2-ancestry mixing, negative discriminant

Percentage Accurate: 98.5% → 100.0%
Time: 7.1s
Alternatives: 3
Speedup: 1.1×

Specification

?
\[\begin{array}{l} \\ 2 \cdot \cos \left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \end{array} \]
(FPCore (g h)
 :precision binary64
 (* 2.0 (cos (+ (/ (* 2.0 (PI)) 3.0) (/ (acos (/ (- g) h)) 3.0)))))
\begin{array}{l}

\\
2 \cdot \cos \left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right)
\end{array}

Sampling outcomes in binary64 precision:

Local Percentage Accuracy vs ?

The average percentage accuracy by input value. Horizontal axis shows value of an input variable; the variable is choosen in the title. Vertical axis is accuracy; higher is better. Red represent the original program, while blue represents Herbie's suggestion. These can be toggled with buttons below the plot. The line is an average while dots represent individual samples.

Accuracy vs Speed?

Herbie found 3 alternatives:

AlternativeAccuracySpeedup
The accuracy (vertical axis) and speed (horizontal axis) of each alternatives. Up and to the right is better. The red square shows the initial program, and each blue circle shows an alternative.The line shows the best available speed-accuracy tradeoffs.

Initial Program: 98.5% accurate, 1.0× speedup?

\[\begin{array}{l} \\ 2 \cdot \cos \left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \end{array} \]
(FPCore (g h)
 :precision binary64
 (* 2.0 (cos (+ (/ (* 2.0 (PI)) 3.0) (/ (acos (/ (- g) h)) 3.0)))))
\begin{array}{l}

\\
2 \cdot \cos \left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right)
\end{array}

Alternative 1: 100.0% accurate, 0.4× speedup?

\[\begin{array}{l} \\ \begin{array}{l} t_0 := \cos^{-1} \left(\frac{-g}{h}\right)\\ -\mathsf{fma}\left(\sin \left(t\_0 \cdot 0.3333333333333333\right) \cdot \sin \left(\mathsf{PI}\left(\right) \cdot 0.6666666666666666\right), 2, \cos \left(t\_0 \cdot -0.3333333333333333\right)\right) \end{array} \end{array} \]
(FPCore (g h)
 :precision binary64
 (let* ((t_0 (acos (/ (- g) h))))
   (-
    (fma
     (* (sin (* t_0 0.3333333333333333)) (sin (* (PI) 0.6666666666666666)))
     2.0
     (cos (* t_0 -0.3333333333333333))))))
\begin{array}{l}

\\
\begin{array}{l}
t_0 := \cos^{-1} \left(\frac{-g}{h}\right)\\
-\mathsf{fma}\left(\sin \left(t\_0 \cdot 0.3333333333333333\right) \cdot \sin \left(\mathsf{PI}\left(\right) \cdot 0.6666666666666666\right), 2, \cos \left(t\_0 \cdot -0.3333333333333333\right)\right)
\end{array}
\end{array}
Derivation
  1. Initial program 98.5%

    \[2 \cdot \cos \left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
  2. Add Preprocessing
  3. Applied rewrites100.0%

    \[\leadsto 2 \cdot \color{blue}{\mathsf{fma}\left(\cos \left(-0.3333333333333333 \cdot \cos^{-1} \left(\frac{-g}{h}\right)\right), -0.5, \left(-\sin \left(0.6666666666666666 \cdot \mathsf{PI}\left(\right)\right)\right) \cdot \sin \left(0.3333333333333333 \cdot \cos^{-1} \left(\frac{-g}{h}\right)\right)\right)} \]
  4. Taylor expanded in g around 0

    \[\leadsto \color{blue}{2 \cdot \left(-1 \cdot \left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right) + \frac{-1}{2} \cdot \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right)\right)} \]
  5. Step-by-step derivation
    1. distribute-lft-inN/A

      \[\leadsto \color{blue}{2 \cdot \left(-1 \cdot \left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right)\right) + 2 \cdot \left(\frac{-1}{2} \cdot \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right)\right)} \]
    2. *-commutativeN/A

      \[\leadsto \color{blue}{\left(-1 \cdot \left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right)\right) \cdot 2} + 2 \cdot \left(\frac{-1}{2} \cdot \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right)\right) \]
    3. associate-*l*N/A

      \[\leadsto \color{blue}{-1 \cdot \left(\left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right) \cdot 2\right)} + 2 \cdot \left(\frac{-1}{2} \cdot \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right)\right) \]
    4. associate-*r*N/A

      \[\leadsto -1 \cdot \left(\left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right) \cdot 2\right) + \color{blue}{\left(2 \cdot \frac{-1}{2}\right) \cdot \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right)} \]
    5. metadata-evalN/A

      \[\leadsto -1 \cdot \left(\left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right) \cdot 2\right) + \color{blue}{-1} \cdot \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \]
    6. distribute-lft-outN/A

      \[\leadsto \color{blue}{-1 \cdot \left(\left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right) \cdot 2 + \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right)\right)} \]
    7. mul-1-negN/A

      \[\leadsto \color{blue}{\mathsf{neg}\left(\left(\left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right) \cdot 2 + \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right)\right)\right)} \]
    8. lower-neg.f64N/A

      \[\leadsto \color{blue}{-\left(\left(\sin \left(\frac{1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right) \cdot \sin \left(\frac{2}{3} \cdot \mathsf{PI}\left(\right)\right)\right) \cdot 2 + \cos \left(\frac{-1}{3} \cdot \cos^{-1} \left(-1 \cdot \frac{g}{h}\right)\right)\right)} \]
  6. Applied rewrites98.5%

    \[\leadsto \color{blue}{-\mathsf{fma}\left(\sin \left(\mathsf{PI}\left(\right) \cdot 0.6666666666666666\right) \cdot 2, \sin \left(\cos^{-1} \left(\frac{-g}{h}\right) \cdot 0.3333333333333333\right), \cos \left(-0.3333333333333333 \cdot \cos^{-1} \left(\frac{-g}{h}\right)\right)\right)} \]
  7. Step-by-step derivation
    1. Applied rewrites100.0%

      \[\leadsto -\mathsf{fma}\left(\sin \left(\cos^{-1} \left(\frac{-g}{h}\right) \cdot 0.3333333333333333\right) \cdot \sin \left(\mathsf{PI}\left(\right) \cdot 0.6666666666666666\right), 2, \cos \left(\cos^{-1} \left(\frac{-g}{h}\right) \cdot -0.3333333333333333\right)\right) \]
    2. Add Preprocessing

    Alternative 2: 98.5% accurate, 1.0× speedup?

    \[\begin{array}{l} \\ 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, 6 \cdot \mathsf{PI}\left(\right)\right)}{9}\right) \end{array} \]
    (FPCore (g h)
     :precision binary64
     (* 2.0 (cos (/ (fma (acos (/ (- g) h)) 3.0 (* 6.0 (PI))) 9.0))))
    \begin{array}{l}
    
    \\
    2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, 6 \cdot \mathsf{PI}\left(\right)\right)}{9}\right)
    \end{array}
    
    Derivation
    1. Initial program 98.5%

      \[2 \cdot \cos \left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
    2. Add Preprocessing
    3. Step-by-step derivation
      1. lift-+.f64N/A

        \[\leadsto 2 \cdot \cos \color{blue}{\left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right)} \]
      2. lift-/.f64N/A

        \[\leadsto 2 \cdot \cos \left(\color{blue}{\frac{2 \cdot \mathsf{PI}\left(\right)}{3}} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
      3. lift-/.f64N/A

        \[\leadsto 2 \cdot \cos \left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \color{blue}{\frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}}\right) \]
      4. frac-addN/A

        \[\leadsto 2 \cdot \cos \color{blue}{\left(\frac{\left(2 \cdot \mathsf{PI}\left(\right)\right) \cdot 3 + 3 \cdot \cos^{-1} \left(\frac{-g}{h}\right)}{3 \cdot 3}\right)} \]
      5. lower-/.f64N/A

        \[\leadsto 2 \cdot \cos \color{blue}{\left(\frac{\left(2 \cdot \mathsf{PI}\left(\right)\right) \cdot 3 + 3 \cdot \cos^{-1} \left(\frac{-g}{h}\right)}{3 \cdot 3}\right)} \]
      6. +-commutativeN/A

        \[\leadsto 2 \cdot \cos \left(\frac{\color{blue}{3 \cdot \cos^{-1} \left(\frac{-g}{h}\right) + \left(2 \cdot \mathsf{PI}\left(\right)\right) \cdot 3}}{3 \cdot 3}\right) \]
      7. *-commutativeN/A

        \[\leadsto 2 \cdot \cos \left(\frac{\color{blue}{\cos^{-1} \left(\frac{-g}{h}\right) \cdot 3} + \left(2 \cdot \mathsf{PI}\left(\right)\right) \cdot 3}{3 \cdot 3}\right) \]
      8. lower-fma.f64N/A

        \[\leadsto 2 \cdot \cos \left(\frac{\color{blue}{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, \left(2 \cdot \mathsf{PI}\left(\right)\right) \cdot 3\right)}}{3 \cdot 3}\right) \]
      9. *-commutativeN/A

        \[\leadsto 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, \color{blue}{3 \cdot \left(2 \cdot \mathsf{PI}\left(\right)\right)}\right)}{3 \cdot 3}\right) \]
      10. lift-*.f64N/A

        \[\leadsto 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, 3 \cdot \color{blue}{\left(2 \cdot \mathsf{PI}\left(\right)\right)}\right)}{3 \cdot 3}\right) \]
      11. associate-*r*N/A

        \[\leadsto 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, \color{blue}{\left(3 \cdot 2\right) \cdot \mathsf{PI}\left(\right)}\right)}{3 \cdot 3}\right) \]
      12. metadata-evalN/A

        \[\leadsto 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, \color{blue}{6} \cdot \mathsf{PI}\left(\right)\right)}{3 \cdot 3}\right) \]
      13. metadata-evalN/A

        \[\leadsto 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, \color{blue}{\left(2 \cdot 3\right)} \cdot \mathsf{PI}\left(\right)\right)}{3 \cdot 3}\right) \]
      14. lower-*.f64N/A

        \[\leadsto 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, \color{blue}{\left(2 \cdot 3\right) \cdot \mathsf{PI}\left(\right)}\right)}{3 \cdot 3}\right) \]
      15. metadata-evalN/A

        \[\leadsto 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, \color{blue}{6} \cdot \mathsf{PI}\left(\right)\right)}{3 \cdot 3}\right) \]
      16. metadata-eval98.5

        \[\leadsto 2 \cdot \cos \left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, 6 \cdot \mathsf{PI}\left(\right)\right)}{\color{blue}{9}}\right) \]
    4. Applied rewrites98.5%

      \[\leadsto 2 \cdot \cos \color{blue}{\left(\frac{\mathsf{fma}\left(\cos^{-1} \left(\frac{-g}{h}\right), 3, 6 \cdot \mathsf{PI}\left(\right)\right)}{9}\right)} \]
    5. Add Preprocessing

    Alternative 3: 98.5% accurate, 1.1× speedup?

    \[\begin{array}{l} \\ 2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), 0.6666666666666666, 0.3333333333333333 \cdot \cos^{-1} \left(\frac{-g}{h}\right)\right)\right) \end{array} \]
    (FPCore (g h)
     :precision binary64
     (*
      2.0
      (cos
       (fma (PI) 0.6666666666666666 (* 0.3333333333333333 (acos (/ (- g) h)))))))
    \begin{array}{l}
    
    \\
    2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), 0.6666666666666666, 0.3333333333333333 \cdot \cos^{-1} \left(\frac{-g}{h}\right)\right)\right)
    \end{array}
    
    Derivation
    1. Initial program 98.5%

      \[2 \cdot \cos \left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
    2. Add Preprocessing
    3. Step-by-step derivation
      1. lift-+.f64N/A

        \[\leadsto 2 \cdot \cos \color{blue}{\left(\frac{2 \cdot \mathsf{PI}\left(\right)}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right)} \]
      2. lift-/.f64N/A

        \[\leadsto 2 \cdot \cos \left(\color{blue}{\frac{2 \cdot \mathsf{PI}\left(\right)}{3}} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
      3. div-invN/A

        \[\leadsto 2 \cdot \cos \left(\color{blue}{\left(2 \cdot \mathsf{PI}\left(\right)\right) \cdot \frac{1}{3}} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
      4. lift-*.f64N/A

        \[\leadsto 2 \cdot \cos \left(\color{blue}{\left(2 \cdot \mathsf{PI}\left(\right)\right)} \cdot \frac{1}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
      5. *-commutativeN/A

        \[\leadsto 2 \cdot \cos \left(\color{blue}{\left(\mathsf{PI}\left(\right) \cdot 2\right)} \cdot \frac{1}{3} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
      6. associate-*l*N/A

        \[\leadsto 2 \cdot \cos \left(\color{blue}{\mathsf{PI}\left(\right) \cdot \left(2 \cdot \frac{1}{3}\right)} + \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right) \]
      7. lower-fma.f64N/A

        \[\leadsto 2 \cdot \cos \color{blue}{\left(\mathsf{fma}\left(\mathsf{PI}\left(\right), 2 \cdot \frac{1}{3}, \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right)\right)} \]
      8. metadata-evalN/A

        \[\leadsto 2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), 2 \cdot \color{blue}{\frac{1}{3}}, \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right)\right) \]
      9. metadata-eval98.5

        \[\leadsto 2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), \color{blue}{0.6666666666666666}, \frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}\right)\right) \]
      10. lift-/.f64N/A

        \[\leadsto 2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), \frac{2}{3}, \color{blue}{\frac{\cos^{-1} \left(\frac{-g}{h}\right)}{3}}\right)\right) \]
      11. clear-numN/A

        \[\leadsto 2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), \frac{2}{3}, \color{blue}{\frac{1}{\frac{3}{\cos^{-1} \left(\frac{-g}{h}\right)}}}\right)\right) \]
      12. associate-/r/N/A

        \[\leadsto 2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), \frac{2}{3}, \color{blue}{\frac{1}{3} \cdot \cos^{-1} \left(\frac{-g}{h}\right)}\right)\right) \]
      13. lower-*.f64N/A

        \[\leadsto 2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), \frac{2}{3}, \color{blue}{\frac{1}{3} \cdot \cos^{-1} \left(\frac{-g}{h}\right)}\right)\right) \]
      14. metadata-eval98.5

        \[\leadsto 2 \cdot \cos \left(\mathsf{fma}\left(\mathsf{PI}\left(\right), 0.6666666666666666, \color{blue}{0.3333333333333333} \cdot \cos^{-1} \left(\frac{-g}{h}\right)\right)\right) \]
    4. Applied rewrites98.5%

      \[\leadsto 2 \cdot \cos \color{blue}{\left(\mathsf{fma}\left(\mathsf{PI}\left(\right), 0.6666666666666666, 0.3333333333333333 \cdot \cos^{-1} \left(\frac{-g}{h}\right)\right)\right)} \]
    5. Add Preprocessing

    Reproduce

    ?
    herbie shell --seed 2024322 
    (FPCore (g h)
      :name "2-ancestry mixing, negative discriminant"
      :precision binary64
      (* 2.0 (cos (+ (/ (* 2.0 (PI)) 3.0) (/ (acos (/ (- g) h)) 3.0)))))