ReportTypes:explainFloat from gipeda-0.1.2.1

Average Error: 0 → 0

Time: 2.7s

Precision: binary64

Cost: 320

Math

\[100 \cdot \frac{x - x}{x} \]

↓

\[100 \cdot \frac{0}{x} \]

FPCore

(FPCore (x) :precision binary64 (* 100.0 (/ (- x x) x)))

↓

(FPCore (x) :precision binary64 (* 100.0 (/ 0.0 x)))

C

double code(double x) {
	return 100.0 * ((x - x) / x);
}

↓

double code(double x) {
	return 100.0 * (0.0 / x);
}

Fortran

real(8) function code(x)
    real(8), intent (in) :: x
    code = 100.0d0 * ((x - x) / x)
end function

↓

real(8) function code(x)
    real(8), intent (in) :: x
    code = 100.0d0 * (0.0d0 / x)
end function

Java

public static double code(double x) {
	return 100.0 * ((x - x) / x);
}

↓

public static double code(double x) {
	return 100.0 * (0.0 / x);
}

Python

def code(x):
	return 100.0 * ((x - x) / x)

↓

def code(x):
	return 100.0 * (0.0 / x)

Julia

function code(x)
	return Float64(100.0 * Float64(Float64(x - x) / x))
end

↓

function code(x)
	return Float64(100.0 * Float64(0.0 / x))
end

MATLAB

function tmp = code(x)
	tmp = 100.0 * ((x - x) / x);
end

↓

function tmp = code(x)
	tmp = 100.0 * (0.0 / x);
end

Wolfram

code[x_] := N[(100.0 * N[(N[(x - x), $MachinePrecision] / x), $MachinePrecision]), $MachinePrecision]

↓

code[x_] := N[(100.0 * N[(0.0 / x), $MachinePrecision]), $MachinePrecision]

Target

Original	0
Target	0
Herbie	0

\[0 \]

Derivation

1. Initial program 0

   \[100 \cdot \frac{x - x}{x} \]

2. Simplified 0

   \[\leadsto \color{blue}{100 \cdot \frac{0}{x}} \]
   Proof

   [Start] 0	\[ 100 \cdot \frac{x - x}{x} \]
   rational.json-simplify-3 [=>] 0	\[ 100 \cdot \frac{\color{blue}{0}}{x} \]

3. Final simplification 0

   \[\leadsto 100 \cdot \frac{0}{x} \]

Reproduce

herbie shell --seed 2023077 
(FPCore (x)
  :name "ReportTypes:explainFloat from gipeda-0.1.2.1"
  :precision binary64

  :herbie-target
  0.0

  (* 100.0 (/ (- x x) x)))