Expression 1, p15

Average Error: 0.4 → 0.3

Time: 3.3s

Precision: 64

\[1 \le a \le 2 \le b \le 4 \le c \le 8 \le d \le 16 \le e \le 32\]

Math

\[\left(\left(\left(e + d\right) + c\right) + b\right) + a\]

↓

\[\left(d + \left(e + c\right)\right) + \left(b + a\right)\]

C

double f(double a, double b, double c, double d, double e) {
        double r96324 = e;
        double r96325 = d;
        double r96326 = r96324 + r96325;
        double r96327 = c;
        double r96328 = r96326 + r96327;
        double r96329 = b;
        double r96330 = r96328 + r96329;
        double r96331 = a;
        double r96332 = r96330 + r96331;
        return r96332;
}

↓

double f(double a, double b, double c, double d, double e) {
        double r96333 = d;
        double r96334 = e;
        double r96335 = c;
        double r96336 = r96334 + r96335;
        double r96337 = r96333 + r96336;
        double r96338 = b;
        double r96339 = a;
        double r96340 = r96338 + r96339;
        double r96341 = r96337 + r96340;
        return r96341;
}

Error

Bits error versus a

Bits error versus b

Bits error versus c

Bits error versus d

Bits error versus e

Target

Original	0.4
Target	0.2
Herbie	0.3

\[\left(d + \left(c + \left(a + b\right)\right)\right) + e\]

Derivation

1. Initial program 0.4

   \[\left(\left(\left(e + d\right) + c\right) + b\right) + a\]

2. Taylor expanded around 0 0.4

   \[\leadsto \left(\color{blue}{\left(d + \left(e + c\right)\right)} + b\right) + a\]
3. Using strategy rm

4. Applied associate-+l+ 0.3

   \[\leadsto \color{blue}{\left(d + \left(e + c\right)\right) + \left(b + a\right)}\]

5. Final simplification 0.3

   \[\leadsto \left(d + \left(e + c\right)\right) + \left(b + a\right)\]

Reproduce

herbie shell --seed 2020039 +o rules:numerics
(FPCore (a b c d e)
  :name "Expression 1, p15"
  :precision binary64
  :pre (<= 1 a 2 b 4 c 8 d 16 e 32)

  :herbie-target
  (+ (+ d (+ c (+ a b))) e)

  (+ (+ (+ (+ e d) c) b) a))