time-to-botec

Benchmark sampling in different programming languages
Log | Files | Refs | README

repl.txt (2073B)

      1 
      2 {{alias}}( x, y[, scale] )
      3     Computes the relative difference of two real numbers in units of double-
      4     precision floating-point epsilon.
      5 
      6     By default, the function scales the absolute difference by dividing the
      7     absolute difference by the maximum absolute value of `x` and `y`. To scale
      8     by a different function, specify a scale function name.
      9 
     10     The following `scale` functions are supported:
     11 
     12     - 'max-abs': maximum absolute value of `x` and `y` (default).
     13     - 'max': maximum value of `x` and `y`.
     14     - 'min-abs': minimum absolute value of `x` and `y`.
     15     - 'min': minimum value of `x` and `y`.
     16     - 'mean-abs': arithmetic mean of the absolute values of `x` and `y`.
     17     - 'mean': arithmetic mean of `x` and `y`.
     18     - 'x': `x` (*noncommutative*).
     19     - 'y': `y` (*noncommutative*).
     20 
     21     To use a custom scale function, provide a function which accepts two numeric
     22     arguments `x` and `y`.
     23 
     24     If computing the relative difference in units of epsilon will result in
     25     overflow, the function returns the maximum double-precision floating-point
     26     number.
     27 
     28     If the absolute difference of `x` and `y` is `0`, the relative difference is
     29     always `0`.
     30 
     31     If `|x| = |y| = infinity`, the function returns `NaN`.
     32 
     33     If `|x| = |-y| = infinity`, the relative difference is `+infinity`.
     34 
     35     If a `scale` function returns `0`, the function returns `NaN`.
     36 
     37     Parameters
     38     ----------
     39     x: number
     40         First number.
     41 
     42     y: number
     43         Second number.
     44 
     45     scale: string|Function (optional)
     46         Scale function. Default: `'max-abs'`.
     47 
     48     Returns
     49     -------
     50     out: number
     51         Relative difference in units of double-precision floating-point epsilon.
     52 
     53     Examples
     54     --------
     55     > var d = {{alias}}( 12.15, 12.149999999999999 )
     56     ~0.658
     57     > d = {{alias}}( 2.4341309458983933, 2.4341309458633909, 'mean-abs' )
     58     ~64761.512
     59 
     60     // Custom scale function:
     61     > function scale( x, y ) { return ( x > y ) ? y : x; };
     62     > d = {{alias}}( 1.0000000000000002, 1.0000000000000100, scale )
     63     ~44
     64 
     65     See Also
     66     --------
     67