Linear.Quaternion:$c/ from linear-1.19.1.3, A

Average Accuracy: 99.8% → 99.9%

Time: 8.1s

Precision: binary64

Cost: 7104

Math

\[\left(\left(x \cdot y + z \cdot z\right) + z \cdot z\right) + z \cdot z \]

↓

\[\mathsf{fma}\left(z, z, x \cdot y + 2 \cdot \left(z \cdot z\right)\right) \]

FPCore

(FPCore (x y z)
 :precision binary64
 (+ (+ (+ (* x y) (* z z)) (* z z)) (* z z)))

↓

(FPCore (x y z) :precision binary64 (fma z z (+ (* x y) (* 2.0 (* z z)))))

C

double code(double x, double y, double z) {
	return (((x * y) + (z * z)) + (z * z)) + (z * z);
}

↓

double code(double x, double y, double z) {
	return fma(z, z, ((x * y) + (2.0 * (z * z))));
}

Julia

function code(x, y, z)
	return Float64(Float64(Float64(Float64(x * y) + Float64(z * z)) + Float64(z * z)) + Float64(z * z))
end

↓

function code(x, y, z)
	return fma(z, z, Float64(Float64(x * y) + Float64(2.0 * Float64(z * z))))
end

Wolfram

code[x_, y_, z_] := N[(N[(N[(N[(x * y), $MachinePrecision] + N[(z * z), $MachinePrecision]), $MachinePrecision] + N[(z * z), $MachinePrecision]), $MachinePrecision] + N[(z * z), $MachinePrecision]), $MachinePrecision]

↓

code[x_, y_, z_] := N[(z * z + N[(N[(x * y), $MachinePrecision] + N[(2.0 * N[(z * z), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]), $MachinePrecision]

Target

Original	99.8%
Target	99.8%
Herbie	99.9%

\[\left(3 \cdot z\right) \cdot z + y \cdot x \]

Derivation

1. Initial program 99.8%

   \[\left(\left(x \cdot y + z \cdot z\right) + z \cdot z\right) + z \cdot z \]

2. Simplified 99.9%

   \[\leadsto \color{blue}{\mathsf{fma}\left(z, z, \mathsf{fma}\left(x, y, 2 \cdot \left(z \cdot z\right)\right)\right)} \]
   Proof

   [Start] 99.8	\[ \left(\left(x \cdot y + z \cdot z\right) + z \cdot z\right) + z \cdot z \]
   +-commutative [=>] 99.8	\[ \color{blue}{z \cdot z + \left(\left(x \cdot y + z \cdot z\right) + z \cdot z\right)} \]
   fma-def [=>] 99.9	\[ \color{blue}{\mathsf{fma}\left(z, z, \left(x \cdot y + z \cdot z\right) + z \cdot z\right)} \]
   associate-+l+ [=>] 99.9	\[ \mathsf{fma}\left(z, z, \color{blue}{x \cdot y + \left(z \cdot z + z \cdot z\right)}\right) \]
   fma-def [=>] 99.9	\[ \mathsf{fma}\left(z, z, \color{blue}{\mathsf{fma}\left(x, y, z \cdot z + z \cdot z\right)}\right) \]
   count-2 [=>] 99.9	\[ \mathsf{fma}\left(z, z, \mathsf{fma}\left(x, y, \color{blue}{2 \cdot \left(z \cdot z\right)}\right)\right) \]

3. Applied egg-rr 99.9%

   \[\leadsto \mathsf{fma}\left(z, z, \color{blue}{x \cdot y + 2 \cdot \left(z \cdot z\right)}\right) \]
   Proof

   [Start] 99.9	\[ \mathsf{fma}\left(z, z, \mathsf{fma}\left(x, y, 2 \cdot \left(z \cdot z\right)\right)\right) \]
   fma-udef [=>] 99.9	\[ \mathsf{fma}\left(z, z, \color{blue}{x \cdot y + 2 \cdot \left(z \cdot z\right)}\right) \]

4. Final simplification 99.9%

   \[\leadsto \mathsf{fma}\left(z, z, x \cdot y + 2 \cdot \left(z \cdot z\right)\right) \]

Alternatives

Alternative 1
Accuracy	82.8%
Cost	585

\[\begin{array}{l} \mathbf{if}\;z \leq -5 \cdot 10^{-33} \lor \neg \left(z \leq 2.5 \cdot 10^{-19}\right):\\ \;\;\;\;\left(z \cdot z\right) \cdot 3\\ \mathbf{else}:\\ \;\;\;\;x \cdot y\\ \end{array} \]

Alternative 2
Accuracy	99.8%
Cost	576

\[x \cdot y + \left(z \cdot z\right) \cdot 3 \]

Alternative 3
Accuracy	62.5%
Cost	192

\[x \cdot y \]

Reproduce

herbie shell --seed 2023137 
(FPCore (x y z)
  :name "Linear.Quaternion:$c/ from linear-1.19.1.3, A"
  :precision binary64

  :herbie-target
  (+ (* (* 3.0 z) z) (* y x))

  (+ (+ (+ (* x y) (* z z)) (* z z)) (* z z)))