Sudoku Py / C++11 / D?

Era Scarecrow rtcvb32 at yahoo.com
Wed Aug 15 14:15:30 PDT 2012


On Wednesday, 15 August 2012 at 20:28:19 UTC, Jonathan M Davis 
wrote:
> Brute force is so fast that there's no really any point in 
> trying to solve it any other way except for the challenge of 
> doing so. I answered a question on this using D at 
> codegolf.stackexchange.com a while back:
>
  
http://codegolf.stackexchange.com/questions/378/implement-a-brute-force-sudoku-solver/402#402
>
> and the code is lightning fast. It would probably have to be 
> tweaked to match whatever Bearophile's code does though as far 
> is input goes (I haven't looked at the code that he linked to). 
> It also makes no attempt at being compact (e.g. it actually 
> checks the command line arguments). It's at just over 150 lines 
> and could be much shorter if I really tried to properly golf it 
> rather than just solve the problem.

  Interesting... Against the same input that brute force only one 
succeeded in 2 seconds vs my 9-12. And on the puzzle supplied on 
the Page, about 250ms compared to mine at 400ms.

  If I add a few lines to remove the only real bottle-neck (cache 
result of 4 functions) I'm sure mine would easily out-perform 
that one; But I wasn't going for absolute speed and keeping 
things simpler.


More information about the Digitalmars-d-learn mailing list