That makes sense.

Thanks,

-- 
Raul

On Sun, Oct 9, 2022 at 7:17 PM 'Michael Day' via General
<[email protected]> wrote:
>
> I don't know about the time for {.a,  but this modification helps a bit
>
> earlier on in the brute force process:
>
>     all=: {{ A#~x=".@>A=.   ~.   ,arith^:5 ":each y}}
>
> Putting in a few diagnostics,  I found in one run:
>
>     task ''
>
> terms: 4 7 7 8 9 10
> target: 315
> ┌─────┬──────┬──────┐
> │size │934358│339567│
> └─────┴──────┴──────┘
> #solutions: 428
> for example: ((7+(4+7))*(8+10))-9
>
> ... where size reports the dimension of A before and after taking its
> nub,  ~.
>
> I'm afraid I didn't time that one,  but taking the nub appears to have
> saved
>
> 600 000 odd calculations out of 900 000 odd,  as well as 600 000 odd
> lines/boxes
>
> of calculands.
>
> Cheers,
>
> Mike
>
>
> On 09/10/2022 18:14, Raul Miller wrote:
> > https://rosettacode.org/wiki/Countdown#J
> >
> >     task''
> > terms: 4 7 8 9 10 10
> > target: 300
> > #solutions: 1213
> > for example: (10+(9*10))*((4+7)-8)
> >
> > The #solutions line takes approaching 20 seconds. That's not the problem 
> > here.
> >
> > The 'for example' line takes several seconds in J903 (and J904). And
> > the expression is
> >     echo 'for example: ',;{.a
> >
> > The only thing I can think of is that this time is occupied by garbage
> > collection.
> >
> > FYI,
> >
> ----------------------------------------------------------------------
> 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