speed of statistics.m

[David Bateman]

Francesco Potorti` wrote:
>>> contrary to my expectations, statistics is much slower in Octave 3.1:
>>>
>>> octave3.0> t=cputime;statistics(1:35000);cputime-t
>>> ans =  0.26002
>>>
>>> octave3.1> t=cputime;statistics(1:35000);cputime-t
>>> ans =  1.6521
>>>       
>> `statistics' computes a number of characteristics. Could you identify
>> which one of them is causing the slowdown?
>>     
>
> Sorry, bad report :(
>
> It turns out that while the nan package does not significantly affect
> the speed of statistics() in 3.0, it severely slows them down in 3.1.
>
> In fact, without the nan package, 3.1 is much faster, as I expected by
> looking at the code.
>   

Why the nan package should make things much slower is of course a 
different bug report, and would be interesting to get to the bottom of.

D.


-- 
David Bateman                                David.Bateman at motorola.com
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 6 72 01 06 33 (Mob) 
91193 Gif-Sur-Yvette FRANCE                  +33 1 69 35 77 01 (Fax) 

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary