2sin (example 3.3)

?

Percentage Accurate: 42.5% → 99.4%
Time: 13.1s
Precision: binary64
Cost: 45184

?

\[\sin \left(x + \varepsilon\right) - \sin x \]
↓
\[\mathsf{fma}\left(\sin \varepsilon, \cos x, \mathsf{fma}\left(\sin x, \cos \varepsilon, -\sin x\right)\right) \]
(FPCore (x eps) :precision binary64 (- (sin (+ x eps)) (sin x)))
↓
(FPCore (x eps)
 :precision binary64
 (fma (sin eps) (cos x) (fma (sin x) (cos eps) (- (sin x)))))
double code(double x, double eps) {
	return sin((x + eps)) - sin(x);
}
↓
double code(double x, double eps) {
	return fma(sin(eps), cos(x), fma(sin(x), cos(eps), -sin(x)));
}

function code(x, eps)
	return Float64(sin(Float64(x + eps)) - sin(x))
end
↓
function code(x, eps)
	return fma(sin(eps), cos(x), fma(sin(x), cos(eps), Float64(-sin(x))))
end

code[x_, eps_] := N[(N[Sin[N[(x + eps), $MachinePrecision]], $MachinePrecision] - N[Sin[x], $MachinePrecision]), $MachinePrecision]
↓
code[x_, eps_] := N[(N[Sin[eps], $MachinePrecision] * N[Cos[x], $MachinePrecision] + N[(N[Sin[x], $MachinePrecision] * N[Cos[eps], $MachinePrecision] + (-N[Sin[x], $MachinePrecision])), $MachinePrecision]), $MachinePrecision]

\sin \left(x + \varepsilon\right) - \sin x
↓
\mathsf{fma}\left(\sin \varepsilon, \cos x, \mathsf{fma}\left(\sin x, \cos \varepsilon, -\sin x\right)\right)

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.

Herbie found 11 alternatives:

AlternativeAccuracySpeedup

Accuracy vs Speed

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.

Bogosity?

Bogosity

Target

Original42.5%
Target76.9%
Herbie99.4%
\[2 \cdot \left(\cos \left(x + \frac{\varepsilon}{2}\right) \cdot \sin \left(\frac{\varepsilon}{2}\right)\right) \]

Derivation?

  1. Initial program 41.6%

    \[\sin \left(x + \varepsilon\right) - \sin x \]

  2. Applied egg-rr67.0%

    \[\leadsto \color{blue}{\sin x \cdot \cos \varepsilon + \left(\cos x \cdot \sin \varepsilon - \sin x\right)} \]
    Step-by-step derivation

    [Start]41.6%

    \[ \sin \left(x + \varepsilon\right) - \sin x \]

    sin-sum [=>]67.0%

    \[ \color{blue}{\left(\sin x \cdot \cos \varepsilon + \cos x \cdot \sin \varepsilon\right)} - \sin x \]

    associate--l+ [=>]67.0%

    \[ \color{blue}{\sin x \cdot \cos \varepsilon + \left(\cos x \cdot \sin \varepsilon - \sin x\right)} \]

  3. Simplified99.4%

    \[\leadsto \color{blue}{\sin \varepsilon \cdot \cos x + \sin x \cdot \left(\cos \varepsilon + -1\right)} \]
    Step-by-step derivation

    [Start]67.0%

    \[ \sin x \cdot \cos \varepsilon + \left(\cos x \cdot \sin \varepsilon - \sin x\right) \]

    +-commutative [=>]67.0%

    \[ \color{blue}{\left(\cos x \cdot \sin \varepsilon - \sin x\right) + \sin x \cdot \cos \varepsilon} \]

    sub-neg [=>]67.0%

    \[ \color{blue}{\left(\cos x \cdot \sin \varepsilon + \left(-\sin x\right)\right)} + \sin x \cdot \cos \varepsilon \]

    associate-+l+ [=>]99.4%

    \[ \color{blue}{\cos x \cdot \sin \varepsilon + \left(\left(-\sin x\right) + \sin x \cdot \cos \varepsilon\right)} \]

    *-commutative [=>]99.4%

    \[ \color{blue}{\sin \varepsilon \cdot \cos x} + \left(\left(-\sin x\right) + \sin x \cdot \cos \varepsilon\right) \]

    neg-mul-1 [=>]99.4%

    \[ \sin \varepsilon \cdot \cos x + \left(\color{blue}{-1 \cdot \sin x} + \sin x \cdot \cos \varepsilon\right) \]

    *-commutative [=>]99.4%

    \[ \sin \varepsilon \cdot \cos x + \left(-1 \cdot \sin x + \color{blue}{\cos \varepsilon \cdot \sin x}\right) \]

    distribute-rgt-out [=>]99.4%

    \[ \sin \varepsilon \cdot \cos x + \color{blue}{\sin x \cdot \left(-1 + \cos \varepsilon\right)} \]

    +-commutative [<=]99.4%

    \[ \sin \varepsilon \cdot \cos x + \sin x \cdot \color{blue}{\left(\cos \varepsilon + -1\right)} \]

  4. Applied egg-rr99.4%

    \[\leadsto \color{blue}{\mathsf{fma}\left(\sin \varepsilon, \cos x, \left(\cos \varepsilon + -1\right) \cdot \sin x\right)} \]
    Step-by-step derivation

    [Start]99.4%

    \[ \sin \varepsilon \cdot \cos x + \sin x \cdot \left(\cos \varepsilon + -1\right) \]

    fma-def [=>]99.4%

    \[ \color{blue}{\mathsf{fma}\left(\sin \varepsilon, \cos x, \sin x \cdot \left(\cos \varepsilon + -1\right)\right)} \]

    *-commutative [=>]99.4%

    \[ \mathsf{fma}\left(\sin \varepsilon, \cos x, \color{blue}{\left(\cos \varepsilon + -1\right) \cdot \sin x}\right) \]

  5. Applied egg-rr99.4%

    \[\leadsto \mathsf{fma}\left(\sin \varepsilon, \cos x, \color{blue}{\mathsf{fma}\left(\sin x, \cos \varepsilon, -\sin x\right)}\right) \]
    Step-by-step derivation

    [Start]99.4%

    \[ \mathsf{fma}\left(\sin \varepsilon, \cos x, \left(\cos \varepsilon + -1\right) \cdot \sin x\right) \]

    *-commutative [=>]99.4%

    \[ \mathsf{fma}\left(\sin \varepsilon, \cos x, \color{blue}{\sin x \cdot \left(\cos \varepsilon + -1\right)}\right) \]

    distribute-rgt-in [=>]99.4%

    \[ \mathsf{fma}\left(\sin \varepsilon, \cos x, \color{blue}{\cos \varepsilon \cdot \sin x + -1 \cdot \sin x}\right) \]

    neg-mul-1 [<=]99.4%

    \[ \mathsf{fma}\left(\sin \varepsilon, \cos x, \cos \varepsilon \cdot \sin x + \color{blue}{\left(-\sin x\right)}\right) \]

    sub-neg [<=]99.4%

    \[ \mathsf{fma}\left(\sin \varepsilon, \cos x, \color{blue}{\cos \varepsilon \cdot \sin x - \sin x}\right) \]

    *-commutative [=>]99.4%

    \[ \mathsf{fma}\left(\sin \varepsilon, \cos x, \color{blue}{\sin x \cdot \cos \varepsilon} - \sin x\right) \]

    fma-neg [=>]99.4%

    \[ \mathsf{fma}\left(\sin \varepsilon, \cos x, \color{blue}{\mathsf{fma}\left(\sin x, \cos \varepsilon, -\sin x\right)}\right) \]

  6. Final simplification99.4%

    \[\leadsto \mathsf{fma}\left(\sin \varepsilon, \cos x, \mathsf{fma}\left(\sin x, \cos \varepsilon, -\sin x\right)\right) \]

Alternatives

Alternative 1
Accuracy99.4%
Cost45184
\[\mathsf{fma}\left(\sin \varepsilon, \cos x, \mathsf{fma}\left(\sin x, \cos \varepsilon, -\sin x\right)\right) \]
Alternative 2
Accuracy99.4%
Cost32448
\[\mathsf{fma}\left(\sin \varepsilon, \cos x, \sin x \cdot \left(\cos \varepsilon + -1\right)\right) \]
Alternative 3
Accuracy99.4%
Cost32448
\[\mathsf{fma}\left(\cos \varepsilon + -1, \sin x, \sin \varepsilon \cdot \cos x\right) \]
Alternative 4
Accuracy99.4%
Cost26176
\[\sin \varepsilon \cdot \cos x + \sin x \cdot \left(\cos \varepsilon + -1\right) \]
Alternative 5
Accuracy78.1%
Cost25920
\[\mathsf{fma}\left(0, \sin x, \sin \varepsilon \cdot \cos x\right) \]
Alternative 6
Accuracy77.4%
Cost13769
\[\begin{array}{l} \mathbf{if}\;\varepsilon \leq -0.0012 \lor \neg \left(\varepsilon \leq 235000000000\right):\\ \;\;\;\;\sin \varepsilon - \sin x\\ \mathbf{else}:\\ \;\;\;\;-0.5 \cdot \left(\sin x \cdot \left(\varepsilon \cdot \varepsilon\right)\right) + \varepsilon \cdot \cos x\\ \end{array} \]
Alternative 7
Accuracy76.8%
Cost13632
\[\cos \left(0.5 \cdot \left(\varepsilon + \left(x + x\right)\right)\right) \cdot \left(2 \cdot \sin \left(\varepsilon \cdot 0.5\right)\right) \]
Alternative 8
Accuracy77.5%
Cost13257
\[\begin{array}{l} \mathbf{if}\;\varepsilon \leq -1.2 \cdot 10^{-5} \lor \neg \left(\varepsilon \leq 0.18\right):\\ \;\;\;\;\sin \varepsilon - \sin x\\ \mathbf{else}:\\ \;\;\;\;\varepsilon \cdot \cos x\\ \end{array} \]
Alternative 9
Accuracy76.8%
Cost6856
\[\begin{array}{l} \mathbf{if}\;\varepsilon \leq -9.5 \cdot 10^{-5}:\\ \;\;\;\;\sin \varepsilon\\ \mathbf{elif}\;\varepsilon \leq 0.18:\\ \;\;\;\;\varepsilon \cdot \cos x\\ \mathbf{else}:\\ \;\;\;\;\sin \varepsilon\\ \end{array} \]
Alternative 10
Accuracy55.3%
Cost6464
\[\sin \varepsilon \]
Alternative 11
Accuracy29.3%
Cost64
\[\varepsilon \]

Reproduce?

herbie shell --seed 2023271 
(FPCore (x eps)
  :name "2sin (example 3.3)"
  :precision binary64

  :herbie-target
  (* 2.0 (* (cos (+ x (/ eps 2.0))) (sin (/ eps 2.0))))

  (- (sin (+ x eps)) (sin x)))