[R] question about precision, floor, and powers of two.

[Patrick Connolly]

On Fri, 04-Nov-2005 at 04:58PM +0100, Peter Dalgaard wrote:


|> In this particular case, it is slightly odd that we can't get an exact
|> answer for operations that could in principle be carried out using
|> integer arithmetic, but we're actually calculating log(8)/log(2).
|> 
|> (Curiously, the same effect is not seen on Linux or Solaris until 
|> 
|>  > log2(2^29)-29
|> [1] 3.552714e-15

"Until" is not quite the word:

> x <- 2:100
> data.frame(x, D = log2(2^x) -x)
     x            D
1    2 0.000000e+00
2    3 0.000000e+00
3    4 0.000000e+00
4    5 0.000000e+00
5    6 0.000000e+00
6    7 0.000000e+00
7    8 0.000000e+00
8    9 0.000000e+00
9   10 0.000000e+00
10  11 0.000000e+00
11  12 0.000000e+00
12  13 0.000000e+00
13  14 0.000000e+00
14  15 0.000000e+00
15  16 0.000000e+00
16  17 0.000000e+00
17  18 0.000000e+00
18  19 0.000000e+00
19  20 0.000000e+00
20  21 0.000000e+00
21  22 0.000000e+00
22  23 0.000000e+00
23  24 0.000000e+00
24  25 0.000000e+00
25  26 0.000000e+00
26  27 0.000000e+00
27  28 0.000000e+00
28  29 3.552714e-15
29  30 0.000000e+00
30  31 3.552714e-15
31  32 0.000000e+00
32  33 0.000000e+00
33  34 0.000000e+00
34  35 0.000000e+00
35  36 0.000000e+00
36  37 0.000000e+00
37  38 0.000000e+00
38  39 7.105427e-15
39  40 0.000000e+00
40  41 0.000000e+00
41  42 0.000000e+00
42  43 0.000000e+00
43  44 0.000000e+00
44  45 0.000000e+00
45  46 0.000000e+00
46  47 7.105427e-15
47  48 0.000000e+00
48  49 0.000000e+00
49  50 0.000000e+00
50  51 7.105427e-15
51  52 0.000000e+00
52  53 0.000000e+00
53  54 0.000000e+00
54  55 7.105427e-15
55  56 0.000000e+00
56  57 0.000000e+00
57  58 7.105427e-15
58  59 7.105427e-15
59  60 0.000000e+00
60  61 0.000000e+00
61  62 7.105427e-15
62  63 0.000000e+00
63  64 0.000000e+00
64  65 0.000000e+00
65  66 0.000000e+00
66  67 0.000000e+00
67  68 0.000000e+00
68  69 0.000000e+00
69  70 0.000000e+00
70  71 0.000000e+00
71  72 0.000000e+00
72  73 0.000000e+00
73  74 0.000000e+00
74  75 0.000000e+00
75  76 0.000000e+00
76  77 0.000000e+00
77  78 1.421085e-14
78  79 0.000000e+00
79  80 0.000000e+00
80  81 0.000000e+00
81  82 0.000000e+00
82  83 0.000000e+00
83  84 0.000000e+00
84  85 0.000000e+00
85  86 0.000000e+00
86  87 0.000000e+00
87  88 0.000000e+00
88  89 0.000000e+00
89  90 0.000000e+00
90  91 0.000000e+00
91  92 0.000000e+00
92  93 1.421085e-14
93  94 1.421085e-14
94  95 1.421085e-14
95  96 0.000000e+00
96  97 0.000000e+00
97  98 0.000000e+00
98  99 0.000000e+00
99 100 0.000000e+00

Is there a pattern?


-- 
Patrick Connolly
HortResearch
Mt Albert
Auckland
New Zealand 
Ph: +64-9 815 4200 x 7188
~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~
I have the world`s largest collection of seashells. I keep it on all
the beaches of the world ... Perhaps you`ve seen it.  ---Steven Wright 
~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~