On Fri, Jan 06, 2006 at 03:29:41PM -0800, Patrick R. Wade wrote:
> >You'd need a lot more than three sites to test before you'd be able to
> >reliably measure a difference in miliseconds per thousand messages.
>
> Assume that i have many more; three is the number for the example, and
> the number for the example shall be three. Four shalt thou not use in
> this example, nor yet two, excepting thou procedest thereafter to three.
> Five is *right out*.
>
> >Does that answer your question?
>
> Not really, but thanks.
Okay then, ignoring pattern compile time for one rule versus three (it's
hard to quantify in the abstract and will pretty much even out), it should
take an ia32 processor longer to establish a recipe context than to test a
pattern, by about a couple dozen clock cycles, depending on comipler,
optimizations, etc.
--
"We are what we repeatedly do. Excellence, therefore, is not an act,
but a habit."
-- Aristotle
_______________________________________________
EUGLUG mailing list
[email protected]
http://www.euglug.org/mailman/listinfo/euglug
