See also  http://www.jsoftware.com/jwiki/Essays/Sorting_versus_Grading
As well, more benchmarks should be done before drawing
firm conclusions about relative efficiencies.  The data used so
far is rather special and makes possible algorithms not available
in more general cases. e.g.

   ts=: 6!:2 , 7!:[EMAIL PROTECTED]
   v0=: (*:1024) [EMAIL PROTECTED] 1024
   v1=: (*:1024) [EMAIL PROTECTED] 2e9
   v2=: (*:1024) [EMAIL PROTECTED] *:1024
   ts '/:~ v0'
0.0243629 8.39757e6
   ts '/:~ v1'
0.49494 1.73023e7
   ts '/:~ v2'
0.229462 1.6778e7



----- Original Message -----
From: "Sherlock, Ric" <[EMAIL PROTECTED]>
Date: Friday, March 14, 2008 2:55
Subject: [Jchat] RE: Alternate Quick Sort
To: Chat forum <[email protected]>

> ---Alex Rufon wrote:
> > So I thought about how I actually do sorting
> > which is this way:
> > $v=: (1024 * 1024) [EMAIL PROTECTED]
> ...
> >    timespacex 'v{~/:v'
> > 0.0476471870028174 25166528
> 
> Simpler, faster and leaner.
> 
>    ts 'v{~/:v'
> 0.0769561494547 16777920
>    ts '/:~v'
> 0.0175374942905 8397504
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to