On Thu, 17 Apr 2008 16:04:06 -0700, Jon Skeet [C# MVP] <sk***@pobox.com>
wrote:
>Mind you, depending on your definition of "fast", processing 1 million
of
anything isn't necessarily ever going to actually be fast. :)
Ooh, I don't know about that. A little test app I wrote using
HashSet<stringtakes about 100ms on my laptop with N=1000000 and
M=50000.
Like I said, it depends on your definition of "fast". In some scenarios,
100ms is incredibly slow.
That's of course ignoring the question as to whether your test was even a
valid performance measurement. Depending on the hash implementation,
there could be a significant difference between the best-case (no
collisions) and worst-case scenarios (many collisions, no strings actually
in the intersection). But for the sake of this discussion, I'm willing to
grant you that. :)
Pete