A small programming challenge

[Geoff Canyon]

I worked on this further over the weekend. I'll post again after I've
formatted everything, but the result now handles any number of cubes from 3
to 26, although anything more than about 14 will probably be too slow. For
4 cube puzzles the new code is about four times faster than my old code.
For 12 cube puzzles it's >100 times faster. I came up with a nice short cut.


On Sun, Apr 14, 2013 at 4:58 PM, Alex Tweedly <alex at tweedly.net> wrote:

>
> Yes, a few interesting aspects.
>
> I too was surprised that the (intuitive) array method was slower than the
> string-based method, so I looked into it some more.
>
> The string method took 35-37 msec on my machine.
> The array method took 84-85 msecs.
>
> But i realized that as well as changin string -> array, you had
> effectively changed from column-first to row-first processing, and
> row-first means you have a larger initial time before you can get a
> duplicate. I changed the array method to be column-first, which decreased
> the time to 54-56 msecs -- still much slower.
>
> But of course, if you do it in column order, there is no need to use an
> array, so I switched to s scalar variable, and decreased the time to 29-32
> msecs; roughly 15% improvement.  Thereby showing that "contains" is
> marginally faster than "replace ... with empty". Hmmm - no real surprise :-(
>
> Other observation is that in the "optimized" version, we already know
> there are no duplicates in the existing lines - only the last line can
> possibly duplicate with an earlier one. So you can modify "checkit" to use
> that fact, and get the time down to 24-25 msecs.
>
> Let's see that gave me twenty minutes of fun coding to save 12 millisecs.
> Just as well I enjoyed it, because the time saving didn't justify the
> effort.
>
> -- Alex.
>
>
>
> On 12/04/2013 16:34, Geoff Canyon wrote:
>
>> http://gcanyon.wordpress.com/**2013/04/12/brute-force-**insanity/<http://gcanyon.wordpress.com/2013/04/12/brute-force-insanity/>
>>
>> I posted my answer, so don't scroll down if you want to try yourself.
>> There
>> were some interesting aspects to the code.
>>
>> gc
>> ______________________________**_________________
>> use-livecode mailing list
>> use-livecode at lists.runrev.com
>> Please visit this url to subscribe, unsubscribe and manage your
>> subscription preferences:
>> http://lists.runrev.com/**mailman/listinfo/use-livecode<http://lists.runrev.com/mailman/listinfo/use-livecode>
>>
>
>
> ______________________________**_________________
> use-livecode mailing list
> use-livecode at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your
> subscription preferences:
> http://lists.runrev.com/**mailman/listinfo/use-livecode<http://lists.runrev.com/mailman/listinfo/use-livecode>
>