[Haskell-cafe] Re: (flawed?) benchmark : sort

[Ketil Malde]

"Krzysztof Skrzętnicki" <gtener at gmail.com> writes:

> The above results are for 1000000 Ints x 10 runs, but I don't expect any
> drastic changes in longer run. I leave the interpretation up to you.

Might I suggest (also) testing with numbers of smaller magnitude?
Lots of collisions is another killer for the naïve quicksort (albeit
easily remedied, of course), and something a general sorting algorithm
should handle. 

-k
-- 
If I haven't seen further, it is by standing in the footprints of giants