math.cos on complex, real part

Percentage Accurate: 100.0% → 100.0%

Time: 3.3s

Alternatives: 8

Speedup: 1.3×

• 50.3% of points produce a very large (infinite) output. You may want to add a precondition.

Specification

Math

\[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

FPCore

(FPCore (re im)
  :precision binary64
  (* (* 0.5 (cos re)) (+ (exp (- im)) (exp im))))

C

double code(double re, double im) {
	return (0.5 * cos(re)) * (exp(-im) + exp(im));
}

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(re, im)
use fmin_fmax_functions
    real(8), intent (in) :: re
    real(8), intent (in) :: im
    code = (0.5d0 * cos(re)) * (exp(-im) + exp(im))
end function

Java

public static double code(double re, double im) {
	return (0.5 * Math.cos(re)) * (Math.exp(-im) + Math.exp(im));
}

Python

def code(re, im):
	return (0.5 * math.cos(re)) * (math.exp(-im) + math.exp(im))

Julia

function code(re, im)
	return Float64(Float64(0.5 * cos(re)) * Float64(exp(Float64(-im)) + exp(im)))
end

MATLAB

function tmp = code(re, im)
	tmp = (0.5 * cos(re)) * (exp(-im) + exp(im));
end

Wolfram

code[re_, im_] := N[(N[(0.5 * N[Cos[re], $MachinePrecision]), $MachinePrecision] * N[(N[Exp[(-im)], $MachinePrecision] + N[Exp[im], $MachinePrecision]), $MachinePrecision]), $MachinePrecision]

Initial Program: 100.0% accurate, 1.0× speedup

Math

\[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

FPCore

(FPCore (re im)
  :precision binary64
  (* (* 0.5 (cos re)) (+ (exp (- im)) (exp im))))

C

double code(double re, double im) {
	return (0.5 * cos(re)) * (exp(-im) + exp(im));
}

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(re, im)
use fmin_fmax_functions
    real(8), intent (in) :: re
    real(8), intent (in) :: im
    code = (0.5d0 * cos(re)) * (exp(-im) + exp(im))
end function

Java

public static double code(double re, double im) {
	return (0.5 * Math.cos(re)) * (Math.exp(-im) + Math.exp(im));
}

Python

def code(re, im):
	return (0.5 * math.cos(re)) * (math.exp(-im) + math.exp(im))

Julia

function code(re, im)
	return Float64(Float64(0.5 * cos(re)) * Float64(exp(Float64(-im)) + exp(im)))
end

MATLAB

function tmp = code(re, im)
	tmp = (0.5 * cos(re)) * (exp(-im) + exp(im));
end

Wolfram

code[re_, im_] := N[(N[(0.5 * N[Cos[re], $MachinePrecision]), $MachinePrecision] * N[(N[Exp[(-im)], $MachinePrecision] + N[Exp[im], $MachinePrecision]), $MachinePrecision]), $MachinePrecision]

Alternative 1: 100.0% accurate, 1.3× speedup

Math

\[\cosh im \cdot \cos re \]

FPCore

(FPCore (re im)
  :precision binary64
  (* (cosh im) (cos re)))

C

double code(double re, double im) {
	return cosh(im) * cos(re);
}

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(re, im)
use fmin_fmax_functions
    real(8), intent (in) :: re
    real(8), intent (in) :: im
    code = cosh(im) * cos(re)
end function

Java

public static double code(double re, double im) {
	return Math.cosh(im) * Math.cos(re);
}

Python

def code(re, im):
	return math.cosh(im) * math.cos(re)

Julia

function code(re, im)
	return Float64(cosh(im) * cos(re))
end

MATLAB

function tmp = code(re, im)
	tmp = cosh(im) * cos(re);
end

Wolfram

code[re_, im_] := N[(N[Cosh[im], $MachinePrecision] * N[Cos[re], $MachinePrecision]), $MachinePrecision]

Derivation

1. Initial program 100.0%

   \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]
2. Step-by-step derivation

   1. lift-*.f64 N/A

      \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right)} \]

   2. *-commutative N/A

      \[\leadsto \color{blue}{\left(e^{-im} + e^{im}\right) \cdot \left(\frac{1}{2} \cdot \cos re\right)} \]

   3. lift-*.f64 N/A

      \[\leadsto \left(e^{-im} + e^{im}\right) \cdot \color{blue}{\left(\frac{1}{2} \cdot \cos re\right)} \]

   4. associate-*r* N/A

      \[\leadsto \color{blue}{\left(\left(e^{-im} + e^{im}\right) \cdot \frac{1}{2}\right) \cdot \cos re} \]

   5. metadata-eval N/A

      \[\leadsto \left(\left(e^{-im} + e^{im}\right) \cdot \color{blue}{\frac{1}{2}}\right) \cdot \cos re \]

   6. mult-flip-rev N/A

      \[\leadsto \color{blue}{\frac{e^{-im} + e^{im}}{2}} \cdot \cos re \]

   7. lift-+.f64 N/A

      \[\leadsto \frac{\color{blue}{e^{-im} + e^{im}}}{2} \cdot \cos re \]

   8. +-commutative N/A

      \[\leadsto \frac{\color{blue}{e^{im} + e^{-im}}}{2} \cdot \cos re \]

   9. lift-exp.f64 N/A

      \[\leadsto \frac{\color{blue}{e^{im}} + e^{-im}}{2} \cdot \cos re \]

   10. lift-exp.f64 N/A

       \[\leadsto \frac{e^{im} + \color{blue}{e^{-im}}}{2} \cdot \cos re \]

   11. lift-neg.f64 N/A

       \[\leadsto \frac{e^{im} + e^{\color{blue}{\mathsf{neg}\left(im\right)}}}{2} \cdot \cos re \]

   12. cosh-def N/A

       \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

   13. lower-*.f64 N/A

       \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

   14. lower-cosh.f64 100.0%

       \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

3. Applied rewrites 100.0%

   \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]
4. Add Preprocessing

Alternative 2: 99.5% accurate, 0.4× speedup

Math

\[\begin{array}{l} t_0 := 0.5 \cdot \cos re\\ t_1 := t\_0 \cdot \left(e^{-im} + e^{im}\right)\\ \mathbf{if}\;t\_1 \leq -\infty:\\ \;\;\;\;\cosh im \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right)\\ \mathbf{elif}\;t\_1 \leq 0.9999999999999997:\\ \;\;\;\;t\_0 \cdot 2\\ \mathbf{else}:\\ \;\;\;\;\cosh im \cdot 1\\ \end{array} \]

FPCore

(FPCore (re im)
  :precision binary64
  (let* ((t_0 (* 0.5 (cos re))) (t_1 (* t_0 (+ (exp (- im)) (exp im)))))
  (if (<= t_1 (- INFINITY))
    (* (cosh im) (fma (* re re) -0.5 1.0))
    (if (<= t_1 0.9999999999999997) (* t_0 2.0) (* (cosh im) 1.0)))))

C

double code(double re, double im) {
	double t_0 = 0.5 * cos(re);
	double t_1 = t_0 * (exp(-im) + exp(im));
	double tmp;
	if (t_1 <= -((double) INFINITY)) {
		tmp = cosh(im) * fma((re * re), -0.5, 1.0);
	} else if (t_1 <= 0.9999999999999997) {
		tmp = t_0 * 2.0;
	} else {
		tmp = cosh(im) * 1.0;
	}
	return tmp;
}

Julia

function code(re, im)
	t_0 = Float64(0.5 * cos(re))
	t_1 = Float64(t_0 * Float64(exp(Float64(-im)) + exp(im)))
	tmp = 0.0
	if (t_1 <= Float64(-Inf))
		tmp = Float64(cosh(im) * fma(Float64(re * re), -0.5, 1.0));
	elseif (t_1 <= 0.9999999999999997)
		tmp = Float64(t_0 * 2.0);
	else
		tmp = Float64(cosh(im) * 1.0);
	end
	return tmp
end

Wolfram

code[re_, im_] := Block[{t$95$0 = N[(0.5 * N[Cos[re], $MachinePrecision]), $MachinePrecision]}, Block[{t$95$1 = N[(t$95$0 * N[(N[Exp[(-im)], $MachinePrecision] + N[Exp[im], $MachinePrecision]), $MachinePrecision]), $MachinePrecision]}, If[LessEqual[t$95$1, (-Infinity)], N[(N[Cosh[im], $MachinePrecision] * N[(N[(re * re), $MachinePrecision] * -0.5 + 1.0), $MachinePrecision]), $MachinePrecision], If[LessEqual[t$95$1, 0.9999999999999997], N[(t$95$0 * 2.0), $MachinePrecision], N[(N[Cosh[im], $MachinePrecision] * 1.0), $MachinePrecision]]]]]

Derivation

1. Split input into 3 regimes

2. if (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im))) < -inf.0

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]
   2. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right)} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{\left(e^{-im} + e^{im}\right) \cdot \left(\frac{1}{2} \cdot \cos re\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \left(e^{-im} + e^{im}\right) \cdot \color{blue}{\left(\frac{1}{2} \cdot \cos re\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(\left(e^{-im} + e^{im}\right) \cdot \frac{1}{2}\right) \cdot \cos re} \]

      5. metadata-eval N/A

         \[\leadsto \left(\left(e^{-im} + e^{im}\right) \cdot \color{blue}{\frac{1}{2}}\right) \cdot \cos re \]

      6. mult-flip-rev N/A

         \[\leadsto \color{blue}{\frac{e^{-im} + e^{im}}{2}} \cdot \cos re \]

      7. lift-+.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{-im} + e^{im}}}{2} \cdot \cos re \]

      8. +-commutative N/A

         \[\leadsto \frac{\color{blue}{e^{im} + e^{-im}}}{2} \cdot \cos re \]

      9. lift-exp.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{im}} + e^{-im}}{2} \cdot \cos re \]

      10. lift-exp.f64 N/A

          \[\leadsto \frac{e^{im} + \color{blue}{e^{-im}}}{2} \cdot \cos re \]

      11. lift-neg.f64 N/A

          \[\leadsto \frac{e^{im} + e^{\color{blue}{\mathsf{neg}\left(im\right)}}}{2} \cdot \cos re \]

      12. cosh-def N/A

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

      13. lower-*.f64 N/A

          \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

      14. lower-cosh.f64 100.0%

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

   4. Taylor expanded in re around 0

      \[\leadsto \cosh im \cdot \color{blue}{\left(1 + \frac{-1}{2} \cdot {re}^{2}\right)} \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \cosh im \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \cosh im \cdot \left(1 + \frac{-1}{2} \cdot \color{blue}{{re}^{2}}\right) \]

      3. lower-pow.f64 63.1%

         \[\leadsto \cosh im \cdot \left(1 + -0.5 \cdot {re}^{\color{blue}{2}}\right) \]

   6. Applied rewrites 63.1%

      \[\leadsto \cosh im \cdot \color{blue}{\left(1 + -0.5 \cdot {re}^{2}\right)} \]
   7. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \cosh im \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right) \]

      2. +-commutative N/A

         \[\leadsto \cosh im \cdot \left(\frac{-1}{2} \cdot {re}^{2} + \color{blue}{1}\right) \]

      3. lift-*.f64 N/A

         \[\leadsto \cosh im \cdot \left(\frac{-1}{2} \cdot {re}^{2} + 1\right) \]

      4. *-commutative N/A

         \[\leadsto \cosh im \cdot \left({re}^{2} \cdot \frac{-1}{2} + 1\right) \]

      5. lower-fma.f64 63.1%

         \[\leadsto \cosh im \cdot \mathsf{fma}\left({re}^{2}, \color{blue}{-0.5}, 1\right) \]

      6. lift-pow.f64 N/A

         \[\leadsto \cosh im \cdot \mathsf{fma}\left({re}^{2}, \frac{-1}{2}, 1\right) \]

      7. unpow2 N/A

         \[\leadsto \cosh im \cdot \mathsf{fma}\left(re \cdot re, \frac{-1}{2}, 1\right) \]

      8. lower-*.f64 63.1%

         \[\leadsto \cosh im \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right) \]

   8. Applied rewrites 63.1%

      \[\leadsto \cosh im \cdot \mathsf{fma}\left(re \cdot re, \color{blue}{-0.5}, 1\right) \]

   if -inf.0 < (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im))) < 0.99999999999999967

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

   2. Taylor expanded in im around 0

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]
   3. Step-by-step derivation

   4. Applied rewrites 50.8%

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]

   if 0.99999999999999967 < (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im)))

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]
   2. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right)} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{\left(e^{-im} + e^{im}\right) \cdot \left(\frac{1}{2} \cdot \cos re\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \left(e^{-im} + e^{im}\right) \cdot \color{blue}{\left(\frac{1}{2} \cdot \cos re\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(\left(e^{-im} + e^{im}\right) \cdot \frac{1}{2}\right) \cdot \cos re} \]

      5. metadata-eval N/A

         \[\leadsto \left(\left(e^{-im} + e^{im}\right) \cdot \color{blue}{\frac{1}{2}}\right) \cdot \cos re \]

      6. mult-flip-rev N/A

         \[\leadsto \color{blue}{\frac{e^{-im} + e^{im}}{2}} \cdot \cos re \]

      7. lift-+.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{-im} + e^{im}}}{2} \cdot \cos re \]

      8. +-commutative N/A

         \[\leadsto \frac{\color{blue}{e^{im} + e^{-im}}}{2} \cdot \cos re \]

      9. lift-exp.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{im}} + e^{-im}}{2} \cdot \cos re \]

      10. lift-exp.f64 N/A

          \[\leadsto \frac{e^{im} + \color{blue}{e^{-im}}}{2} \cdot \cos re \]

      11. lift-neg.f64 N/A

          \[\leadsto \frac{e^{im} + e^{\color{blue}{\mathsf{neg}\left(im\right)}}}{2} \cdot \cos re \]

      12. cosh-def N/A

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

      13. lower-*.f64 N/A

          \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

      14. lower-cosh.f64 100.0%

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

   4. Taylor expanded in re around 0

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
   5. Step-by-step derivation

   6. Applied rewrites 64.9%

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
3. Recombined 3 regimes into one program.
4. Add Preprocessing

Alternative 3: 77.8% accurate, 0.7× speedup

Math

\[\begin{array}{l} \mathbf{if}\;\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \leq -0.01:\\ \;\;\;\;\cosh im \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right)\\ \mathbf{else}:\\ \;\;\;\;\cosh im \cdot 1\\ \end{array} \]

FPCore

(FPCore (re im)
  :precision binary64
  (if (<= (* (* 0.5 (cos re)) (+ (exp (- im)) (exp im))) -0.01)
  (* (cosh im) (fma (* re re) -0.5 1.0))
  (* (cosh im) 1.0)))

C

double code(double re, double im) {
	double tmp;
	if (((0.5 * cos(re)) * (exp(-im) + exp(im))) <= -0.01) {
		tmp = cosh(im) * fma((re * re), -0.5, 1.0);
	} else {
		tmp = cosh(im) * 1.0;
	}
	return tmp;
}

Julia

function code(re, im)
	tmp = 0.0
	if (Float64(Float64(0.5 * cos(re)) * Float64(exp(Float64(-im)) + exp(im))) <= -0.01)
		tmp = Float64(cosh(im) * fma(Float64(re * re), -0.5, 1.0));
	else
		tmp = Float64(cosh(im) * 1.0);
	end
	return tmp
end

Wolfram

code[re_, im_] := If[LessEqual[N[(N[(0.5 * N[Cos[re], $MachinePrecision]), $MachinePrecision] * N[(N[Exp[(-im)], $MachinePrecision] + N[Exp[im], $MachinePrecision]), $MachinePrecision]), $MachinePrecision], -0.01], N[(N[Cosh[im], $MachinePrecision] * N[(N[(re * re), $MachinePrecision] * -0.5 + 1.0), $MachinePrecision]), $MachinePrecision], N[(N[Cosh[im], $MachinePrecision] * 1.0), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im))) < -0.01

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]
   2. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right)} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{\left(e^{-im} + e^{im}\right) \cdot \left(\frac{1}{2} \cdot \cos re\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \left(e^{-im} + e^{im}\right) \cdot \color{blue}{\left(\frac{1}{2} \cdot \cos re\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(\left(e^{-im} + e^{im}\right) \cdot \frac{1}{2}\right) \cdot \cos re} \]

      5. metadata-eval N/A

         \[\leadsto \left(\left(e^{-im} + e^{im}\right) \cdot \color{blue}{\frac{1}{2}}\right) \cdot \cos re \]

      6. mult-flip-rev N/A

         \[\leadsto \color{blue}{\frac{e^{-im} + e^{im}}{2}} \cdot \cos re \]

      7. lift-+.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{-im} + e^{im}}}{2} \cdot \cos re \]

      8. +-commutative N/A

         \[\leadsto \frac{\color{blue}{e^{im} + e^{-im}}}{2} \cdot \cos re \]

      9. lift-exp.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{im}} + e^{-im}}{2} \cdot \cos re \]

      10. lift-exp.f64 N/A

          \[\leadsto \frac{e^{im} + \color{blue}{e^{-im}}}{2} \cdot \cos re \]

      11. lift-neg.f64 N/A

          \[\leadsto \frac{e^{im} + e^{\color{blue}{\mathsf{neg}\left(im\right)}}}{2} \cdot \cos re \]

      12. cosh-def N/A

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

      13. lower-*.f64 N/A

          \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

      14. lower-cosh.f64 100.0%

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

   4. Taylor expanded in re around 0

      \[\leadsto \cosh im \cdot \color{blue}{\left(1 + \frac{-1}{2} \cdot {re}^{2}\right)} \]
   5. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \cosh im \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right) \]

      2. lower-*.f64 N/A

         \[\leadsto \cosh im \cdot \left(1 + \frac{-1}{2} \cdot \color{blue}{{re}^{2}}\right) \]

      3. lower-pow.f64 63.1%

         \[\leadsto \cosh im \cdot \left(1 + -0.5 \cdot {re}^{\color{blue}{2}}\right) \]

   6. Applied rewrites 63.1%

      \[\leadsto \cosh im \cdot \color{blue}{\left(1 + -0.5 \cdot {re}^{2}\right)} \]
   7. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \cosh im \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right) \]

      2. +-commutative N/A

         \[\leadsto \cosh im \cdot \left(\frac{-1}{2} \cdot {re}^{2} + \color{blue}{1}\right) \]

      3. lift-*.f64 N/A

         \[\leadsto \cosh im \cdot \left(\frac{-1}{2} \cdot {re}^{2} + 1\right) \]

      4. *-commutative N/A

         \[\leadsto \cosh im \cdot \left({re}^{2} \cdot \frac{-1}{2} + 1\right) \]

      5. lower-fma.f64 63.1%

         \[\leadsto \cosh im \cdot \mathsf{fma}\left({re}^{2}, \color{blue}{-0.5}, 1\right) \]

      6. lift-pow.f64 N/A

         \[\leadsto \cosh im \cdot \mathsf{fma}\left({re}^{2}, \frac{-1}{2}, 1\right) \]

      7. unpow2 N/A

         \[\leadsto \cosh im \cdot \mathsf{fma}\left(re \cdot re, \frac{-1}{2}, 1\right) \]

      8. lower-*.f64 63.1%

         \[\leadsto \cosh im \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right) \]

   8. Applied rewrites 63.1%

      \[\leadsto \cosh im \cdot \mathsf{fma}\left(re \cdot re, \color{blue}{-0.5}, 1\right) \]

   if -0.01 < (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im)))

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]
   2. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right)} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{\left(e^{-im} + e^{im}\right) \cdot \left(\frac{1}{2} \cdot \cos re\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \left(e^{-im} + e^{im}\right) \cdot \color{blue}{\left(\frac{1}{2} \cdot \cos re\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(\left(e^{-im} + e^{im}\right) \cdot \frac{1}{2}\right) \cdot \cos re} \]

      5. metadata-eval N/A

         \[\leadsto \left(\left(e^{-im} + e^{im}\right) \cdot \color{blue}{\frac{1}{2}}\right) \cdot \cos re \]

      6. mult-flip-rev N/A

         \[\leadsto \color{blue}{\frac{e^{-im} + e^{im}}{2}} \cdot \cos re \]

      7. lift-+.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{-im} + e^{im}}}{2} \cdot \cos re \]

      8. +-commutative N/A

         \[\leadsto \frac{\color{blue}{e^{im} + e^{-im}}}{2} \cdot \cos re \]

      9. lift-exp.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{im}} + e^{-im}}{2} \cdot \cos re \]

      10. lift-exp.f64 N/A

          \[\leadsto \frac{e^{im} + \color{blue}{e^{-im}}}{2} \cdot \cos re \]

      11. lift-neg.f64 N/A

          \[\leadsto \frac{e^{im} + e^{\color{blue}{\mathsf{neg}\left(im\right)}}}{2} \cdot \cos re \]

      12. cosh-def N/A

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

      13. lower-*.f64 N/A

          \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

      14. lower-cosh.f64 100.0%

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

   4. Taylor expanded in re around 0

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
   5. Step-by-step derivation

   6. Applied rewrites 64.9%

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 4: 74.8% accurate, 0.7× speedup

Math

\[\begin{array}{l} \mathbf{if}\;\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \leq -0.01:\\ \;\;\;\;\left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(\sqrt{\left(\left(re \cdot re\right) \cdot re\right) \cdot re}, -0.5, 1\right)\\ \mathbf{else}:\\ \;\;\;\;\cosh im \cdot 1\\ \end{array} \]

FPCore

(FPCore (re im)
  :precision binary64
  (if (<= (* (* 0.5 (cos re)) (+ (exp (- im)) (exp im))) -0.01)
  (* (* 2.0 0.5) (fma (sqrt (* (* (* re re) re) re)) -0.5 1.0))
  (* (cosh im) 1.0)))

C

double code(double re, double im) {
	double tmp;
	if (((0.5 * cos(re)) * (exp(-im) + exp(im))) <= -0.01) {
		tmp = (2.0 * 0.5) * fma(sqrt((((re * re) * re) * re)), -0.5, 1.0);
	} else {
		tmp = cosh(im) * 1.0;
	}
	return tmp;
}

Julia

function code(re, im)
	tmp = 0.0
	if (Float64(Float64(0.5 * cos(re)) * Float64(exp(Float64(-im)) + exp(im))) <= -0.01)
		tmp = Float64(Float64(2.0 * 0.5) * fma(sqrt(Float64(Float64(Float64(re * re) * re) * re)), -0.5, 1.0));
	else
		tmp = Float64(cosh(im) * 1.0);
	end
	return tmp
end

Wolfram

code[re_, im_] := If[LessEqual[N[(N[(0.5 * N[Cos[re], $MachinePrecision]), $MachinePrecision] * N[(N[Exp[(-im)], $MachinePrecision] + N[Exp[im], $MachinePrecision]), $MachinePrecision]), $MachinePrecision], -0.01], N[(N[(2.0 * 0.5), $MachinePrecision] * N[(N[Sqrt[N[(N[(N[(re * re), $MachinePrecision] * re), $MachinePrecision] * re), $MachinePrecision]], $MachinePrecision] * -0.5 + 1.0), $MachinePrecision]), $MachinePrecision], N[(N[Cosh[im], $MachinePrecision] * 1.0), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im))) < -0.01

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

   2. Taylor expanded in im around 0

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]
   3. Step-by-step derivation

   4. Applied rewrites 50.8%

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]

   5. Taylor expanded in re around 0

      \[\leadsto \left(0.5 \cdot \color{blue}{\left(1 + \frac{-1}{2} \cdot {re}^{2}\right)}\right) \cdot 2 \]
   6. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \left(\frac{1}{2} \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right)\right) \cdot 2 \]

      2. lower-*.f64 N/A

         \[\leadsto \left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot \color{blue}{{re}^{2}}\right)\right) \cdot 2 \]

      3. lower-pow.f64 32.9%

         \[\leadsto \left(0.5 \cdot \left(1 + -0.5 \cdot {re}^{\color{blue}{2}}\right)\right) \cdot 2 \]

   7. Applied rewrites 32.9%

      \[\leadsto \left(0.5 \cdot \color{blue}{\left(1 + -0.5 \cdot {re}^{2}\right)}\right) \cdot 2 \]
   8. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right) \cdot 2} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{2 \cdot \left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto 2 \cdot \color{blue}{\left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)} \]

      5. lower-*.f64 N/A

         \[\leadsto \color{blue}{\left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)} \]

      6. lower-*.f64 32.9%

         \[\leadsto \color{blue}{\left(2 \cdot 0.5\right)} \cdot \left(1 + -0.5 \cdot {re}^{2}\right) \]

      7. lift-+.f64 N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right) \]

      8. +-commutative N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(\frac{-1}{2} \cdot {re}^{2} + \color{blue}{1}\right) \]

      9. lift-*.f64 N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(\frac{-1}{2} \cdot {re}^{2} + 1\right) \]

      10. *-commutative N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left({re}^{2} \cdot \frac{-1}{2} + 1\right) \]

      11. lower-fma.f64 32.9%

          \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left({re}^{2}, \color{blue}{-0.5}, 1\right) \]

      12. lift-pow.f64 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left({re}^{2}, \frac{-1}{2}, 1\right) \]

      13. unpow2 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(re \cdot re, \frac{-1}{2}, 1\right) \]

      14. lower-*.f64 32.9%

          \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right) \]

   9. Applied rewrites 32.9%

      \[\leadsto \color{blue}{\left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right)} \]
   10. Step-by-step derivation

       1. rem-square-sqrt N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{re \cdot re} \cdot \sqrt{re \cdot re}, \frac{-1}{2}, 1\right) \]

       2. sqrt-unprod N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{\left(re \cdot re\right) \cdot \left(re \cdot re\right)}, \frac{-1}{2}, 1\right) \]

       3. lift-*.f64 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{\left(re \cdot re\right) \cdot \left(re \cdot re\right)}, \frac{-1}{2}, 1\right) \]

       4. lift-sqrt.f64 35.6%

          \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(\sqrt{\left(re \cdot re\right) \cdot \left(re \cdot re\right)}, -0.5, 1\right) \]

       5. lift-*.f64 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{\left(re \cdot re\right) \cdot \left(re \cdot re\right)}, \frac{-1}{2}, 1\right) \]

       6. lift-*.f64 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{\left(re \cdot re\right) \cdot \left(re \cdot re\right)}, \frac{-1}{2}, 1\right) \]

       7. associate-*l* N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{re \cdot \left(re \cdot \left(re \cdot re\right)\right)}, \frac{-1}{2}, 1\right) \]

       8. *-commutative N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{\left(re \cdot \left(re \cdot re\right)\right) \cdot re}, \frac{-1}{2}, 1\right) \]

       9. lower-*.f64 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{\left(re \cdot \left(re \cdot re\right)\right) \cdot re}, \frac{-1}{2}, 1\right) \]

       10. *-commutative N/A

           \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(\sqrt{\left(\left(re \cdot re\right) \cdot re\right) \cdot re}, \frac{-1}{2}, 1\right) \]

       11. lower-*.f64 35.6%

           \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(\sqrt{\left(\left(re \cdot re\right) \cdot re\right) \cdot re}, -0.5, 1\right) \]

   11. Applied rewrites 35.6%

       \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(\sqrt{\left(\left(re \cdot re\right) \cdot re\right) \cdot re}, -0.5, 1\right) \]

   if -0.01 < (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im)))

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]
   2. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right)} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{\left(e^{-im} + e^{im}\right) \cdot \left(\frac{1}{2} \cdot \cos re\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \left(e^{-im} + e^{im}\right) \cdot \color{blue}{\left(\frac{1}{2} \cdot \cos re\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(\left(e^{-im} + e^{im}\right) \cdot \frac{1}{2}\right) \cdot \cos re} \]

      5. metadata-eval N/A

         \[\leadsto \left(\left(e^{-im} + e^{im}\right) \cdot \color{blue}{\frac{1}{2}}\right) \cdot \cos re \]

      6. mult-flip-rev N/A

         \[\leadsto \color{blue}{\frac{e^{-im} + e^{im}}{2}} \cdot \cos re \]

      7. lift-+.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{-im} + e^{im}}}{2} \cdot \cos re \]

      8. +-commutative N/A

         \[\leadsto \frac{\color{blue}{e^{im} + e^{-im}}}{2} \cdot \cos re \]

      9. lift-exp.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{im}} + e^{-im}}{2} \cdot \cos re \]

      10. lift-exp.f64 N/A

          \[\leadsto \frac{e^{im} + \color{blue}{e^{-im}}}{2} \cdot \cos re \]

      11. lift-neg.f64 N/A

          \[\leadsto \frac{e^{im} + e^{\color{blue}{\mathsf{neg}\left(im\right)}}}{2} \cdot \cos re \]

      12. cosh-def N/A

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

      13. lower-*.f64 N/A

          \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

      14. lower-cosh.f64 100.0%

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

   4. Taylor expanded in re around 0

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
   5. Step-by-step derivation

   6. Applied rewrites 64.9%

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 5: 72.1% accurate, 0.8× speedup

Math

\[\begin{array}{l} \mathbf{if}\;\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \leq 0.92:\\ \;\;\;\;0.5 \cdot \left(\left(2 + 2 \cdot im\right) \cdot \left(1 + -1 \cdot im\right)\right)\\ \mathbf{else}:\\ \;\;\;\;\cosh im \cdot 1\\ \end{array} \]

FPCore

(FPCore (re im)
  :precision binary64
  (if (<= (* (* 0.5 (cos re)) (+ (exp (- im)) (exp im))) 0.92)
  (* 0.5 (* (+ 2.0 (* 2.0 im)) (+ 1.0 (* -1.0 im))))
  (* (cosh im) 1.0)))

C

double code(double re, double im) {
	double tmp;
	if (((0.5 * cos(re)) * (exp(-im) + exp(im))) <= 0.92) {
		tmp = 0.5 * ((2.0 + (2.0 * im)) * (1.0 + (-1.0 * im)));
	} else {
		tmp = cosh(im) * 1.0;
	}
	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(re, im)
use fmin_fmax_functions
    real(8), intent (in) :: re
    real(8), intent (in) :: im
    real(8) :: tmp
    if (((0.5d0 * cos(re)) * (exp(-im) + exp(im))) <= 0.92d0) then
        tmp = 0.5d0 * ((2.0d0 + (2.0d0 * im)) * (1.0d0 + ((-1.0d0) * im)))
    else
        tmp = cosh(im) * 1.0d0
    end if
    code = tmp
end function

Java

public static double code(double re, double im) {
	double tmp;
	if (((0.5 * Math.cos(re)) * (Math.exp(-im) + Math.exp(im))) <= 0.92) {
		tmp = 0.5 * ((2.0 + (2.0 * im)) * (1.0 + (-1.0 * im)));
	} else {
		tmp = Math.cosh(im) * 1.0;
	}
	return tmp;
}

Python

def code(re, im):
	tmp = 0
	if ((0.5 * math.cos(re)) * (math.exp(-im) + math.exp(im))) <= 0.92:
		tmp = 0.5 * ((2.0 + (2.0 * im)) * (1.0 + (-1.0 * im)))
	else:
		tmp = math.cosh(im) * 1.0
	return tmp

Julia

function code(re, im)
	tmp = 0.0
	if (Float64(Float64(0.5 * cos(re)) * Float64(exp(Float64(-im)) + exp(im))) <= 0.92)
		tmp = Float64(0.5 * Float64(Float64(2.0 + Float64(2.0 * im)) * Float64(1.0 + Float64(-1.0 * im))));
	else
		tmp = Float64(cosh(im) * 1.0);
	end
	return tmp
end

MATLAB

function tmp_2 = code(re, im)
	tmp = 0.0;
	if (((0.5 * cos(re)) * (exp(-im) + exp(im))) <= 0.92)
		tmp = 0.5 * ((2.0 + (2.0 * im)) * (1.0 + (-1.0 * im)));
	else
		tmp = cosh(im) * 1.0;
	end
	tmp_2 = tmp;
end

Wolfram

code[re_, im_] := If[LessEqual[N[(N[(0.5 * N[Cos[re], $MachinePrecision]), $MachinePrecision] * N[(N[Exp[(-im)], $MachinePrecision] + N[Exp[im], $MachinePrecision]), $MachinePrecision]), $MachinePrecision], 0.92], N[(0.5 * N[(N[(2.0 + N[(2.0 * im), $MachinePrecision]), $MachinePrecision] * N[(1.0 + N[(-1.0 * im), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision], N[(N[Cosh[im], $MachinePrecision] * 1.0), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im))) < 0.92000000000000004

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

   2. Taylor expanded in re around 0

      \[\leadsto \color{blue}{\frac{1}{2}} \cdot \left(e^{-im} + e^{im}\right) \]
   3. Step-by-step derivation

   4. Applied rewrites 64.9%

      \[\leadsto \color{blue}{0.5} \cdot \left(e^{-im} + e^{im}\right) \]
   5. Step-by-step derivation

      1. lift-+.f64 N/A

         \[\leadsto \frac{1}{2} \cdot \color{blue}{\left(e^{-im} + e^{im}\right)} \]

      2. sum-to-mult-rev N/A

         \[\leadsto \frac{1}{2} \cdot \color{blue}{\left(\left(1 + \frac{e^{im}}{e^{-im}}\right) \cdot e^{-im}\right)} \]

      3. lift-exp.f64 N/A

         \[\leadsto \frac{1}{2} \cdot \left(\left(1 + \frac{\color{blue}{e^{im}}}{e^{-im}}\right) \cdot e^{-im}\right) \]

      4. lift-exp.f64 N/A

         \[\leadsto \frac{1}{2} \cdot \left(\left(1 + \frac{e^{im}}{\color{blue}{e^{-im}}}\right) \cdot e^{-im}\right) \]

      5. exp-diff N/A

         \[\leadsto \frac{1}{2} \cdot \left(\left(1 + \color{blue}{e^{im - \left(-im\right)}}\right) \cdot e^{-im}\right) \]

      6. lift-neg.f64 N/A

         \[\leadsto \frac{1}{2} \cdot \left(\left(1 + e^{im - \color{blue}{\left(\mathsf{neg}\left(im\right)\right)}}\right) \cdot e^{-im}\right) \]

      7. add-flip N/A

         \[\leadsto \frac{1}{2} \cdot \left(\left(1 + e^{\color{blue}{im + im}}\right) \cdot e^{-im}\right) \]

      8. lift-+.f64 N/A

         \[\leadsto \frac{1}{2} \cdot \left(\left(1 + e^{\color{blue}{im + im}}\right) \cdot e^{-im}\right) \]

      9. lift-exp.f64 N/A

         \[\leadsto \frac{1}{2} \cdot \left(\left(1 + \color{blue}{e^{im + im}}\right) \cdot e^{-im}\right) \]

      10. +-commutative N/A

          \[\leadsto \frac{1}{2} \cdot \left(\color{blue}{\left(e^{im + im} + 1\right)} \cdot e^{-im}\right) \]

      11. metadata-eval N/A

          \[\leadsto \frac{1}{2} \cdot \left(\left(e^{im + im} + \color{blue}{\left(\mathsf{neg}\left(-1\right)\right)}\right) \cdot e^{-im}\right) \]

      12. sub-flip N/A

          \[\leadsto \frac{1}{2} \cdot \left(\color{blue}{\left(e^{im + im} - -1\right)} \cdot e^{-im}\right) \]

      13. lift--.f64 N/A

          \[\leadsto \frac{1}{2} \cdot \left(\color{blue}{\left(e^{im + im} - -1\right)} \cdot e^{-im}\right) \]

      14. lower-*.f64 46.3%

          \[\leadsto 0.5 \cdot \color{blue}{\left(\left(e^{im + im} - -1\right) \cdot e^{-im}\right)} \]

   6. Applied rewrites 46.3%

      \[\leadsto 0.5 \cdot \color{blue}{\left(\left(e^{im + im} - -1\right) \cdot e^{-im}\right)} \]

   7. Taylor expanded in im around 0

      \[\leadsto 0.5 \cdot \left(\color{blue}{\left(2 + 2 \cdot im\right)} \cdot e^{-im}\right) \]
   8. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \frac{1}{2} \cdot \left(\left(2 + \color{blue}{2 \cdot im}\right) \cdot e^{-im}\right) \]

      2. lower-*.f64 33.5%

         \[\leadsto 0.5 \cdot \left(\left(2 + 2 \cdot \color{blue}{im}\right) \cdot e^{-im}\right) \]

   9. Applied rewrites 33.5%

      \[\leadsto 0.5 \cdot \left(\color{blue}{\left(2 + 2 \cdot im\right)} \cdot e^{-im}\right) \]

   10. Taylor expanded in im around 0

       \[\leadsto 0.5 \cdot \left(\left(2 + 2 \cdot im\right) \cdot \color{blue}{\left(1 + -1 \cdot im\right)}\right) \]
   11. Step-by-step derivation

       1. lower-+.f64 N/A

          \[\leadsto \frac{1}{2} \cdot \left(\left(2 + 2 \cdot im\right) \cdot \left(1 + \color{blue}{-1 \cdot im}\right)\right) \]

       2. lower-*.f64 34.1%

          \[\leadsto 0.5 \cdot \left(\left(2 + 2 \cdot im\right) \cdot \left(1 + -1 \cdot \color{blue}{im}\right)\right) \]

   12. Applied rewrites 34.1%

       \[\leadsto 0.5 \cdot \left(\left(2 + 2 \cdot im\right) \cdot \color{blue}{\left(1 + -1 \cdot im\right)}\right) \]

   if 0.92000000000000004 < (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im)))

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]
   2. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right)} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{\left(e^{-im} + e^{im}\right) \cdot \left(\frac{1}{2} \cdot \cos re\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \left(e^{-im} + e^{im}\right) \cdot \color{blue}{\left(\frac{1}{2} \cdot \cos re\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(\left(e^{-im} + e^{im}\right) \cdot \frac{1}{2}\right) \cdot \cos re} \]

      5. metadata-eval N/A

         \[\leadsto \left(\left(e^{-im} + e^{im}\right) \cdot \color{blue}{\frac{1}{2}}\right) \cdot \cos re \]

      6. mult-flip-rev N/A

         \[\leadsto \color{blue}{\frac{e^{-im} + e^{im}}{2}} \cdot \cos re \]

      7. lift-+.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{-im} + e^{im}}}{2} \cdot \cos re \]

      8. +-commutative N/A

         \[\leadsto \frac{\color{blue}{e^{im} + e^{-im}}}{2} \cdot \cos re \]

      9. lift-exp.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{im}} + e^{-im}}{2} \cdot \cos re \]

      10. lift-exp.f64 N/A

          \[\leadsto \frac{e^{im} + \color{blue}{e^{-im}}}{2} \cdot \cos re \]

      11. lift-neg.f64 N/A

          \[\leadsto \frac{e^{im} + e^{\color{blue}{\mathsf{neg}\left(im\right)}}}{2} \cdot \cos re \]

      12. cosh-def N/A

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

      13. lower-*.f64 N/A

          \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

      14. lower-cosh.f64 100.0%

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

   4. Taylor expanded in re around 0

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
   5. Step-by-step derivation

   6. Applied rewrites 64.9%

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 6: 71.7% accurate, 0.8× speedup

Math

\[\begin{array}{l} \mathbf{if}\;\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \leq -0.01:\\ \;\;\;\;\left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right)\\ \mathbf{else}:\\ \;\;\;\;\cosh im \cdot 1\\ \end{array} \]

FPCore

(FPCore (re im)
  :precision binary64
  (if (<= (* (* 0.5 (cos re)) (+ (exp (- im)) (exp im))) -0.01)
  (* (* 2.0 0.5) (fma (* re re) -0.5 1.0))
  (* (cosh im) 1.0)))

C

double code(double re, double im) {
	double tmp;
	if (((0.5 * cos(re)) * (exp(-im) + exp(im))) <= -0.01) {
		tmp = (2.0 * 0.5) * fma((re * re), -0.5, 1.0);
	} else {
		tmp = cosh(im) * 1.0;
	}
	return tmp;
}

Julia

function code(re, im)
	tmp = 0.0
	if (Float64(Float64(0.5 * cos(re)) * Float64(exp(Float64(-im)) + exp(im))) <= -0.01)
		tmp = Float64(Float64(2.0 * 0.5) * fma(Float64(re * re), -0.5, 1.0));
	else
		tmp = Float64(cosh(im) * 1.0);
	end
	return tmp
end

Wolfram

code[re_, im_] := If[LessEqual[N[(N[(0.5 * N[Cos[re], $MachinePrecision]), $MachinePrecision] * N[(N[Exp[(-im)], $MachinePrecision] + N[Exp[im], $MachinePrecision]), $MachinePrecision]), $MachinePrecision], -0.01], N[(N[(2.0 * 0.5), $MachinePrecision] * N[(N[(re * re), $MachinePrecision] * -0.5 + 1.0), $MachinePrecision]), $MachinePrecision], N[(N[Cosh[im], $MachinePrecision] * 1.0), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im))) < -0.01

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

   2. Taylor expanded in im around 0

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]
   3. Step-by-step derivation

   4. Applied rewrites 50.8%

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]

   5. Taylor expanded in re around 0

      \[\leadsto \left(0.5 \cdot \color{blue}{\left(1 + \frac{-1}{2} \cdot {re}^{2}\right)}\right) \cdot 2 \]
   6. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \left(\frac{1}{2} \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right)\right) \cdot 2 \]

      2. lower-*.f64 N/A

         \[\leadsto \left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot \color{blue}{{re}^{2}}\right)\right) \cdot 2 \]

      3. lower-pow.f64 32.9%

         \[\leadsto \left(0.5 \cdot \left(1 + -0.5 \cdot {re}^{\color{blue}{2}}\right)\right) \cdot 2 \]

   7. Applied rewrites 32.9%

      \[\leadsto \left(0.5 \cdot \color{blue}{\left(1 + -0.5 \cdot {re}^{2}\right)}\right) \cdot 2 \]
   8. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right) \cdot 2} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{2 \cdot \left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto 2 \cdot \color{blue}{\left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)} \]

      5. lower-*.f64 N/A

         \[\leadsto \color{blue}{\left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)} \]

      6. lower-*.f64 32.9%

         \[\leadsto \color{blue}{\left(2 \cdot 0.5\right)} \cdot \left(1 + -0.5 \cdot {re}^{2}\right) \]

      7. lift-+.f64 N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right) \]

      8. +-commutative N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(\frac{-1}{2} \cdot {re}^{2} + \color{blue}{1}\right) \]

      9. lift-*.f64 N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(\frac{-1}{2} \cdot {re}^{2} + 1\right) \]

      10. *-commutative N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left({re}^{2} \cdot \frac{-1}{2} + 1\right) \]

      11. lower-fma.f64 32.9%

          \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left({re}^{2}, \color{blue}{-0.5}, 1\right) \]

      12. lift-pow.f64 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left({re}^{2}, \frac{-1}{2}, 1\right) \]

      13. unpow2 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(re \cdot re, \frac{-1}{2}, 1\right) \]

      14. lower-*.f64 32.9%

          \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right) \]

   9. Applied rewrites 32.9%

      \[\leadsto \color{blue}{\left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right)} \]

   if -0.01 < (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im)))

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]
   2. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right)} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{\left(e^{-im} + e^{im}\right) \cdot \left(\frac{1}{2} \cdot \cos re\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto \left(e^{-im} + e^{im}\right) \cdot \color{blue}{\left(\frac{1}{2} \cdot \cos re\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(\left(e^{-im} + e^{im}\right) \cdot \frac{1}{2}\right) \cdot \cos re} \]

      5. metadata-eval N/A

         \[\leadsto \left(\left(e^{-im} + e^{im}\right) \cdot \color{blue}{\frac{1}{2}}\right) \cdot \cos re \]

      6. mult-flip-rev N/A

         \[\leadsto \color{blue}{\frac{e^{-im} + e^{im}}{2}} \cdot \cos re \]

      7. lift-+.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{-im} + e^{im}}}{2} \cdot \cos re \]

      8. +-commutative N/A

         \[\leadsto \frac{\color{blue}{e^{im} + e^{-im}}}{2} \cdot \cos re \]

      9. lift-exp.f64 N/A

         \[\leadsto \frac{\color{blue}{e^{im}} + e^{-im}}{2} \cdot \cos re \]

      10. lift-exp.f64 N/A

          \[\leadsto \frac{e^{im} + \color{blue}{e^{-im}}}{2} \cdot \cos re \]

      11. lift-neg.f64 N/A

          \[\leadsto \frac{e^{im} + e^{\color{blue}{\mathsf{neg}\left(im\right)}}}{2} \cdot \cos re \]

      12. cosh-def N/A

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

      13. lower-*.f64 N/A

          \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

      14. lower-cosh.f64 100.0%

          \[\leadsto \color{blue}{\cosh im} \cdot \cos re \]

   3. Applied rewrites 100.0%

      \[\leadsto \color{blue}{\cosh im \cdot \cos re} \]

   4. Taylor expanded in re around 0

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
   5. Step-by-step derivation

   6. Applied rewrites 64.9%

      \[\leadsto \cosh im \cdot \color{blue}{1} \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 7: 35.6% accurate, 0.8× speedup

Math

\[\begin{array}{l} \mathbf{if}\;\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \leq -0.01:\\ \;\;\;\;\left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right)\\ \mathbf{else}:\\ \;\;\;\;\left(0.5 \cdot 1\right) \cdot 2\\ \end{array} \]

FPCore

(FPCore (re im)
  :precision binary64
  (if (<= (* (* 0.5 (cos re)) (+ (exp (- im)) (exp im))) -0.01)
  (* (* 2.0 0.5) (fma (* re re) -0.5 1.0))
  (* (* 0.5 1.0) 2.0)))

C

double code(double re, double im) {
	double tmp;
	if (((0.5 * cos(re)) * (exp(-im) + exp(im))) <= -0.01) {
		tmp = (2.0 * 0.5) * fma((re * re), -0.5, 1.0);
	} else {
		tmp = (0.5 * 1.0) * 2.0;
	}
	return tmp;
}

Julia

function code(re, im)
	tmp = 0.0
	if (Float64(Float64(0.5 * cos(re)) * Float64(exp(Float64(-im)) + exp(im))) <= -0.01)
		tmp = Float64(Float64(2.0 * 0.5) * fma(Float64(re * re), -0.5, 1.0));
	else
		tmp = Float64(Float64(0.5 * 1.0) * 2.0);
	end
	return tmp
end

Wolfram

code[re_, im_] := If[LessEqual[N[(N[(0.5 * N[Cos[re], $MachinePrecision]), $MachinePrecision] * N[(N[Exp[(-im)], $MachinePrecision] + N[Exp[im], $MachinePrecision]), $MachinePrecision]), $MachinePrecision], -0.01], N[(N[(2.0 * 0.5), $MachinePrecision] * N[(N[(re * re), $MachinePrecision] * -0.5 + 1.0), $MachinePrecision]), $MachinePrecision], N[(N[(0.5 * 1.0), $MachinePrecision] * 2.0), $MachinePrecision]]

Derivation

1. Split input into 2 regimes

2. if (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im))) < -0.01

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

   2. Taylor expanded in im around 0

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]
   3. Step-by-step derivation

   4. Applied rewrites 50.8%

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]

   5. Taylor expanded in re around 0

      \[\leadsto \left(0.5 \cdot \color{blue}{\left(1 + \frac{-1}{2} \cdot {re}^{2}\right)}\right) \cdot 2 \]
   6. Step-by-step derivation

      1. lower-+.f64 N/A

         \[\leadsto \left(\frac{1}{2} \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right)\right) \cdot 2 \]

      2. lower-*.f64 N/A

         \[\leadsto \left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot \color{blue}{{re}^{2}}\right)\right) \cdot 2 \]

      3. lower-pow.f64 32.9%

         \[\leadsto \left(0.5 \cdot \left(1 + -0.5 \cdot {re}^{\color{blue}{2}}\right)\right) \cdot 2 \]

   7. Applied rewrites 32.9%

      \[\leadsto \left(0.5 \cdot \color{blue}{\left(1 + -0.5 \cdot {re}^{2}\right)}\right) \cdot 2 \]
   8. Step-by-step derivation

      1. lift-*.f64 N/A

         \[\leadsto \color{blue}{\left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right) \cdot 2} \]

      2. *-commutative N/A

         \[\leadsto \color{blue}{2 \cdot \left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right)} \]

      3. lift-*.f64 N/A

         \[\leadsto 2 \cdot \color{blue}{\left(\frac{1}{2} \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)\right)} \]

      4. associate-*r* N/A

         \[\leadsto \color{blue}{\left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)} \]

      5. lower-*.f64 N/A

         \[\leadsto \color{blue}{\left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \frac{-1}{2} \cdot {re}^{2}\right)} \]

      6. lower-*.f64 32.9%

         \[\leadsto \color{blue}{\left(2 \cdot 0.5\right)} \cdot \left(1 + -0.5 \cdot {re}^{2}\right) \]

      7. lift-+.f64 N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(1 + \color{blue}{\frac{-1}{2} \cdot {re}^{2}}\right) \]

      8. +-commutative N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(\frac{-1}{2} \cdot {re}^{2} + \color{blue}{1}\right) \]

      9. lift-*.f64 N/A

         \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left(\frac{-1}{2} \cdot {re}^{2} + 1\right) \]

      10. *-commutative N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \left({re}^{2} \cdot \frac{-1}{2} + 1\right) \]

      11. lower-fma.f64 32.9%

          \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left({re}^{2}, \color{blue}{-0.5}, 1\right) \]

      12. lift-pow.f64 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left({re}^{2}, \frac{-1}{2}, 1\right) \]

      13. unpow2 N/A

          \[\leadsto \left(2 \cdot \frac{1}{2}\right) \cdot \mathsf{fma}\left(re \cdot re, \frac{-1}{2}, 1\right) \]

      14. lower-*.f64 32.9%

          \[\leadsto \left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right) \]

   9. Applied rewrites 32.9%

      \[\leadsto \color{blue}{\left(2 \cdot 0.5\right) \cdot \mathsf{fma}\left(re \cdot re, -0.5, 1\right)} \]

   if -0.01 < (*.f64 (*.f64 #s(literal 1/2 binary64) (cos.f64 re)) (+.f64 (exp.f64 (neg.f64 im)) (exp.f64 im)))

   1. Initial program 100.0%

      \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

   2. Taylor expanded in im around 0

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]
   3. Step-by-step derivation

   4. Applied rewrites 50.8%

      \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]

   5. Taylor expanded in re around 0

      \[\leadsto \left(0.5 \cdot \color{blue}{1}\right) \cdot 2 \]
   6. Step-by-step derivation

   7. Applied rewrites 28.5%

      \[\leadsto \left(0.5 \cdot \color{blue}{1}\right) \cdot 2 \]
3. Recombined 2 regimes into one program.
4. Add Preprocessing

Alternative 8: 28.5% accurate, 9.5× speedup

Math

\[\left(0.5 \cdot 1\right) \cdot 2 \]

FPCore

(FPCore (re im)
  :precision binary64
  (* (* 0.5 1.0) 2.0))

C

double code(double re, double im) {
	return (0.5 * 1.0) * 2.0;
}

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(re, im)
use fmin_fmax_functions
    real(8), intent (in) :: re
    real(8), intent (in) :: im
    code = (0.5d0 * 1.0d0) * 2.0d0
end function

Java

public static double code(double re, double im) {
	return (0.5 * 1.0) * 2.0;
}

Python

def code(re, im):
	return (0.5 * 1.0) * 2.0

Julia

function code(re, im)
	return Float64(Float64(0.5 * 1.0) * 2.0)
end

MATLAB

function tmp = code(re, im)
	tmp = (0.5 * 1.0) * 2.0;
end

Wolfram

code[re_, im_] := N[(N[(0.5 * 1.0), $MachinePrecision] * 2.0), $MachinePrecision]

Derivation

1. Initial program 100.0%

   \[\left(0.5 \cdot \cos re\right) \cdot \left(e^{-im} + e^{im}\right) \]

2. Taylor expanded in im around 0

   \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]
3. Step-by-step derivation

4. Applied rewrites 50.8%

   \[\leadsto \left(0.5 \cdot \cos re\right) \cdot \color{blue}{2} \]

5. Taylor expanded in re around 0

   \[\leadsto \left(0.5 \cdot \color{blue}{1}\right) \cdot 2 \]
6. Step-by-step derivation

7. Applied rewrites 28.5%

   \[\leadsto \left(0.5 \cdot \color{blue}{1}\right) \cdot 2 \]
8. Add Preprocessing

Reproduce

herbie shell --seed 2025212 
(FPCore (re im)
  :name "math.cos on complex, real part"
  :precision binary64
  (* (* 0.5 (cos re)) (+ (exp (- im)) (exp im))))