NMSE Section 6.1 mentioned, B

Percentage Accurate: 79.2% → 99.6%
Time: 8.0s
Alternatives: 7
Speedup: 2.4×

Specification

?
\[\begin{array}{l} \\ \left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \end{array} \]
(FPCore (a b)
 :precision binary64
 (* (* (/ (PI) 2.0) (/ 1.0 (- (* b b) (* a a)))) (- (/ 1.0 a) (/ 1.0 b))))
\begin{array}{l}

\\
\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\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 7 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: 79.2% accurate, 1.0× speedup?

\[\begin{array}{l} \\ \left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \end{array} \]
(FPCore (a b)
 :precision binary64
 (* (* (/ (PI) 2.0) (/ 1.0 (- (* b b) (* a a)))) (- (/ 1.0 a) (/ 1.0 b))))
\begin{array}{l}

\\
\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right)
\end{array}

Alternative 1: 99.6% accurate, 2.0× speedup?

\[\begin{array}{l} \\ \frac{\frac{0.5}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{a + b} \end{array} \]
(FPCore (a b) :precision binary64 (/ (* (/ 0.5 (* a b)) (PI)) (+ a b)))
\begin{array}{l}

\\
\frac{\frac{0.5}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{a + b}
\end{array}
Derivation
  1. Initial program 75.5%

    \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
  2. Add Preprocessing
  3. Applied rewrites87.4%

    \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{PI}\left(\right)}{a + b}, \frac{0.5}{b - a} \cdot {a}^{-1}, \frac{\frac{0.5}{a + b} \cdot \frac{\mathsf{PI}\left(\right)}{b - a}}{-b}\right)} \]
  4. Applied rewrites99.1%

    \[\leadsto \color{blue}{\frac{1}{\frac{b \cdot a}{0.5 \cdot \mathsf{PI}\left(\right)} \cdot \left(b + a\right)}} \]
  5. Step-by-step derivation
    1. lift-/.f64N/A

      \[\leadsto \color{blue}{\frac{1}{\frac{b \cdot a}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)} \cdot \left(b + a\right)}} \]
    2. lift-*.f64N/A

      \[\leadsto \frac{1}{\color{blue}{\frac{b \cdot a}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)} \cdot \left(b + a\right)}} \]
    3. associate-/r*N/A

      \[\leadsto \color{blue}{\frac{\frac{1}{\frac{b \cdot a}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}}}{b + a}} \]
    4. lift-/.f64N/A

      \[\leadsto \frac{\frac{1}{\color{blue}{\frac{b \cdot a}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}}}}{b + a} \]
    5. clear-numN/A

      \[\leadsto \frac{\color{blue}{\frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{b \cdot a}}}{b + a} \]
    6. lower-/.f64N/A

      \[\leadsto \color{blue}{\frac{\frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{b \cdot a}}{b + a}} \]
    7. lift-*.f64N/A

      \[\leadsto \frac{\frac{\color{blue}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}}{b \cdot a}}{b + a} \]
    8. *-commutativeN/A

      \[\leadsto \frac{\frac{\color{blue}{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}}{b \cdot a}}{b + a} \]
    9. associate-*r/N/A

      \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right) \cdot \frac{\frac{1}{2}}{b \cdot a}}}{b + a} \]
    10. lift-/.f64N/A

      \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \color{blue}{\frac{\frac{1}{2}}{b \cdot a}}}{b + a} \]
    11. *-commutativeN/A

      \[\leadsto \frac{\color{blue}{\frac{\frac{1}{2}}{b \cdot a} \cdot \mathsf{PI}\left(\right)}}{b + a} \]
    12. lower-*.f6499.7

      \[\leadsto \frac{\color{blue}{\frac{0.5}{b \cdot a} \cdot \mathsf{PI}\left(\right)}}{b + a} \]
    13. lift-*.f64N/A

      \[\leadsto \frac{\frac{\frac{1}{2}}{\color{blue}{b \cdot a}} \cdot \mathsf{PI}\left(\right)}{b + a} \]
    14. *-commutativeN/A

      \[\leadsto \frac{\frac{\frac{1}{2}}{\color{blue}{a \cdot b}} \cdot \mathsf{PI}\left(\right)}{b + a} \]
    15. lower-*.f6499.7

      \[\leadsto \frac{\frac{0.5}{\color{blue}{a \cdot b}} \cdot \mathsf{PI}\left(\right)}{b + a} \]
    16. lift-+.f64N/A

      \[\leadsto \frac{\frac{\frac{1}{2}}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{\color{blue}{b + a}} \]
    17. +-commutativeN/A

      \[\leadsto \frac{\frac{\frac{1}{2}}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{\color{blue}{a + b}} \]
    18. lower-+.f6499.7

      \[\leadsto \frac{\frac{0.5}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{\color{blue}{a + b}} \]
  6. Applied rewrites99.7%

    \[\leadsto \color{blue}{\frac{\frac{0.5}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{a + b}} \]
  7. Add Preprocessing

Alternative 2: 74.7% accurate, 2.2× speedup?

\[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;a \leq -1.75 \cdot 10^{-33}:\\ \;\;\;\;\frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5\\ \mathbf{else}:\\ \;\;\;\;\frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot b} \cdot 0.5\\ \end{array} \end{array} \]
(FPCore (a b)
 :precision binary64
 (if (<= a -1.75e-33)
   (* (/ (PI) (* (* b a) a)) 0.5)
   (* (/ (PI) (* (* b a) b)) 0.5)))
\begin{array}{l}

\\
\begin{array}{l}
\mathbf{if}\;a \leq -1.75 \cdot 10^{-33}:\\
\;\;\;\;\frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5\\

\mathbf{else}:\\
\;\;\;\;\frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot b} \cdot 0.5\\


\end{array}
\end{array}
Derivation
  1. Split input into 2 regimes
  2. if a < -1.7499999999999999e-33

    1. Initial program 69.3%

      \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
    2. Add Preprocessing
    3. Taylor expanded in a around inf

      \[\leadsto \color{blue}{\frac{1}{2} \cdot \frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \]
    4. Step-by-step derivation
      1. *-commutativeN/A

        \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
      2. lower-*.f64N/A

        \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
      3. lower-/.f64N/A

        \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
      4. lower-PI.f64N/A

        \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right)}}{{a}^{2} \cdot b} \cdot \frac{1}{2} \]
      5. lower-*.f64N/A

        \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
      6. unpow2N/A

        \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot \frac{1}{2} \]
      7. lower-*.f6480.1

        \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot 0.5 \]
    5. Applied rewrites80.1%

      \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{\left(a \cdot a\right) \cdot b} \cdot 0.5} \]
    6. Step-by-step derivation
      1. Applied rewrites92.2%

        \[\leadsto \frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5 \]

      if -1.7499999999999999e-33 < a

      1. Initial program 78.1%

        \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
      2. Add Preprocessing
      3. Applied rewrites88.4%

        \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{PI}\left(\right)}{a + b}, \frac{0.5}{b - a} \cdot {a}^{-1}, \frac{\frac{0.5}{a + b} \cdot \frac{\mathsf{PI}\left(\right)}{b - a}}{-b}\right)} \]
      4. Applied rewrites99.4%

        \[\leadsto \color{blue}{\frac{1}{\frac{b \cdot a}{0.5 \cdot \mathsf{PI}\left(\right)} \cdot \left(b + a\right)}} \]
      5. Taylor expanded in a around 0

        \[\leadsto \color{blue}{\frac{1}{2} \cdot \frac{\mathsf{PI}\left(\right)}{a \cdot {b}^{2}}} \]
      6. Step-by-step derivation
        1. *-commutativeN/A

          \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{a \cdot {b}^{2}} \cdot \frac{1}{2}} \]
        2. lower-*.f64N/A

          \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{a \cdot {b}^{2}} \cdot \frac{1}{2}} \]
        3. lower-/.f64N/A

          \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{a \cdot {b}^{2}}} \cdot \frac{1}{2} \]
        4. lower-PI.f64N/A

          \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right)}}{a \cdot {b}^{2}} \cdot \frac{1}{2} \]
        5. *-commutativeN/A

          \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{b}^{2} \cdot a}} \cdot \frac{1}{2} \]
        6. lower-*.f64N/A

          \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{b}^{2} \cdot a}} \cdot \frac{1}{2} \]
        7. unpow2N/A

          \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(b \cdot b\right)} \cdot a} \cdot \frac{1}{2} \]
        8. lower-*.f6466.7

          \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(b \cdot b\right)} \cdot a} \cdot 0.5 \]
      7. Applied rewrites66.7%

        \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{\left(b \cdot b\right) \cdot a} \cdot 0.5} \]
      8. Step-by-step derivation
        1. Applied rewrites72.9%

          \[\leadsto \frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot b} \cdot 0.5 \]
      9. Recombined 2 regimes into one program.
      10. Add Preprocessing

      Alternative 3: 68.9% accurate, 2.2× speedup?

      \[\begin{array}{l} \\ \begin{array}{l} \mathbf{if}\;a \leq -1.55 \cdot 10^{-33}:\\ \;\;\;\;\frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5\\ \mathbf{else}:\\ \;\;\;\;\frac{\mathsf{PI}\left(\right)}{\left(b \cdot b\right) \cdot a} \cdot 0.5\\ \end{array} \end{array} \]
      (FPCore (a b)
       :precision binary64
       (if (<= a -1.55e-33)
         (* (/ (PI) (* (* b a) a)) 0.5)
         (* (/ (PI) (* (* b b) a)) 0.5)))
      \begin{array}{l}
      
      \\
      \begin{array}{l}
      \mathbf{if}\;a \leq -1.55 \cdot 10^{-33}:\\
      \;\;\;\;\frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5\\
      
      \mathbf{else}:\\
      \;\;\;\;\frac{\mathsf{PI}\left(\right)}{\left(b \cdot b\right) \cdot a} \cdot 0.5\\
      
      
      \end{array}
      \end{array}
      
      Derivation
      1. Split input into 2 regimes
      2. if a < -1.54999999999999998e-33

        1. Initial program 69.3%

          \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
        2. Add Preprocessing
        3. Taylor expanded in a around inf

          \[\leadsto \color{blue}{\frac{1}{2} \cdot \frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \]
        4. Step-by-step derivation
          1. *-commutativeN/A

            \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
          2. lower-*.f64N/A

            \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
          3. lower-/.f64N/A

            \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
          4. lower-PI.f64N/A

            \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right)}}{{a}^{2} \cdot b} \cdot \frac{1}{2} \]
          5. lower-*.f64N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
          6. unpow2N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot \frac{1}{2} \]
          7. lower-*.f6480.1

            \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot 0.5 \]
        5. Applied rewrites80.1%

          \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{\left(a \cdot a\right) \cdot b} \cdot 0.5} \]
        6. Step-by-step derivation
          1. Applied rewrites92.2%

            \[\leadsto \frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5 \]

          if -1.54999999999999998e-33 < a

          1. Initial program 78.1%

            \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
          2. Add Preprocessing
          3. Taylor expanded in a around 0

            \[\leadsto \color{blue}{\frac{1}{2} \cdot \frac{\mathsf{PI}\left(\right)}{a \cdot {b}^{2}}} \]
          4. Step-by-step derivation
            1. *-commutativeN/A

              \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{a \cdot {b}^{2}} \cdot \frac{1}{2}} \]
            2. lower-*.f64N/A

              \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{a \cdot {b}^{2}} \cdot \frac{1}{2}} \]
            3. lower-/.f64N/A

              \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{a \cdot {b}^{2}}} \cdot \frac{1}{2} \]
            4. lower-PI.f64N/A

              \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right)}}{a \cdot {b}^{2}} \cdot \frac{1}{2} \]
            5. *-commutativeN/A

              \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{b}^{2} \cdot a}} \cdot \frac{1}{2} \]
            6. lower-*.f64N/A

              \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{b}^{2} \cdot a}} \cdot \frac{1}{2} \]
            7. unpow2N/A

              \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(b \cdot b\right)} \cdot a} \cdot \frac{1}{2} \]
            8. lower-*.f6466.7

              \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(b \cdot b\right)} \cdot a} \cdot 0.5 \]
          5. Applied rewrites66.7%

            \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{\left(b \cdot b\right) \cdot a} \cdot 0.5} \]
        7. Recombined 2 regimes into one program.
        8. Add Preprocessing

        Alternative 4: 99.1% accurate, 2.4× speedup?

        \[\begin{array}{l} \\ \frac{\mathsf{PI}\left(\right) \cdot 0.5}{\left(b + a\right) \cdot \left(b \cdot a\right)} \end{array} \]
        (FPCore (a b) :precision binary64 (/ (* (PI) 0.5) (* (+ b a) (* b a))))
        \begin{array}{l}
        
        \\
        \frac{\mathsf{PI}\left(\right) \cdot 0.5}{\left(b + a\right) \cdot \left(b \cdot a\right)}
        \end{array}
        
        Derivation
        1. Initial program 75.5%

          \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
        2. Add Preprocessing
        3. Applied rewrites87.4%

          \[\leadsto \color{blue}{\mathsf{fma}\left(\frac{\mathsf{PI}\left(\right)}{a + b}, \frac{0.5}{b - a} \cdot {a}^{-1}, \frac{\frac{0.5}{a + b} \cdot \frac{\mathsf{PI}\left(\right)}{b - a}}{-b}\right)} \]
        4. Applied rewrites99.1%

          \[\leadsto \color{blue}{\frac{1}{\frac{b \cdot a}{0.5 \cdot \mathsf{PI}\left(\right)} \cdot \left(b + a\right)}} \]
        5. Step-by-step derivation
          1. lift-/.f64N/A

            \[\leadsto \color{blue}{\frac{1}{\frac{b \cdot a}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)} \cdot \left(b + a\right)}} \]
          2. lift-*.f64N/A

            \[\leadsto \frac{1}{\color{blue}{\frac{b \cdot a}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)} \cdot \left(b + a\right)}} \]
          3. associate-/r*N/A

            \[\leadsto \color{blue}{\frac{\frac{1}{\frac{b \cdot a}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}}}{b + a}} \]
          4. lift-/.f64N/A

            \[\leadsto \frac{\frac{1}{\color{blue}{\frac{b \cdot a}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}}}}{b + a} \]
          5. clear-numN/A

            \[\leadsto \frac{\color{blue}{\frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{b \cdot a}}}{b + a} \]
          6. lower-/.f64N/A

            \[\leadsto \color{blue}{\frac{\frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{b \cdot a}}{b + a}} \]
          7. lift-*.f64N/A

            \[\leadsto \frac{\frac{\color{blue}{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}}{b \cdot a}}{b + a} \]
          8. *-commutativeN/A

            \[\leadsto \frac{\frac{\color{blue}{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}}{b \cdot a}}{b + a} \]
          9. associate-*r/N/A

            \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right) \cdot \frac{\frac{1}{2}}{b \cdot a}}}{b + a} \]
          10. lift-/.f64N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \color{blue}{\frac{\frac{1}{2}}{b \cdot a}}}{b + a} \]
          11. *-commutativeN/A

            \[\leadsto \frac{\color{blue}{\frac{\frac{1}{2}}{b \cdot a} \cdot \mathsf{PI}\left(\right)}}{b + a} \]
          12. lower-*.f6499.7

            \[\leadsto \frac{\color{blue}{\frac{0.5}{b \cdot a} \cdot \mathsf{PI}\left(\right)}}{b + a} \]
          13. lift-*.f64N/A

            \[\leadsto \frac{\frac{\frac{1}{2}}{\color{blue}{b \cdot a}} \cdot \mathsf{PI}\left(\right)}{b + a} \]
          14. *-commutativeN/A

            \[\leadsto \frac{\frac{\frac{1}{2}}{\color{blue}{a \cdot b}} \cdot \mathsf{PI}\left(\right)}{b + a} \]
          15. lower-*.f6499.7

            \[\leadsto \frac{\frac{0.5}{\color{blue}{a \cdot b}} \cdot \mathsf{PI}\left(\right)}{b + a} \]
          16. lift-+.f64N/A

            \[\leadsto \frac{\frac{\frac{1}{2}}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{\color{blue}{b + a}} \]
          17. +-commutativeN/A

            \[\leadsto \frac{\frac{\frac{1}{2}}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{\color{blue}{a + b}} \]
          18. lower-+.f6499.7

            \[\leadsto \frac{\frac{0.5}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{\color{blue}{a + b}} \]
        6. Applied rewrites99.7%

          \[\leadsto \color{blue}{\frac{\frac{0.5}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{a + b}} \]
        7. Step-by-step derivation
          1. lift-/.f64N/A

            \[\leadsto \color{blue}{\frac{\frac{\frac{1}{2}}{a \cdot b} \cdot \mathsf{PI}\left(\right)}{a + b}} \]
          2. lift-*.f64N/A

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

            \[\leadsto \color{blue}{\frac{\frac{1}{2}}{a \cdot b} \cdot \frac{\mathsf{PI}\left(\right)}{a + b}} \]
          4. lift-/.f64N/A

            \[\leadsto \color{blue}{\frac{\frac{1}{2}}{a \cdot b}} \cdot \frac{\mathsf{PI}\left(\right)}{a + b} \]
          5. frac-timesN/A

            \[\leadsto \color{blue}{\frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{\left(a \cdot b\right) \cdot \left(a + b\right)}} \]
          6. lift-PI.f64N/A

            \[\leadsto \frac{\frac{1}{2} \cdot \color{blue}{\mathsf{PI}\left(\right)}}{\left(a \cdot b\right) \cdot \left(a + b\right)} \]
          7. lift-*.f64N/A

            \[\leadsto \frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot b\right)} \cdot \left(a + b\right)} \]
          8. *-commutativeN/A

            \[\leadsto \frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{\color{blue}{\left(b \cdot a\right)} \cdot \left(a + b\right)} \]
          9. lift-+.f64N/A

            \[\leadsto \frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot \color{blue}{\left(a + b\right)}} \]
          10. +-commutativeN/A

            \[\leadsto \frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot \color{blue}{\left(b + a\right)}} \]
          11. lower-/.f64N/A

            \[\leadsto \color{blue}{\frac{\frac{1}{2} \cdot \mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot \left(b + a\right)}} \]
          12. lift-PI.f64N/A

            \[\leadsto \frac{\frac{1}{2} \cdot \color{blue}{\mathsf{PI}\left(\right)}}{\left(b \cdot a\right) \cdot \left(b + a\right)} \]
          13. *-commutativeN/A

            \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}}{\left(b \cdot a\right) \cdot \left(b + a\right)} \]
          14. lower-*.f64N/A

            \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}}{\left(b \cdot a\right) \cdot \left(b + a\right)} \]
          15. *-commutativeN/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\color{blue}{\left(a \cdot b\right)} \cdot \left(b + a\right)} \]
          16. lift-*.f64N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\color{blue}{\left(a \cdot b\right)} \cdot \left(b + a\right)} \]
          17. +-commutativeN/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\left(a \cdot b\right) \cdot \color{blue}{\left(a + b\right)}} \]
          18. lift-+.f64N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\left(a \cdot b\right) \cdot \color{blue}{\left(a + b\right)}} \]
          19. *-commutativeN/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\color{blue}{\left(a + b\right) \cdot \left(a \cdot b\right)}} \]
          20. lower-*.f6498.8

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot 0.5}{\color{blue}{\left(a + b\right) \cdot \left(a \cdot b\right)}} \]
          21. lift-+.f64N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\color{blue}{\left(a + b\right)} \cdot \left(a \cdot b\right)} \]
          22. +-commutativeN/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\color{blue}{\left(b + a\right)} \cdot \left(a \cdot b\right)} \]
          23. lower-+.f6498.8

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot 0.5}{\color{blue}{\left(b + a\right)} \cdot \left(a \cdot b\right)} \]
          24. lift-*.f64N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\left(b + a\right) \cdot \color{blue}{\left(a \cdot b\right)}} \]
          25. *-commutativeN/A

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot \frac{1}{2}}{\left(b + a\right) \cdot \color{blue}{\left(b \cdot a\right)}} \]
          26. lower-*.f6498.8

            \[\leadsto \frac{\mathsf{PI}\left(\right) \cdot 0.5}{\left(b + a\right) \cdot \color{blue}{\left(b \cdot a\right)}} \]
        8. Applied rewrites98.8%

          \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right) \cdot 0.5}{\left(b + a\right) \cdot \left(b \cdot a\right)}} \]
        9. Add Preprocessing

        Alternative 5: 63.3% accurate, 2.6× speedup?

        \[\begin{array}{l} \\ \frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5 \end{array} \]
        (FPCore (a b) :precision binary64 (* (/ (PI) (* (* b a) a)) 0.5))
        \begin{array}{l}
        
        \\
        \frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5
        \end{array}
        
        Derivation
        1. Initial program 75.5%

          \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
        2. Add Preprocessing
        3. Taylor expanded in a around inf

          \[\leadsto \color{blue}{\frac{1}{2} \cdot \frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \]
        4. Step-by-step derivation
          1. *-commutativeN/A

            \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
          2. lower-*.f64N/A

            \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
          3. lower-/.f64N/A

            \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
          4. lower-PI.f64N/A

            \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right)}}{{a}^{2} \cdot b} \cdot \frac{1}{2} \]
          5. lower-*.f64N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
          6. unpow2N/A

            \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot \frac{1}{2} \]
          7. lower-*.f6459.3

            \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot 0.5 \]
        5. Applied rewrites59.3%

          \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{\left(a \cdot a\right) \cdot b} \cdot 0.5} \]
        6. Step-by-step derivation
          1. Applied rewrites66.8%

            \[\leadsto \frac{\mathsf{PI}\left(\right)}{\left(b \cdot a\right) \cdot a} \cdot 0.5 \]
          2. Add Preprocessing

          Alternative 6: 63.2% accurate, 2.6× speedup?

          \[\begin{array}{l} \\ \mathsf{PI}\left(\right) \cdot \frac{0.5}{\left(a \cdot b\right) \cdot a} \end{array} \]
          (FPCore (a b) :precision binary64 (* (PI) (/ 0.5 (* (* a b) a))))
          \begin{array}{l}
          
          \\
          \mathsf{PI}\left(\right) \cdot \frac{0.5}{\left(a \cdot b\right) \cdot a}
          \end{array}
          
          Derivation
          1. Initial program 75.5%

            \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
          2. Add Preprocessing
          3. Taylor expanded in a around inf

            \[\leadsto \color{blue}{\frac{1}{2} \cdot \frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \]
          4. Step-by-step derivation
            1. *-commutativeN/A

              \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
            2. lower-*.f64N/A

              \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
            3. lower-/.f64N/A

              \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
            4. lower-PI.f64N/A

              \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right)}}{{a}^{2} \cdot b} \cdot \frac{1}{2} \]
            5. lower-*.f64N/A

              \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
            6. unpow2N/A

              \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot \frac{1}{2} \]
            7. lower-*.f6459.3

              \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot 0.5 \]
          5. Applied rewrites59.3%

            \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{\left(a \cdot a\right) \cdot b} \cdot 0.5} \]
          6. Step-by-step derivation
            1. Applied rewrites59.3%

              \[\leadsto \mathsf{PI}\left(\right) \cdot \color{blue}{\frac{0.5}{\left(a \cdot a\right) \cdot b}} \]
            2. Step-by-step derivation
              1. Applied rewrites66.8%

                \[\leadsto \mathsf{PI}\left(\right) \cdot \frac{0.5}{\left(a \cdot b\right) \cdot \color{blue}{a}} \]
              2. Add Preprocessing

              Alternative 7: 57.8% accurate, 2.6× speedup?

              \[\begin{array}{l} \\ \mathsf{PI}\left(\right) \cdot \frac{0.5}{\left(a \cdot a\right) \cdot b} \end{array} \]
              (FPCore (a b) :precision binary64 (* (PI) (/ 0.5 (* (* a a) b))))
              \begin{array}{l}
              
              \\
              \mathsf{PI}\left(\right) \cdot \frac{0.5}{\left(a \cdot a\right) \cdot b}
              \end{array}
              
              Derivation
              1. Initial program 75.5%

                \[\left(\frac{\mathsf{PI}\left(\right)}{2} \cdot \frac{1}{b \cdot b - a \cdot a}\right) \cdot \left(\frac{1}{a} - \frac{1}{b}\right) \]
              2. Add Preprocessing
              3. Taylor expanded in a around inf

                \[\leadsto \color{blue}{\frac{1}{2} \cdot \frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \]
              4. Step-by-step derivation
                1. *-commutativeN/A

                  \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
                2. lower-*.f64N/A

                  \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b} \cdot \frac{1}{2}} \]
                3. lower-/.f64N/A

                  \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
                4. lower-PI.f64N/A

                  \[\leadsto \frac{\color{blue}{\mathsf{PI}\left(\right)}}{{a}^{2} \cdot b} \cdot \frac{1}{2} \]
                5. lower-*.f64N/A

                  \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{{a}^{2} \cdot b}} \cdot \frac{1}{2} \]
                6. unpow2N/A

                  \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot \frac{1}{2} \]
                7. lower-*.f6459.3

                  \[\leadsto \frac{\mathsf{PI}\left(\right)}{\color{blue}{\left(a \cdot a\right)} \cdot b} \cdot 0.5 \]
              5. Applied rewrites59.3%

                \[\leadsto \color{blue}{\frac{\mathsf{PI}\left(\right)}{\left(a \cdot a\right) \cdot b} \cdot 0.5} \]
              6. Step-by-step derivation
                1. Applied rewrites59.3%

                  \[\leadsto \mathsf{PI}\left(\right) \cdot \color{blue}{\frac{0.5}{\left(a \cdot a\right) \cdot b}} \]
                2. Add Preprocessing

                Reproduce

                ?
                herbie shell --seed 2024313 
                (FPCore (a b)
                  :name "NMSE Section 6.1 mentioned, B"
                  :precision binary64
                  (* (* (/ (PI) 2.0) (/ 1.0 (- (* b b) (* a a)))) (- (/ 1.0 a) (/ 1.0 b))))