On Thursday 09 December 2004 11:36 am, Jim Menard wrote:
| trans. (T. Onoma) wrote:
| > On Thursday 09 December 2004 10:52 am, Brian Schröder wrote:
| > | Well, you're right. I can't imagine anybody wants to solve a crossword
| > | of the size whose layout would take noteworthy time to calculate. I
| > | only got the feeling that all this transpose, gsub, etc... stuff should
| > | take a lot of time. On the other hand ruby performance is nearly always
| > | different from what I expect.
| >
| > I don't expect mine to be anywhere near "fast" considering the approach I
| > took. But just for the sake of inquiry I put a quick benchmark together.
| > Should be trivial for others to fill in with their own. (see attached)
| >
| > My Results:
| >
| >   Tom's Solution 1000 runs:
| >                        user     system      total        real
| >   Default Layout:  3.390000   0.210000   3.600000 (  3.716914)
| >     Jim's Layout: 21.490000   1.470000  22.960000 ( 23.339115)
|
| Ouch. Of course, I don't plan to generate 1,000 crossword puzzles any time
| soon. I focused on elegant code, not needless optimizations :-)

Nay, nay. That's _my_ solution running your layout. Not your solution.

Feel better? :)

T.