Average Error: 5.3 → 1.6
Time: 7.2m
Precision: 64
Internal Precision: 384
\[\left(\left(\left(\left(\left(\left(x \cdot 18.0\right) \cdot y\right) \cdot z\right) \cdot t - \left(a \cdot 4.0\right) \cdot t\right) + b \cdot c\right) - \left(x \cdot 4.0\right) \cdot i\right) - \left(j \cdot 27.0\right) \cdot k\]
↓
\[\begin{array}{l} \mathbf{if}\;t \le -0.0010632815105304718:\\ \;\;\;\;\left(b \cdot c - 4.0 \cdot \left(x \cdot i + a \cdot t\right)\right) + \left(t \cdot \left(\left(\left(18.0 \cdot x\right) \cdot y\right) \cdot z\right) - 27.0 \cdot \left(k \cdot j\right)\right)\\ \mathbf{if}\;t \le 2.979751374511743 \cdot 10^{-48}:\\ \;\;\;\;\left(b \cdot c - 4.0 \cdot \left(x \cdot i + a \cdot t\right)\right) + \left(18.0 \cdot \left(z \cdot \left(\left(y \cdot t\right) \cdot x\right)\right) - 27.0 \cdot \left(k \cdot j\right)\right)\\ \mathbf{else}:\\ \;\;\;\;\left(b \cdot c - 4.0 \cdot \left(x \cdot i + a \cdot t\right)\right) + \left(t \cdot \left(\left(\left(18.0 \cdot x\right) \cdot y\right) \cdot z\right) - 27.0 \cdot \left(k \cdot j\right)\right)\\ \end{array}\]

Error

Bits error versus x

Bits error versus y

Bits error versus z

Bits error versus t

Bits error versus a

Bits error versus b

Bits error versus c

Bits error versus i

Bits error versus j

Bits error versus k

Derivation

  1. Split input into 2 regimes

  2. if t < -0.0010632815105304718 or 2.979751374511743e-48 < t

    1. Initial program 2.0

      \[\left(\left(\left(\left(\left(\left(x \cdot 18.0\right) \cdot y\right) \cdot z\right) \cdot t - \left(a \cdot 4.0\right) \cdot t\right) + b \cdot c\right) - \left(x \cdot 4.0\right) \cdot i\right) - \left(j \cdot 27.0\right) \cdot k\]

    2. Applied simplify2.2

      \[\leadsto \color{blue}{\left(b \cdot c - 4.0 \cdot \left(x \cdot i + a \cdot t\right)\right) + \left(t \cdot \left(\left(18.0 \cdot x\right) \cdot \left(y \cdot z\right)\right) - 27.0 \cdot \left(k \cdot j\right)\right)}\]
    3. Using strategy rm

    4. Applied associate-*r*2.0

      \[\leadsto \left(b \cdot c - 4.0 \cdot \left(x \cdot i + a \cdot t\right)\right) + \left(t \cdot \color{blue}{\left(\left(\left(18.0 \cdot x\right) \cdot y\right) \cdot z\right)} - 27.0 \cdot \left(k \cdot j\right)\right)\]

    if -0.0010632815105304718 < t < 2.979751374511743e-48

    1. Initial program 7.7

      \[\left(\left(\left(\left(\left(\left(x \cdot 18.0\right) \cdot y\right) \cdot z\right) \cdot t - \left(a \cdot 4.0\right) \cdot t\right) + b \cdot c\right) - \left(x \cdot 4.0\right) \cdot i\right) - \left(j \cdot 27.0\right) \cdot k\]

    2. Applied simplify8.7

      \[\leadsto \color{blue}{\left(b \cdot c - 4.0 \cdot \left(x \cdot i + a \cdot t\right)\right) + \left(t \cdot \left(\left(18.0 \cdot x\right) \cdot \left(y \cdot z\right)\right) - 27.0 \cdot \left(k \cdot j\right)\right)}\]

    3. Taylor expanded around inf 1.9

      \[\leadsto \left(b \cdot c - 4.0 \cdot \left(x \cdot i + a \cdot t\right)\right) + \left(\color{blue}{18.0 \cdot \left(z \cdot \left(y \cdot \left(t \cdot x\right)\right)\right)} - 27.0 \cdot \left(k \cdot j\right)\right)\]
    4. Using strategy rm

    5. Applied associate-*r*1.3

      \[\leadsto \left(b \cdot c - 4.0 \cdot \left(x \cdot i + a \cdot t\right)\right) + \left(18.0 \cdot \left(z \cdot \color{blue}{\left(\left(y \cdot t\right) \cdot x\right)}\right) - 27.0 \cdot \left(k \cdot j\right)\right)\]
  3. Recombined 2 regimes into one program.
  4. Removed slow pow expressions.

Runtime

Time bar (total: 7.2m)Debug logProfile

herbie shell --seed '#(1062930989 876886121 3990119081 3032829768 3060892583 1929069376)' 
(FPCore (x y z t a b c i j k)
  :name "Diagrams.Solve.Polynomial:cubForm  from diagrams-solve-0.1"
  (- (- (+ (- (* (* (* (* x 18.0) y) z) t) (* (* a 4.0) t)) (* b c)) (* (* x 4.0) i)) (* (* j 27.0) k)))