Average Error: 0.6 → 0.8
Time: 26.2s
Precision: 64
Internal Precision: 128
\[\cos^{-1} \left(\frac{1 - 5 \cdot \left(v \cdot v\right)}{v \cdot v - 1}\right)\]
\[e^{\log \left((e^{(e^{\log_* (1 + \log_* (1 + \cos^{-1} \left((4 \cdot \left((v \cdot v + \left({v}^{4}\right))_*\right) + -1)_*\right)))} - 1)^*} - 1)^*\right)}\]

Error

Bits error versus v

Derivation

  1. Initial program 0.6

    \[\cos^{-1} \left(\frac{1 - 5 \cdot \left(v \cdot v\right)}{v \cdot v - 1}\right)\]
  2. Initial simplification0.6

    \[\leadsto \cos^{-1} \left(\frac{(\left(-5 \cdot v\right) \cdot v + 1)_*}{(v \cdot v + -1)_*}\right)\]
  3. Taylor expanded around 0 0.8

    \[\leadsto \cos^{-1} \color{blue}{\left(\left(4 \cdot {v}^{4} + 4 \cdot {v}^{2}\right) - 1\right)}\]
  4. Simplified0.8

    \[\leadsto \cos^{-1} \color{blue}{\left((4 \cdot \left((v \cdot v + \left({v}^{4}\right))_*\right) + -1)_*\right)}\]
  5. Using strategy rm
  6. Applied expm1-log1p-u0.8

    \[\leadsto \color{blue}{(e^{\log_* (1 + \cos^{-1} \left((4 \cdot \left((v \cdot v + \left({v}^{4}\right))_*\right) + -1)_*\right))} - 1)^*}\]
  7. Using strategy rm
  8. Applied expm1-log1p-u0.8

    \[\leadsto (e^{\color{blue}{(e^{\log_* (1 + \log_* (1 + \cos^{-1} \left((4 \cdot \left((v \cdot v + \left({v}^{4}\right))_*\right) + -1)_*\right)))} - 1)^*}} - 1)^*\]
  9. Using strategy rm
  10. Applied add-exp-log0.8

    \[\leadsto \color{blue}{e^{\log \left((e^{(e^{\log_* (1 + \log_* (1 + \cos^{-1} \left((4 \cdot \left((v \cdot v + \left({v}^{4}\right))_*\right) + -1)_*\right)))} - 1)^*} - 1)^*\right)}}\]
  11. Final simplification0.8

    \[\leadsto e^{\log \left((e^{(e^{\log_* (1 + \log_* (1 + \cos^{-1} \left((4 \cdot \left((v \cdot v + \left({v}^{4}\right))_*\right) + -1)_*\right)))} - 1)^*} - 1)^*\right)}\]

Runtime

Time bar (total: 26.2s)Debug logProfile

BaselineHerbieOracleSpan%
Regimes0.80.80.60.20%
herbie shell --seed 2018354 +o rules:numerics
(FPCore (v)
  :name "Falkner and Boettcher, Appendix B, 1"
  (acos (/ (- 1 (* 5 (* v v))) (- (* v v) 1))))