It's because of

 p=. 1,s$0x

It looks like the amend is not in-place when exact arithmetic
is used.

Henry rich 

> -----Original Message-----
> From: [EMAIL PROTECTED] 
> [mailto:[EMAIL PROTECTED] On Behalf Of Boyko Bantchev
> Sent: Monday, September 18, 2006 12:15 PM
> To: Programming forum
> Subject: Re: [Jprogramming] linearity or ...?
> 
> On 9/18/06, Pascal Jasmin <[EMAIL PROTECTED]> wrote:
> > I don't understand what pnkd does ........
> 
> It counts partitions, as described on the cited wiki page.
> But that is unrelated to the question I asked, which is:
> `why pnkd appears, according to the measurement, to be
> non-linear in x when y is 1'?
> The outer loop has count 1, so the function should be
> as fast as its inner loop, and that must be linear in the size
> of t which is x.  Or am I wrong?
> Please someone comment on this.
> 
> Regards,
>    Boyko
> ----------------------------------------------------------------------
> For information about J forums see 
> http://www.jsoftware.com/forums.htm

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to