} \]

      2. +-commutative N/A

         \[\leadsto \color{blue}{y \cdot i + z} \]

      3. lift-*.f64 N/A

         \[\leadsto \color{blue}{y \cdot i} + z \]

      4. lower-fma.f64 38.5

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

   6. Applied rewrites 38.5%

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

   if -200 < (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i)) < +inf.0

   1. Initial program 99.9%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in a around inf

      \[\leadsto \color{blue}{a} \]
   3. Step-by-step derivation

   4. Applied rewrites 15.7%

      \[\leadsto \color{blue}{a} \]

   if +inf.0 < (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i))

   1. Initial program 0.0%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in y around inf

      \[\leadsto \color{blue}{i \cdot y} \]
   3. Step-by-step derivation

      1. lower-*.f64 100.0

         \[\leadsto i \cdot \color{blue}{y} \]

   4. Applied rewrites 100.0%

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

Alternative 4: 78.1% accurate, 0.7× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \left(b - 0.5\right) \cdot \log c\\ t_2 := \left(\left(z + a\right) + t\_1\right) + y \cdot i\\ \mathbf{if}\;t\_1 \leq -2 \cdot 10^{+178}:\\ \;\;\;\;t\_2\\ \mathbf{elif}\;t\_1 \leq 5 \cdot 10^{+51}:\\ \;\;\;\;\left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right)\\ \mathbf{else}:\\ \;\;\;\;t\_2\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1 (* (- b 0.5) (log c))) (t_2 (+ (+ (+ z a) t_1) (* y i))))
   (if (<= t_1 -2e+178)
     t_2
     (if (<= t_1 5e+51) (+ (+ a z) (fma i y (* (log y) x))) t_2))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = (b - 0.5) * log(c);
	double t_2 = ((z + a) + t_1) + (y * i);
	double tmp;
	if (t_1 <= -2e+178) {
		tmp = t_2;
	} else if (t_1 <= 5e+51) {
		tmp = (a + z) + fma(i, y, (log(y) * x));
	} else {
		tmp = t_2;
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(Float64(b - 0.5) * log(c))
	t_2 = Float64(Float64(Float64(z + a) + t_1) + Float64(y * i))
	tmp = 0.0
	if (t_1 <= -2e+178)
		tmp = t_2;
	elseif (t_1 <= 5e+51)
		tmp = Float64(Float64(a + z) + fma(i, y, Float64(log(y) * x)));
	else
		tmp = t_2;
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]}, Block[{t$95$2 = N[(N[(N[(z + a), $MachinePrecision] + t$95$1), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$1, -2e+178], t$95$2, If[LessEqual[t$95$1, 5e+51], N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], t$95$2]]]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < -2.0000000000000001e178 or 5e51 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))

   1. Initial program 99.6%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in z around inf

      \[\leadsto \left(\left(\color{blue}{z} + a\right) + \left(b - \frac{1}{2}\right) \cdot \log c\right) + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 80.5%

      \[\leadsto \left(\left(\color{blue}{z} + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   if -2.0000000000000001e178 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < 5e51

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 81.8

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 81.8%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in x around inf

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, x \cdot \log y\right) \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      3. lift-*.f64 77.0

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   7. Applied rewrites 77.0%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 5: 76.8% accurate, 0.7× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \left(b - 0.5\right) \cdot \log c\\ \mathbf{if}\;t\_1 \leq -2 \cdot 10^{+178}:\\ \;\;\;\;\left(\left(t + a\right) + t\_1\right) + y \cdot i\\ \mathbf{elif}\;t\_1 \leq 2 \cdot 10^{+203}:\\ \;\;\;\;\left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right)\\ \mathbf{else}:\\ \;\;\;\;\left(a + t\_1\right) + y \cdot i\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1 (* (- b 0.5) (log c))))
   (if (<= t_1 -2e+178)
     (+ (+ (+ t a) t_1) (* y i))
     (if (<= t_1 2e+203)
       (+ (+ a z) (fma i y (* (log y) x)))
       (+ (+ a t_1) (* y i))))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = (b - 0.5) * log(c);
	double tmp;
	if (t_1 <= -2e+178) {
		tmp = ((t + a) + t_1) + (y * i);
	} else if (t_1 <= 2e+203) {
		tmp = (a + z) + fma(i, y, (log(y) * x));
	} else {
		tmp = (a + t_1) + (y * i);
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(Float64(b - 0.5) * log(c))
	tmp = 0.0
	if (t_1 <= -2e+178)
		tmp = Float64(Float64(Float64(t + a) + t_1) + Float64(y * i));
	elseif (t_1 <= 2e+203)
		tmp = Float64(Float64(a + z) + fma(i, y, Float64(log(y) * x)));
	else
		tmp = Float64(Float64(a + t_1) + Float64(y * i));
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$1, -2e+178], N[(N[(N[(t + a), $MachinePrecision] + t$95$1), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision], If[LessEqual[t$95$1, 2e+203], N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], N[(N[(a + t$95$1), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]]]]

Derivation

1. Split input into 3 regimes

2. if (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < -2.0000000000000001e178

   1. Initial program 99.6%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around inf

      \[\leadsto \left(\left(\color{blue}{t} + a\right) + \left(b - \frac{1}{2}\right) \cdot \log c\right) + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 84.3%

      \[\leadsto \left(\left(\color{blue}{t} + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   if -2.0000000000000001e178 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < 2e203

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 82.6

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 82.6%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in x around inf

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, x \cdot \log y\right) \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      3. lift-*.f64 75.4

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   7. Applied rewrites 75.4%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   if 2e203 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))

   1. Initial program 99.3%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in a around inf

      \[\leadsto \left(\color{blue}{a} + \left(b - \frac{1}{2}\right) \cdot \log c\right) + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 80.3%

      \[\leadsto \left(\color{blue}{a} + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 6: 76.1% accurate, 0.7× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \left(b - 0.5\right) \cdot \log c\\ t_2 := \left(a + t\_1\right) + y \cdot i\\ \mathbf{if}\;t\_1 \leq -2 \cdot 10^{+178}:\\ \;\;\;\;t\_2\\ \mathbf{elif}\;t\_1 \leq 2 \cdot 10^{+203}:\\ \;\;\;\;\left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right)\\ \mathbf{else}:\\ \;\;\;\;t\_2\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1 (* (- b 0.5) (log c))) (t_2 (+ (+ a t_1) (* y i))))
   (if (<= t_1 -2e+178)
     t_2
     (if (<= t_1 2e+203) (+ (+ a z) (fma i y (* (log y) x))) t_2))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = (b - 0.5) * log(c);
	double t_2 = (a + t_1) + (y * i);
	double tmp;
	if (t_1 <= -2e+178) {
		tmp = t_2;
	} else if (t_1 <= 2e+203) {
		tmp = (a + z) + fma(i, y, (log(y) * x));
	} else {
		tmp = t_2;
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(Float64(b - 0.5) * log(c))
	t_2 = Float64(Float64(a + t_1) + Float64(y * i))
	tmp = 0.0
	if (t_1 <= -2e+178)
		tmp = t_2;
	elseif (t_1 <= 2e+203)
		tmp = Float64(Float64(a + z) + fma(i, y, Float64(log(y) * x)));
	else
		tmp = t_2;
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]}, Block[{t$95$2 = N[(N[(a + t$95$1), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$1, -2e+178], t$95$2, If[LessEqual[t$95$1, 2e+203], N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], t$95$2]]]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < -2.0000000000000001e178 or 2e203 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))

   1. Initial program 99.5%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in a around inf

      \[\leadsto \left(\color{blue}{a} + \left(b - \frac{1}{2}\right) \cdot \log c\right) + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 78.9%

      \[\leadsto \left(\color{blue}{a} + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   if -2.0000000000000001e178 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < 2e203

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 82.6

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 82.6%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in x around inf

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, x \cdot \log y\right) \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      3. lift-*.f64 75.4

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   7. Applied rewrites 75.4%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 7: 75.0% accurate, 0.7× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \left(b - 0.5\right) \cdot \log c\\ \mathbf{if}\;t\_1 \leq -2 \cdot 10^{+178}:\\ \;\;\;\;\mathsf{fma}\left(y, i, \mathsf{fma}\left(\log c, b - 0.5, z\right)\right)\\ \mathbf{elif}\;t\_1 \leq 2 \cdot 10^{+249}:\\ \;\;\;\;\left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right)\\ \mathbf{else}:\\ \;\;\;\;\mathsf{fma}\left(y, i, \mathsf{fma}\left(\log c, b - 0.5, t\right)\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1 (* (- b 0.5) (log c))))
   (if (<= t_1 -2e+178)
     (fma y i (fma (log c) (- b 0.5) z))
     (if (<= t_1 2e+249)
       (+ (+ a z) (fma i y (* (log y) x)))
       (fma y i (fma (log c) (- b 0.5) t))))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = (b - 0.5) * log(c);
	double tmp;
	if (t_1 <= -2e+178) {
		tmp = fma(y, i, fma(log(c), (b - 0.5), z));
	} else if (t_1 <= 2e+249) {
		tmp = (a + z) + fma(i, y, (log(y) * x));
	} else {
		tmp = fma(y, i, fma(log(c), (b - 0.5), t));
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(Float64(b - 0.5) * log(c))
	tmp = 0.0
	if (t_1 <= -2e+178)
		tmp = fma(y, i, fma(log(c), Float64(b - 0.5), z));
	elseif (t_1 <= 2e+249)
		tmp = Float64(Float64(a + z) + fma(i, y, Float64(log(y) * x)));
	else
		tmp = fma(y, i, fma(log(c), Float64(b - 0.5), t));
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$1, -2e+178], N[(y * i + N[(N[Log[c], $MachinePrecision] * N[(b - 0.5), $MachinePrecision] + z), $MachinePrecision]), $MachinePrecision], If[LessEqual[t$95$1, 2e+249], N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], N[(y * i + N[(N[Log[c], $MachinePrecision] * N[(b - 0.5), $MachinePrecision] + t), $MachinePrecision]), $MachinePrecision]]]]

Derivation

1. Split input into 3 regimes

2. if (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < -2.0000000000000001e178

   1. Initial program 99.6%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in z around inf

      \[\leadsto \left(\color{blue}{z} + \left(b - \frac{1}{2}\right) \cdot \log c\right) + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 78.0%

      \[\leadsto \left(\color{blue}{z} + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \color{blue}{\left(z + \left(b - \frac{1}{2}\right) \cdot \log c\right) + y \cdot i} \]

      2. +-commutative N/A

         \[\leadsto \color{blue}{y \cdot i + \left(z + \left(b - \frac{1}{2}\right) \cdot \log c\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \color{blue}{y \cdot i} + \left(z + \left(b - \frac{1}{2}\right) \cdot \log c\right) \]

      4. lower-fma.f64 78.0

         \[\leadsto \color{blue}{\mathsf{fma}\left(y, i, z + \left(b - 0.5\right) \cdot \log c\right)} \]

      5. lift-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(y, i, \color{blue}{z + \left(b - \frac{1}{2}\right) \cdot \log c}\right) \]

      6. lift--.f64 N/A

         \[\leadsto \mathsf{fma}\left(y, i, z + \color{blue}{\left(b - \frac{1}{2}\right)} \cdot \log c\right) \]

      7. lift-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(y, i, z + \color{blue}{\left(b - \frac{1}{2}\right) \cdot \log c}\right) \]

      8. lift-log.f64 N/A

         \[\leadsto \mathsf{fma}\left(y, i, z + \left(b - \frac{1}{2}\right) \cdot \color{blue}{\log c}\right) \]

      9. +-commutative N/A

         \[\leadsto \mathsf{fma}\left(y, i, \color{blue}{\left(b - \frac{1}{2}\right) \cdot \log c + z}\right) \]

      10. *-commutative N/A

          \[\leadsto \mathsf{fma}\left(y, i, \color{blue}{\log c \cdot \left(b - \frac{1}{2}\right)} + z\right) \]

      11. lower-fma.f64 N/A

          \[\leadsto \mathsf{fma}\left(y, i, \color{blue}{\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right)}\right) \]

      12. lift-log.f64 N/A

          \[\leadsto \mathsf{fma}\left(y, i, \mathsf{fma}\left(\color{blue}{\log c}, b - \frac{1}{2}, z\right)\right) \]

      13. lift--.f64 78.0

          \[\leadsto \mathsf{fma}\left(y, i, \mathsf{fma}\left(\log c, \color{blue}{b - 0.5}, z\right)\right) \]

   6. Applied rewrites 78.0%

      \[\leadsto \color{blue}{\mathsf{fma}\left(y, i, \mathsf{fma}\left(\log c, b - 0.5, z\right)\right)} \]

   if -2.0000000000000001e178 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < 1.9999999999999998e249

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 82.8

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 82.8%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in x around inf

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, x \cdot \log y\right) \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      3. lift-*.f64 73.9

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   7. Applied rewrites 73.9%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   if 1.9999999999999998e249 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))

   1. Initial program 99.0%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around inf

      \[\leadsto \left(\color{blue}{t} + \left(b - \frac{1}{2}\right) \cdot \log c\right) + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 87.0%

      \[\leadsto \left(\color{blue}{t} + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \color{blue}{\left(t + \left(b - \frac{1}{2}\right) \cdot \log c\right) + y \cdot i} \]

      2. +-commutative N/A

         \[\leadsto \color{blue}{y \cdot i + \left(t + \left(b - \frac{1}{2}\right) \cdot \log c\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \color{blue}{y \cdot i} + \left(t + \left(b - \frac{1}{2}\right) \cdot \log c\right) \]

      4. lower-fma.f64 87.0

         \[\leadsto \color{blue}{\mathsf{fma}\left(y, i, t + \left(b - 0.5\right) \cdot \log c\right)} \]

      5. lift-+.f64 N/A

         \[\leadsto \mathsf{fma}\left(y, i, \color{blue}{t + \left(b - \frac{1}{2}\right) \cdot \log c}\right) \]

      6. lift--.f64 N/A

         \[\leadsto \mathsf{fma}\left(y, i, t + \color{blue}{\left(b - \frac{1}{2}\right)} \cdot \log c\right) \]

      7. lift-*.f64 N/A

         \[\leadsto \mathsf{fma}\left(y, i, t + \color{blue}{\left(b - \frac{1}{2}\right) \cdot \log c}\right) \]

      8. lift-log.f64 N/A

         \[\leadsto \mathsf{fma}\left(y, i, t + \left(b - \frac{1}{2}\right) \cdot \color{blue}{\log c}\right) \]

      9. +-commutative N/A

         \[\leadsto \mathsf{fma}\left(y, i, \color{blue}{\left(b - \frac{1}{2}\right) \cdot \log c + t}\right) \]

      10. *-commutative N/A

          \[\leadsto \mathsf{fma}\left(y, i, \color{blue}{\log c \cdot \left(b - \frac{1}{2}\right)} + t\right) \]

      11. lower-fma.f64 N/A

          \[\leadsto \mathsf{fma}\left(y, i, \color{blue}{\mathsf{fma}\left(\log c, b - \frac{1}{2}, t\right)}\right) \]

      12. lift-log.f64 N/A

          \[\leadsto \mathsf{fma}\left(y, i, \mathsf{fma}\left(\color{blue}{\log c}, b - \frac{1}{2}, t\right)\right) \]

      13. lift--.f64 87.0

          \[\leadsto \mathsf{fma}\left(y, i, \mathsf{fma}\left(\log c, \color{blue}{b - 0.5}, t\right)\right) \]

   6. Applied rewrites 87.0%

      \[\leadsto \color{blue}{\mathsf{fma}\left(y, i, \mathsf{fma}\left(\log c, b - 0.5, t\right)\right)} \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 8: 72.7% accurate, 0.7× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := a + \log c \cdot b\\ t_2 := \left(b - 0.5\right) \cdot \log c\\ \mathbf{if}\;t\_2 \leq -2 \cdot 10^{+194}:\\ \;\;\;\;t\_1\\ \mathbf{elif}\;t\_2 \leq 2 \cdot 10^{+203}:\\ \;\;\;\;\left(\left(z + t\right) + a\right) + y \cdot i\\ \mathbf{else}:\\ \;\;\;\;t\_1\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1 (+ a (* (log c) b))) (t_2 (* (- b 0.5) (log c))))
   (if (<= t_2 -2e+194)
     t_1
     (if (<= t_2 2e+203) (+ (+ (+ z t) a) (* y i)) t_1))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = a + (log(c) * b);
	double t_2 = (b - 0.5) * log(c);
	double tmp;
	if (t_2 <= -2e+194) {
		tmp = t_1;
	} else if (t_2 <= 2e+203) {
		tmp = ((z + t) + a) + (y * i);
	} else {
		tmp = t_1;
	}
	return tmp;
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    real(8) :: t_1
    real(8) :: t_2
    real(8) :: tmp
    t_1 = a + (log(c) * b)
    t_2 = (b - 0.5d0) * log(c)
    if (t_2 <= (-2d+194)) then
        tmp = t_1
    else if (t_2 <= 2d+203) then
        tmp = ((z + t) + a) + (y * i)
    else
        tmp = t_1
    end if
    code = tmp
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = a + (Math.log(c) * b);
	double t_2 = (b - 0.5) * Math.log(c);
	double tmp;
	if (t_2 <= -2e+194) {
		tmp = t_1;
	} else if (t_2 <= 2e+203) {
		tmp = ((z + t) + a) + (y * i);
	} else {
		tmp = t_1;
	}
	return tmp;
}

Python

def code(x, y, z, t, a, b, c, i):
	t_1 = a + (math.log(c) * b)
	t_2 = (b - 0.5) * math.log(c)
	tmp = 0
	if t_2 <= -2e+194:
		tmp = t_1
	elif t_2 <= 2e+203:
		tmp = ((z + t) + a) + (y * i)
	else:
		tmp = t_1
	return tmp

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(a + Float64(log(c) * b))
	t_2 = Float64(Float64(b - 0.5) * log(c))
	tmp = 0.0
	if (t_2 <= -2e+194)
		tmp = t_1;
	elseif (t_2 <= 2e+203)
		tmp = Float64(Float64(Float64(z + t) + a) + Float64(y * i));
	else
		tmp = t_1;
	end
	return tmp
end

MATLAB

function tmp_2 = code(x, y, z, t, a, b, c, i)
	t_1 = a + (log(c) * b);
	t_2 = (b - 0.5) * log(c);
	tmp = 0.0;
	if (t_2 <= -2e+194)
		tmp = t_1;
	elseif (t_2 <= 2e+203)
		tmp = ((z + t) + a) + (y * i);
	else
		tmp = t_1;
	end
	tmp_2 = tmp;
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(a + N[(N[Log[c], $MachinePrecision] * b), $MachinePrecision]), $MachinePrecision]}, Block[{t$95$2 = N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$2, -2e+194], t$95$1, If[LessEqual[t$95$2, 2e+203], N[(N[(N[(z + t), $MachinePrecision] + a), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision], t$95$1]]]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < -1.99999999999999989e194 or 2e203 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))

   1. Initial program 99.4%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 92.9

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 92.9%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]
   5. Step-by-step derivation

      1. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      2. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      3. lift-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

      4. lift-*.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

      5. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

      6. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      7. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      8. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      9. lower-*.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      10. lift-log.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      11. lift--.f64 92.9

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - 0.5\right)\right)\right) \]

   6. Applied rewrites 92.9%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - 0.5\right)\right)\right) \]

   7. Taylor expanded in b around inf

      \[\leadsto \left(a + z\right) + b \cdot \color{blue}{\log c} \]
   8. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \log c \cdot b \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \log c \cdot b \]

      3. lift-*.f64 70.4

         \[\leadsto \left(a + z\right) + \log c \cdot b \]

   9. Applied rewrites 70.4%

      \[\leadsto \left(a + z\right) + \log c \cdot \color{blue}{b} \]

   10. Taylor expanded in z around 0

       \[\leadsto a + \color{blue}{\log c} \cdot b \]
   11. Step-by-step derivation

   12. Applied rewrites 64.0%

       \[\leadsto a + \color{blue}{\log c} \cdot b \]

   if -1.99999999999999989e194 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < 2e203

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in x around 0

      \[\leadsto \color{blue}{\left(a + \left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} + y \cdot i \]
   3. Step-by-step derivation

      1. +-commutative N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      2. lower-+.f64 N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      3. +-commutative N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      4. lower-+.f64 N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      5. +-commutative N/A

         \[\leadsto \left(\left(\left(\log c \cdot \left(b - \frac{1}{2}\right) + z\right) + t\right) + a\right) + y \cdot i \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      7. lift-log.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      8. lift--.f64 82.4

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right) + y \cdot i \]

   4. Applied rewrites 82.4%

      \[\leadsto \color{blue}{\left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right)} + y \cdot i \]

   5. Taylor expanded in z around inf

      \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
   6. Step-by-step derivation

   7. Applied rewrites 74.7%

      \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 9: 71.5% accurate, 0.7× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \log c \cdot b\\ t_2 := \left(b - 0.5\right) \cdot \log c\\ \mathbf{if}\;t\_2 \leq -2 \cdot 10^{+194}:\\ \;\;\;\;t\_1\\ \mathbf{elif}\;t\_2 \leq 2 \cdot 10^{+249}:\\ \;\;\;\;\left(\left(z + t\right) + a\right) + y \cdot i\\ \mathbf{else}:\\ \;\;\;\;t\_1\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1 (* (log c) b)) (t_2 (* (- b 0.5) (log c))))
   (if (<= t_2 -2e+194)
     t_1
     (if (<= t_2 2e+249) (+ (+ (+ z t) a) (* y i)) t_1))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = log(c) * b;
	double t_2 = (b - 0.5) * log(c);
	double tmp;
	if (t_2 <= -2e+194) {
		tmp = t_1;
	} else if (t_2 <= 2e+249) {
		tmp = ((z + t) + a) + (y * i);
	} else {
		tmp = t_1;
	}
	return tmp;
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    real(8) :: t_1
    real(8) :: t_2
    real(8) :: tmp
    t_1 = log(c) * b
    t_2 = (b - 0.5d0) * log(c)
    if (t_2 <= (-2d+194)) then
        tmp = t_1
    else if (t_2 <= 2d+249) then
        tmp = ((z + t) + a) + (y * i)
    else
        tmp = t_1
    end if
    code = tmp
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = Math.log(c) * b;
	double t_2 = (b - 0.5) * Math.log(c);
	double tmp;
	if (t_2 <= -2e+194) {
		tmp = t_1;
	} else if (t_2 <= 2e+249) {
		tmp = ((z + t) + a) + (y * i);
	} else {
		tmp = t_1;
	}
	return tmp;
}

Python

def code(x, y, z, t, a, b, c, i):
	t_1 = math.log(c) * b
	t_2 = (b - 0.5) * math.log(c)
	tmp = 0
	if t_2 <= -2e+194:
		tmp = t_1
	elif t_2 <= 2e+249:
		tmp = ((z + t) + a) + (y * i)
	else:
		tmp = t_1
	return tmp

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(log(c) * b)
	t_2 = Float64(Float64(b - 0.5) * log(c))
	tmp = 0.0
	if (t_2 <= -2e+194)
		tmp = t_1;
	elseif (t_2 <= 2e+249)
		tmp = Float64(Float64(Float64(z + t) + a) + Float64(y * i));
	else
		tmp = t_1;
	end
	return tmp
end

MATLAB

function tmp_2 = code(x, y, z, t, a, b, c, i)
	t_1 = log(c) * b;
	t_2 = (b - 0.5) * log(c);
	tmp = 0.0;
	if (t_2 <= -2e+194)
		tmp = t_1;
	elseif (t_2 <= 2e+249)
		tmp = ((z + t) + a) + (y * i);
	else
		tmp = t_1;
	end
	tmp_2 = tmp;
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[Log[c], $MachinePrecision] * b), $MachinePrecision]}, Block[{t$95$2 = N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$2, -2e+194], t$95$1, If[LessEqual[t$95$2, 2e+249], N[(N[(N[(z + t), $MachinePrecision] + a), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision], t$95$1]]]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < -1.99999999999999989e194 or 1.9999999999999998e249 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))

   1. Initial program 99.4%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in b around inf

      \[\leadsto \color{blue}{b \cdot \log c} \]
   3. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \log c \cdot \color{blue}{b} \]

      2. lower-*.f64 N/A

         \[\leadsto \log c \cdot \color{blue}{b} \]

      3. lift-log.f64 60.8

         \[\leadsto \log c \cdot b \]

   4. Applied rewrites 60.8%

      \[\leadsto \color{blue}{\log c \cdot b} \]

   if -1.99999999999999989e194 < (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c)) < 1.9999999999999998e249

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in x around 0

      \[\leadsto \color{blue}{\left(a + \left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} + y \cdot i \]
   3. Step-by-step derivation

      1. +-commutative N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      2. lower-+.f64 N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      3. +-commutative N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      4. lower-+.f64 N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      5. +-commutative N/A

         \[\leadsto \left(\left(\left(\log c \cdot \left(b - \frac{1}{2}\right) + z\right) + t\right) + a\right) + y \cdot i \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      7. lift-log.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      8. lift--.f64 82.8

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right) + y \cdot i \]

   4. Applied rewrites 82.8%

      \[\leadsto \color{blue}{\left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right)} + y \cdot i \]

   5. Taylor expanded in z around inf

      \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
   6. Step-by-step derivation

   7. Applied rewrites 73.3%

      \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 10: 38.4% accurate, 0.9× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \leq -200:\\ \;\;\;\;\mathsf{fma}\left(y, i, z\right)\\ \mathbf{else}:\\ \;\;\;\;a + y \cdot i\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (if (<=
      (+ (+ (+ (+ (+ (* x (log y)) z) t) a) (* (- b 0.5) (log c))) (* y i))
      -200.0)
   (fma y i z)
   (+ a (* y i))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (((((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i)) <= -200.0) {
		tmp = fma(y, i, z);
	} else {
		tmp = a + (y * i);
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	tmp = 0.0
	if (Float64(Float64(Float64(Float64(Float64(Float64(x * log(y)) + z) + t) + a) + Float64(Float64(b - 0.5) * log(c))) + Float64(y * i)) <= -200.0)
		tmp = fma(y, i, z);
	else
		tmp = Float64(a + Float64(y * i));
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := If[LessEqual[N[(N[(N[(N[(N[(N[(x * N[Log[y], $MachinePrecision]), $MachinePrecision] + z), $MachinePrecision] + t), $MachinePrecision] + a), $MachinePrecision] + N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision], -200.0], N[(y * i + z), $MachinePrecision], N[(a + N[(y * i), $MachinePrecision]), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i)) < -200

   1. Initial program 99.9%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in z around inf

      \[\leadsto \color{blue}{z} + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 38.5%

      \[\leadsto \color{blue}{z} + y \cdot i \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \color{blue}{z + y \cdot i} \]

      2. +-commutative N/A

         \[\leadsto \color{blue}{y \cdot i + z} \]

      3. lift-*.f64 N/A

         \[\leadsto \color{blue}{y \cdot i} + z \]

      4. lower-fma.f64 38.5

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

   6. Applied rewrites 38.5%

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

   if -200 < (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i))

   1. Initial program 99.7%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in a around inf

      \[\leadsto \color{blue}{a} + y \cdot i \]
   3. Step-by-step derivation

   4. Applied rewrites 38.3%

      \[\leadsto \color{blue}{a} + y \cdot i \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 11: 15.7% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \leq -200:\\ \;\;\;\;z\\ \mathbf{else}:\\ \;\;\;\;a\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (if (<=
      (+ (+ (+ (+ (+ (* x (log y)) z) t) a) (* (- b 0.5) (log c))) (* y i))
      -200.0)
   z
   a))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (((((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i)) <= -200.0) {
		tmp = z;
	} else {
		tmp = a;
	}
	return tmp;
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    real(8) :: tmp
    if (((((((x * log(y)) + z) + t) + a) + ((b - 0.5d0) * log(c))) + (y * i)) <= (-200.0d0)) then
        tmp = z
    else
        tmp = a
    end if
    code = tmp
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (((((((x * Math.log(y)) + z) + t) + a) + ((b - 0.5) * Math.log(c))) + (y * i)) <= -200.0) {
		tmp = z;
	} else {
		tmp = a;
	}
	return tmp;
}

Python

def code(x, y, z, t, a, b, c, i):
	tmp = 0
	if ((((((x * math.log(y)) + z) + t) + a) + ((b - 0.5) * math.log(c))) + (y * i)) <= -200.0:
		tmp = z
	else:
		tmp = a
	return tmp

Julia

function code(x, y, z, t, a, b, c, i)
	tmp = 0.0
	if (Float64(Float64(Float64(Float64(Float64(Float64(x * log(y)) + z) + t) + a) + Float64(Float64(b - 0.5) * log(c))) + Float64(y * i)) <= -200.0)
		tmp = z;
	else
		tmp = a;
	end
	return tmp
end

MATLAB

function tmp_2 = code(x, y, z, t, a, b, c, i)
	tmp = 0.0;
	if (((((((x * log(y)) + z) + t) + a) + ((b - 0.5) * log(c))) + (y * i)) <= -200.0)
		tmp = z;
	else
		tmp = a;
	end
	tmp_2 = tmp;
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := If[LessEqual[N[(N[(N[(N[(N[(N[(x * N[Log[y], $MachinePrecision]), $MachinePrecision] + z), $MachinePrecision] + t), $MachinePrecision] + a), $MachinePrecision] + N[(N[(b - 0.5), $MachinePrecision] * N[Log[c], $MachinePrecision]), $MachinePrecision]), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision], -200.0], z, a]

Derivation

1. Split input into 2 regimes

2. if (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i)) < -200

   1. Initial program 99.9%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in z around inf

      \[\leadsto \color{blue}{z} \]
   3. Step-by-step derivation

   4. Applied rewrites 15.7%

      \[\leadsto \color{blue}{z} \]

   if -200 < (+.f64 (+.f64 (+.f64 (+.f64 (+.f64 (*.f64 x (log.f64 y)) z) t) a) (*.f64 (-.f64 b #s(literal 1/2 binary64)) (log.f64 c))) (*.f64 y i))

   1. Initial program 99.7%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in a around inf

      \[\leadsto \color{blue}{a} \]
   3. Step-by-step derivation

   4. Applied rewrites 15.7%

      \[\leadsto \color{blue}{a} \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 12: 85.7% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;y \leq 600000:\\ \;\;\;\;\left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \left(b - 0.5\right)\right)\\ \mathbf{else}:\\ \;\;\;\;\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot -0.5\right)\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (if (<= y 600000.0)
   (+ (+ a t) (+ (fma (log y) x z) (* (log c) (- b 0.5))))
   (+ (+ a z) (fma i y (fma (log y) x (* (log c) -0.5))))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (y <= 600000.0) {
		tmp = (a + t) + (fma(log(y), x, z) + (log(c) * (b - 0.5)));
	} else {
		tmp = (a + z) + fma(i, y, fma(log(y), x, (log(c) * -0.5)));
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	tmp = 0.0
	if (y <= 600000.0)
		tmp = Float64(Float64(a + t) + Float64(fma(log(y), x, z) + Float64(log(c) * Float64(b - 0.5))));
	else
		tmp = Float64(Float64(a + z) + fma(i, y, fma(log(y), x, Float64(log(c) * -0.5))));
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := If[LessEqual[y, 600000.0], N[(N[(a + t), $MachinePrecision] + N[(N[(N[Log[y], $MachinePrecision] * x + z), $MachinePrecision] + N[(N[Log[c], $MachinePrecision] * N[(b - 0.5), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[y], $MachinePrecision] * x + N[(N[Log[c], $MachinePrecision] * -0.5), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if y < 6e5

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in y around 0

      \[\leadsto \color{blue}{a + \left(t + \left(z + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + t\right) + \color{blue}{\left(z + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + t\right) + \color{blue}{\left(z + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + t\right) + \left(\color{blue}{z} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. associate-+r+ N/A

         \[\leadsto \left(a + t\right) + \left(\left(z + x \cdot \log y\right) + \color{blue}{\log c \cdot \left(b - \frac{1}{2}\right)}\right) \]

      5. lower-+.f64 N/A

         \[\leadsto \left(a + t\right) + \left(\left(z + x \cdot \log y\right) + \color{blue}{\log c \cdot \left(b - \frac{1}{2}\right)}\right) \]

      6. +-commutative N/A

         \[\leadsto \left(a + t\right) + \left(\left(x \cdot \log y + z\right) + \color{blue}{\log c} \cdot \left(b - \frac{1}{2}\right)\right) \]

      7. *-commutative N/A

         \[\leadsto \left(a + t\right) + \left(\left(\log y \cdot x + z\right) + \log \color{blue}{c} \cdot \left(b - \frac{1}{2}\right)\right) \]

      8. lower-fma.f64 N/A

         \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \color{blue}{\log c} \cdot \left(b - \frac{1}{2}\right)\right) \]

      9. lift-log.f64 N/A

         \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log \color{blue}{c} \cdot \left(b - \frac{1}{2}\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \color{blue}{\left(b - \frac{1}{2}\right)}\right) \]

      11. lift-log.f64 N/A

          \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \left(\color{blue}{b} - \frac{1}{2}\right)\right) \]

      12. lift--.f64 94.9

          \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \left(b - \color{blue}{0.5}\right)\right) \]

   4. Applied rewrites 94.9%

      \[\leadsto \color{blue}{\left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \left(b - 0.5\right)\right)} \]

   if 6e5 < y

   1. Initial program 99.7%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 88.2

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 88.2%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]
   5. Step-by-step derivation

      1. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      2. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      3. lift-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

      4. lift-*.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

      5. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

      6. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      7. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      8. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      9. lower-*.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      10. lift-log.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      11. lift--.f64 88.2

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - 0.5\right)\right)\right) \]

   6. Applied rewrites 88.2%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - 0.5\right)\right)\right) \]

   7. Taylor expanded in b around 0

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \frac{-1}{2}\right)\right) \]
   8. Step-by-step derivation

   9. Applied rewrites 76.2%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot -0.5\right)\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 13: 85.4% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;y \leq 600000:\\ \;\;\;\;\left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \left(b - 0.5\right)\right)\\ \mathbf{else}:\\ \;\;\;\;\left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (if (<= y 600000.0)
   (+ (+ a t) (+ (fma (log y) x z) (* (log c) (- b 0.5))))
   (+ (+ a z) (fma i y (* (log y) x)))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (y <= 600000.0) {
		tmp = (a + t) + (fma(log(y), x, z) + (log(c) * (b - 0.5)));
	} else {
		tmp = (a + z) + fma(i, y, (log(y) * x));
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	tmp = 0.0
	if (y <= 600000.0)
		tmp = Float64(Float64(a + t) + Float64(fma(log(y), x, z) + Float64(log(c) * Float64(b - 0.5))));
	else
		tmp = Float64(Float64(a + z) + fma(i, y, Float64(log(y) * x)));
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := If[LessEqual[y, 600000.0], N[(N[(a + t), $MachinePrecision] + N[(N[(N[Log[y], $MachinePrecision] * x + z), $MachinePrecision] + N[(N[Log[c], $MachinePrecision] * N[(b - 0.5), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if y < 6e5

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in y around 0

      \[\leadsto \color{blue}{a + \left(t + \left(z + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + t\right) + \color{blue}{\left(z + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + t\right) + \color{blue}{\left(z + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + t\right) + \left(\color{blue}{z} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. associate-+r+ N/A

         \[\leadsto \left(a + t\right) + \left(\left(z + x \cdot \log y\right) + \color{blue}{\log c \cdot \left(b - \frac{1}{2}\right)}\right) \]

      5. lower-+.f64 N/A

         \[\leadsto \left(a + t\right) + \left(\left(z + x \cdot \log y\right) + \color{blue}{\log c \cdot \left(b - \frac{1}{2}\right)}\right) \]

      6. +-commutative N/A

         \[\leadsto \left(a + t\right) + \left(\left(x \cdot \log y + z\right) + \color{blue}{\log c} \cdot \left(b - \frac{1}{2}\right)\right) \]

      7. *-commutative N/A

         \[\leadsto \left(a + t\right) + \left(\left(\log y \cdot x + z\right) + \log \color{blue}{c} \cdot \left(b - \frac{1}{2}\right)\right) \]

      8. lower-fma.f64 N/A

         \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \color{blue}{\log c} \cdot \left(b - \frac{1}{2}\right)\right) \]

      9. lift-log.f64 N/A

         \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log \color{blue}{c} \cdot \left(b - \frac{1}{2}\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \color{blue}{\left(b - \frac{1}{2}\right)}\right) \]

      11. lift-log.f64 N/A

          \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \left(\color{blue}{b} - \frac{1}{2}\right)\right) \]

      12. lift--.f64 94.9

          \[\leadsto \left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \left(b - \color{blue}{0.5}\right)\right) \]

   4. Applied rewrites 94.9%

      \[\leadsto \color{blue}{\left(a + t\right) + \left(\mathsf{fma}\left(\log y, x, z\right) + \log c \cdot \left(b - 0.5\right)\right)} \]

   if 6e5 < y

   1. Initial program 99.7%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 88.2

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 88.2%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in x around inf

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, x \cdot \log y\right) \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      3. lift-*.f64 75.6

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   7. Applied rewrites 75.6%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 14: 76.0% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \log y \cdot x\\ \mathbf{if}\;y \leq 480000:\\ \;\;\;\;\left(a + z\right) + \mathsf{fma}\left(\log c, b - 0.5, t\_1\right)\\ \mathbf{else}:\\ \;\;\;\;\left(a + z\right) + \mathsf{fma}\left(i, y, t\_1\right)\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1 (* (log y) x)))
   (if (<= y 480000.0)
     (+ (+ a z) (fma (log c) (- b 0.5) t_1))
     (+ (+ a z) (fma i y t_1)))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = log(y) * x;
	double tmp;
	if (y <= 480000.0) {
		tmp = (a + z) + fma(log(c), (b - 0.5), t_1);
	} else {
		tmp = (a + z) + fma(i, y, t_1);
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(log(y) * x)
	tmp = 0.0
	if (y <= 480000.0)
		tmp = Float64(Float64(a + z) + fma(log(c), Float64(b - 0.5), t_1));
	else
		tmp = Float64(Float64(a + z) + fma(i, y, t_1));
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]}, If[LessEqual[y, 480000.0], N[(N[(a + z), $MachinePrecision] + N[(N[Log[c], $MachinePrecision] * N[(b - 0.5), $MachinePrecision] + t$95$1), $MachinePrecision]), $MachinePrecision], N[(N[(a + z), $MachinePrecision] + N[(i * y + t$95$1), $MachinePrecision]), $MachinePrecision]]]

Derivation

1. Split input into 2 regimes

2. if y < 4.8e5

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 81.1

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 81.1%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in y around 0

      \[\leadsto \left(a + z\right) + \left(x \cdot \log y + \color{blue}{\log c \cdot \left(b - \frac{1}{2}\right)}\right) \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \left(\log y \cdot x + \log c \cdot \left(\color{blue}{b} - \frac{1}{2}\right)\right) \]

      2. +-commutative N/A

         \[\leadsto \left(a + z\right) + \left(\log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot \color{blue}{x}\right) \]

      3. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

      4. lift-*.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

      5. lift-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(\log c, b - \color{blue}{\frac{1}{2}}, \log y \cdot x\right) \]

      6. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right) \]

      7. lift--.f64 76.4

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right) \]

   7. Applied rewrites 76.4%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(\log c, \color{blue}{b - 0.5}, \log y \cdot x\right) \]

   if 4.8e5 < y

   1. Initial program 99.7%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 88.2

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 88.2%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in x around inf

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, x \cdot \log y\right) \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      3. lift-*.f64 75.6

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   7. Applied rewrites 75.6%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 15: 84.6% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - 0.5\right)\right)\right) \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (+ (+ a z) (fma i y (fma (log y) x (* (log c) (- b 0.5))))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (a + z) + fma(i, y, fma(log(y), x, (log(c) * (b - 0.5))));
}

Julia

function code(x, y, z, t, a, b, c, i)
	return Float64(Float64(a + z) + fma(i, y, fma(log(y), x, Float64(log(c) * Float64(b - 0.5)))))
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[y], $MachinePrecision] * x + N[(N[Log[c], $MachinePrecision] * N[(b - 0.5), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]

Derivation

1. Initial program 99.8%

   \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

2. Taylor expanded in t around 0

   \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
3. Step-by-step derivation

   1. associate-+r+ N/A

      \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

   2. lower-+.f64 N/A

      \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

   3. lower-+.f64 N/A

      \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

   4. lower-fma.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

   5. +-commutative N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

   6. lower-fma.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   7. lift-log.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   8. lift--.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   9. *-commutative N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

   10. lower-*.f64 N/A

       \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

   11. lift-log.f64 84.6

       \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

4. Applied rewrites 84.6%

   \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]
5. Step-by-step derivation

   1. lift-log.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

   2. lift--.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

   3. lift-fma.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

   4. lift-*.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

   5. lift-log.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + \log y \cdot x\right) \]

   6. +-commutative N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

   7. lower-fma.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

   8. lift-log.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

   9. lower-*.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

   10. lift-log.f64 N/A

       \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

   11. lift--.f64 84.6

       \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - 0.5\right)\right)\right) \]

6. Applied rewrites 84.6%

   \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log y, x, \log c \cdot \left(b - 0.5\right)\right)\right) \]
7. Add Preprocessing

Alternative 16: 84.6% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (+ (+ a z) (fma i y (fma (log c) (- b 0.5) (* (log y) x)))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (a + z) + fma(i, y, fma(log(c), (b - 0.5), (log(y) * x)));
}

Julia

function code(x, y, z, t, a, b, c, i)
	return Float64(Float64(a + z) + fma(i, y, fma(log(c), Float64(b - 0.5), Float64(log(y) * x))))
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[c], $MachinePrecision] * N[(b - 0.5), $MachinePrecision] + N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]

Derivation

1. Initial program 99.8%

   \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

2. Taylor expanded in t around 0

   \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
3. Step-by-step derivation

   1. associate-+r+ N/A

      \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

   2. lower-+.f64 N/A

      \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

   3. lower-+.f64 N/A

      \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

   4. lower-fma.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

   5. +-commutative N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

   6. lower-fma.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   7. lift-log.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   8. lift--.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   9. *-commutative N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

   10. lower-*.f64 N/A

       \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

   11. lift-log.f64 84.6

       \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

4. Applied rewrites 84.6%

   \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]
5. Add Preprocessing

Alternative 17: 90.0% accurate, 1.7× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} t_1 := \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right)\\ \mathbf{if}\;x \leq -3.2 \cdot 10^{+162}:\\ \;\;\;\;t\_1\\ \mathbf{elif}\;x \leq 0.14:\\ \;\;\;\;\left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right) + y \cdot i\\ \mathbf{else}:\\ \;\;\;\;t\_1\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (let* ((t_1 (+ (+ a z) (fma i y (* (log y) x)))))
   (if (<= x -3.2e+162)
     t_1
     (if (<= x 0.14) (+ (+ (+ (fma (log c) (- b 0.5) z) t) a) (* y i)) t_1))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double t_1 = (a + z) + fma(i, y, (log(y) * x));
	double tmp;
	if (x <= -3.2e+162) {
		tmp = t_1;
	} else if (x <= 0.14) {
		tmp = ((fma(log(c), (b - 0.5), z) + t) + a) + (y * i);
	} else {
		tmp = t_1;
	}
	return tmp;
}

Julia

function code(x, y, z, t, a, b, c, i)
	t_1 = Float64(Float64(a + z) + fma(i, y, Float64(log(y) * x)))
	tmp = 0.0
	if (x <= -3.2e+162)
		tmp = t_1;
	elseif (x <= 0.14)
		tmp = Float64(Float64(Float64(fma(log(c), Float64(b - 0.5), z) + t) + a) + Float64(y * i));
	else
		tmp = t_1;
	end
	return tmp
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := Block[{t$95$1 = N[(N[(a + z), $MachinePrecision] + N[(i * y + N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]}, If[LessEqual[x, -3.2e+162], t$95$1, If[LessEqual[x, 0.14], N[(N[(N[(N[(N[Log[c], $MachinePrecision] * N[(b - 0.5), $MachinePrecision] + z), $MachinePrecision] + t), $MachinePrecision] + a), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision], t$95$1]]]

Derivation

1. Split input into 2 regimes

2. if x < -3.2000000000000001e162 or 0.14000000000000001 < x

   1. Initial program 99.6%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 88.7

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 88.7%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in x around inf

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, x \cdot \log y\right) \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

      3. lift-*.f64 77.5

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   7. Applied rewrites 77.5%

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log y \cdot x\right) \]

   if -3.2000000000000001e162 < x < 0.14000000000000001

   1. Initial program 99.9%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in x around 0

      \[\leadsto \color{blue}{\left(a + \left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} + y \cdot i \]
   3. Step-by-step derivation

      1. +-commutative N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      2. lower-+.f64 N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      3. +-commutative N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      4. lower-+.f64 N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      5. +-commutative N/A

         \[\leadsto \left(\left(\left(\log c \cdot \left(b - \frac{1}{2}\right) + z\right) + t\right) + a\right) + y \cdot i \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      7. lift-log.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      8. lift--.f64 97.2

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right) + y \cdot i \]

   4. Applied rewrites 97.2%

      \[\leadsto \color{blue}{\left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right)} + y \cdot i \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 18: 64.2% accurate, 2.0× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;y \leq 8 \cdot 10^{+38}:\\ \;\;\;\;\left(a + z\right) + \log y \cdot x\\ \mathbf{else}:\\ \;\;\;\;\left(\left(z + t\right) + a\right) + y \cdot i\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (if (<= y 8e+38) (+ (+ a z) (* (log y) x)) (+ (+ (+ z t) a) (* y i))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (y <= 8e+38) {
		tmp = (a + z) + (log(y) * x);
	} else {
		tmp = ((z + t) + a) + (y * i);
	}
	return tmp;
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    real(8) :: tmp
    if (y <= 8d+38) then
        tmp = (a + z) + (log(y) * x)
    else
        tmp = ((z + t) + a) + (y * i)
    end if
    code = tmp
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (y <= 8e+38) {
		tmp = (a + z) + (Math.log(y) * x);
	} else {
		tmp = ((z + t) + a) + (y * i);
	}
	return tmp;
}

Python

def code(x, y, z, t, a, b, c, i):
	tmp = 0
	if y <= 8e+38:
		tmp = (a + z) + (math.log(y) * x)
	else:
		tmp = ((z + t) + a) + (y * i)
	return tmp

Julia

function code(x, y, z, t, a, b, c, i)
	tmp = 0.0
	if (y <= 8e+38)
		tmp = Float64(Float64(a + z) + Float64(log(y) * x));
	else
		tmp = Float64(Float64(Float64(z + t) + a) + Float64(y * i));
	end
	return tmp
end

MATLAB

function tmp_2 = code(x, y, z, t, a, b, c, i)
	tmp = 0.0;
	if (y <= 8e+38)
		tmp = (a + z) + (log(y) * x);
	else
		tmp = ((z + t) + a) + (y * i);
	end
	tmp_2 = tmp;
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := If[LessEqual[y, 8e+38], N[(N[(a + z), $MachinePrecision] + N[(N[Log[y], $MachinePrecision] * x), $MachinePrecision]), $MachinePrecision], N[(N[(N[(z + t), $MachinePrecision] + a), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if y < 7.99999999999999982e38

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 81.3

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 81.3%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in x around inf

      \[\leadsto \left(a + z\right) + x \cdot \color{blue}{\log y} \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \log y \cdot x \]

      2. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \log y \cdot x \]

      3. lift-*.f64 54.5

         \[\leadsto \left(a + z\right) + \log y \cdot x \]

   7. Applied rewrites 54.5%

      \[\leadsto \left(a + z\right) + \log y \cdot \color{blue}{x} \]

   if 7.99999999999999982e38 < y

   1. Initial program 99.7%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in x around 0

      \[\leadsto \color{blue}{\left(a + \left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} + y \cdot i \]
   3. Step-by-step derivation

      1. +-commutative N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      2. lower-+.f64 N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      3. +-commutative N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      4. lower-+.f64 N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      5. +-commutative N/A

         \[\leadsto \left(\left(\left(\log c \cdot \left(b - \frac{1}{2}\right) + z\right) + t\right) + a\right) + y \cdot i \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      7. lift-log.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      8. lift--.f64 88.3

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right) + y \cdot i \]

   4. Applied rewrites 88.3%

      \[\leadsto \color{blue}{\left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right)} + y \cdot i \]

   5. Taylor expanded in z around inf

      \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
   6. Step-by-step derivation

   7. Applied rewrites 76.5%

      \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 19: 65.4% accurate, 2.0× speedup

Math

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;y \leq 5500000:\\ \;\;\;\;\left(a + z\right) + \log c \cdot b\\ \mathbf{else}:\\ \;\;\;\;\left(\left(z + t\right) + a\right) + y \cdot i\\ \end{array} \end{array} \]

FPCore

(FPCore (x y z t a b c i)
 :precision binary64
 (if (<= y 5500000.0) (+ (+ a z) (* (log c) b)) (+ (+ (+ z t) a) (* y i))))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (y <= 5500000.0) {
		tmp = (a + z) + (log(c) * b);
	} else {
		tmp = ((z + t) + a) + (y * i);
	}
	return tmp;
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    real(8) :: tmp
    if (y <= 5500000.0d0) then
        tmp = (a + z) + (log(c) * b)
    else
        tmp = ((z + t) + a) + (y * i)
    end if
    code = tmp
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	double tmp;
	if (y <= 5500000.0) {
		tmp = (a + z) + (Math.log(c) * b);
	} else {
		tmp = ((z + t) + a) + (y * i);
	}
	return tmp;
}

Python

def code(x, y, z, t, a, b, c, i):
	tmp = 0
	if y <= 5500000.0:
		tmp = (a + z) + (math.log(c) * b)
	else:
		tmp = ((z + t) + a) + (y * i)
	return tmp

Julia

function code(x, y, z, t, a, b, c, i)
	tmp = 0.0
	if (y <= 5500000.0)
		tmp = Float64(Float64(a + z) + Float64(log(c) * b));
	else
		tmp = Float64(Float64(Float64(z + t) + a) + Float64(y * i));
	end
	return tmp
end

MATLAB

function tmp_2 = code(x, y, z, t, a, b, c, i)
	tmp = 0.0;
	if (y <= 5500000.0)
		tmp = (a + z) + (log(c) * b);
	else
		tmp = ((z + t) + a) + (y * i);
	end
	tmp_2 = tmp;
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := If[LessEqual[y, 5500000.0], N[(N[(a + z), $MachinePrecision] + N[(N[Log[c], $MachinePrecision] * b), $MachinePrecision]), $MachinePrecision], N[(N[(N[(z + t), $MachinePrecision] + a), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if y < 5.5e6

   1. Initial program 99.8%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in t around 0

      \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
   3. Step-by-step derivation

      1. associate-+r+ N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      2. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

      3. lower-+.f64 N/A

         \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

      4. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

      5. +-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      7. lift-log.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      8. lift--.f64 N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

      9. *-commutative N/A

         \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      10. lower-*.f64 N/A

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

      11. lift-log.f64 81.1

          \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

   4. Applied rewrites 81.1%

      \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

   5. Taylor expanded in b around inf

      \[\leadsto \left(a + z\right) + b \cdot \color{blue}{\log c} \]
   6. Step-by-step derivation

      1. *-commutative N/A

         \[\leadsto \left(a + z\right) + \log c \cdot b \]

      2. lower-*.f64 N/A

         \[\leadsto \left(a + z\right) + \log c \cdot b \]

      3. lift-log.f64 56.0

         \[\leadsto \left(a + z\right) + \log c \cdot b \]

   7. Applied rewrites 56.0%

      \[\leadsto \left(a + z\right) + \log c \cdot \color{blue}{b} \]

   if 5.5e6 < y

   1. Initial program 99.7%

      \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

   2. Taylor expanded in x around 0

      \[\leadsto \color{blue}{\left(a + \left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} + y \cdot i \]
   3. Step-by-step derivation

      1. +-commutative N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      2. lower-+.f64 N/A

         \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

      3. +-commutative N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      4. lower-+.f64 N/A

         \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

      5. +-commutative N/A

         \[\leadsto \left(\left(\left(\log c \cdot \left(b - \frac{1}{2}\right) + z\right) + t\right) + a\right) + y \cdot i \]

      6. lower-fma.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      7. lift-log.f64 N/A

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

      8. lift--.f64 87.7

         \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right) + y \cdot i \]

   4. Applied rewrites 87.7%

      \[\leadsto \color{blue}{\left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right)} + y \cdot i \]

   5. Taylor expanded in z around inf

      \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
   6. Step-by-step derivation

   7. Applied rewrites 75.2%

      \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 20: 67.4% accurate, 15.6× speedup

Math

\[\begin{array}{l} \\ \left(\left(z + t\right) + a\right) + y \cdot i \end{array} \]

FPCore

(FPCore (x y z t a b c i) :precision binary64 (+ (+ (+ z t) a) (* y i)))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return ((z + t) + a) + (y * i);
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    code = ((z + t) + a) + (y * i)
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return ((z + t) + a) + (y * i);
}

Python

def code(x, y, z, t, a, b, c, i):
	return ((z + t) + a) + (y * i)

Julia

function code(x, y, z, t, a, b, c, i)
	return Float64(Float64(Float64(z + t) + a) + Float64(y * i))
end

MATLAB

function tmp = code(x, y, z, t, a, b, c, i)
	tmp = ((z + t) + a) + (y * i);
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := N[(N[(N[(z + t), $MachinePrecision] + a), $MachinePrecision] + N[(y * i), $MachinePrecision]), $MachinePrecision]

Derivation

1. Initial program 99.8%

   \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

2. Taylor expanded in x around 0

   \[\leadsto \color{blue}{\left(a + \left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} + y \cdot i \]
3. Step-by-step derivation

   1. +-commutative N/A

      \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

   2. lower-+.f64 N/A

      \[\leadsto \left(\left(t + \left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) + \color{blue}{a}\right) + y \cdot i \]

   3. +-commutative N/A

      \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

   4. lower-+.f64 N/A

      \[\leadsto \left(\left(\left(z + \log c \cdot \left(b - \frac{1}{2}\right)\right) + t\right) + a\right) + y \cdot i \]

   5. +-commutative N/A

      \[\leadsto \left(\left(\left(\log c \cdot \left(b - \frac{1}{2}\right) + z\right) + t\right) + a\right) + y \cdot i \]

   6. lower-fma.f64 N/A

      \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

   7. lift-log.f64 N/A

      \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - \frac{1}{2}, z\right) + t\right) + a\right) + y \cdot i \]

   8. lift--.f64 84.3

      \[\leadsto \left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right) + y \cdot i \]

4. Applied rewrites 84.3%

   \[\leadsto \color{blue}{\left(\left(\mathsf{fma}\left(\log c, b - 0.5, z\right) + t\right) + a\right)} + y \cdot i \]

5. Taylor expanded in z around inf

   \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
6. Step-by-step derivation

7. Applied rewrites 67.4%

   \[\leadsto \left(\left(z + t\right) + a\right) + y \cdot i \]
8. Add Preprocessing

Alternative 21: 52.9% accurate, 19.5× speedup

Math

\[\begin{array}{l} \\ \left(a + z\right) + i \cdot y \end{array} \]

FPCore

(FPCore (x y z t a b c i) :precision binary64 (+ (+ a z) (* i y)))

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (a + z) + (i * 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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    code = (a + z) + (i * y)
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return (a + z) + (i * y);
}

Python

def code(x, y, z, t, a, b, c, i):
	return (a + z) + (i * y)

Julia

function code(x, y, z, t, a, b, c, i)
	return Float64(Float64(a + z) + Float64(i * y))
end

MATLAB

function tmp = code(x, y, z, t, a, b, c, i)
	tmp = (a + z) + (i * y);
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := N[(N[(a + z), $MachinePrecision] + N[(i * y), $MachinePrecision]), $MachinePrecision]

Derivation

1. Initial program 99.8%

   \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

2. Taylor expanded in t around 0

   \[\leadsto \color{blue}{a + \left(z + \left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)\right)} \]
3. Step-by-step derivation

   1. associate-+r+ N/A

      \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

   2. lower-+.f64 N/A

      \[\leadsto \left(a + z\right) + \color{blue}{\left(i \cdot y + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right)} \]

   3. lower-+.f64 N/A

      \[\leadsto \left(a + z\right) + \left(\color{blue}{i \cdot y} + \left(x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right)\right) \]

   4. lower-fma.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, \color{blue}{y}, x \cdot \log y + \log c \cdot \left(b - \frac{1}{2}\right)\right) \]

   5. +-commutative N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \log c \cdot \left(b - \frac{1}{2}\right) + x \cdot \log y\right) \]

   6. lower-fma.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   7. lift-log.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   8. lift--.f64 N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, x \cdot \log y\right)\right) \]

   9. *-commutative N/A

      \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

   10. lower-*.f64 N/A

       \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - \frac{1}{2}, \log y \cdot x\right)\right) \]

   11. lift-log.f64 84.6

       \[\leadsto \left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right) \]

4. Applied rewrites 84.6%

   \[\leadsto \color{blue}{\left(a + z\right) + \mathsf{fma}\left(i, y, \mathsf{fma}\left(\log c, b - 0.5, \log y \cdot x\right)\right)} \]

5. Taylor expanded in y around inf

   \[\leadsto \left(a + z\right) + i \cdot \color{blue}{y} \]
6. Step-by-step derivation

   1. lower-*.f64 52.9

      \[\leadsto \left(a + z\right) + i \cdot y \]

7. Applied rewrites 52.9%

   \[\leadsto \left(a + z\right) + i \cdot \color{blue}{y} \]
8. Add Preprocessing

Alternative 22: 16.5% accurate, 234.0× speedup

Math

\[\begin{array}{l} \\ a \end{array} \]

FPCore

(FPCore (x y z t a b c i) :precision binary64 a)

C

double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return a;
}

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, t, a, b, c, i)
use fmin_fmax_functions
    real(8), intent (in) :: x
    real(8), intent (in) :: y
    real(8), intent (in) :: z
    real(8), intent (in) :: t
    real(8), intent (in) :: a
    real(8), intent (in) :: b
    real(8), intent (in) :: c
    real(8), intent (in) :: i
    code = a
end function

Java

public static double code(double x, double y, double z, double t, double a, double b, double c, double i) {
	return a;
}

Python

def code(x, y, z, t, a, b, c, i):
	return a

Julia

function code(x, y, z, t, a, b, c, i)
	return a
end

MATLAB

function tmp = code(x, y, z, t, a, b, c, i)
	tmp = a;
end

Wolfram

code[x_, y_, z_, t_, a_, b_, c_, i_] := a

Derivation

1. Initial program 99.8%

   \[\left(\left(\left(\left(x \cdot \log y + z\right) + t\right) + a\right) + \left(b - 0.5\right) \cdot \log c\right) + y \cdot i \]

2. Taylor expanded in a around inf

   \[\leadsto \color{blue}{a} \]
3. Step-by-step derivation

4. Applied rewrites 16.5%

   \[\leadsto \color{blue}{a} \]
5. Add Preprocessing

Reproduce

herbie shell --seed 2025093 
(FPCore (x y z t a b c i)
  :name "Numeric.SpecFunctions:logBeta from math-functions-0.1.5.2, B"
  :precision binary64
  (+ (+ (+ (+ (+ (* x (log y)) z) t) a) (* (- b 0.5) (log c))) (* y i)))