Average Error: 0.1 → 0.1
Time: 1.0s
Precision: binary64
\[\frac{d}{a} - \frac{0.333333343000000004 \cdot \frac{bc}{a}}{a}\]
↓
\[\frac{d}{a} - \frac{0.333333343000000004 \cdot \frac{bc}{a}}{a}\]
\frac{d}{a} - \frac{0.333333343000000004 \cdot \frac{bc}{a}}{a}
↓
\frac{d}{a} - \frac{0.333333343000000004 \cdot \frac{bc}{a}}{a}
double code(double d, double a, double bc) {
	return ((double) (((double) (d / a)) - ((double) (((double) (0.333333343 * ((double) (bc / a)))) / a))));
}
↓
double code(double d, double a, double bc) {
	return ((double) (((double) (d / a)) - ((double) (((double) (0.333333343 * ((double) (bc / a)))) / a))));
}

Error

Bits error versus d

Bits error versus a

Bits error versus bc

Try it out

Your Program's Arguments

Results

Enter valid numbers for all inputs

Derivation

  1. Initial program 0.1

    \[\frac{d}{a} - \frac{0.333333343000000004 \cdot \frac{bc}{a}}{a}\]

  2. Final simplification0.1

    \[\leadsto \frac{d}{a} - \frac{0.333333343000000004 \cdot \frac{bc}{a}}{a}\]

Reproduce

herbie shell --seed 2020153 
(FPCore (d a bc)
  :name "(- (/ d a) (/ (* 0.333333343 (/ bc a)) a))"
  :precision binary64
  (- (/ d a) (/ (* 0.333333343 (/ bc a)) a)))