On Thu, 19 May 2005, Eric Bloch wrote:

Thanks for the note, Henrik.  I'm looking into setting up a few machines
and running a benchmark (likely using polygraph) on the 3.0PRE build to
see if kqueue is a win.  Has anyone done that yet?  Any general comments
on how the 3.0 code should compare performance-wise to the 2.x stuff?
I'm particularly interested in how well squid deals with lots of
concurrent connections.

There has been some users attempting benchmarking of Squid-3.0, however the code base has been somewhat buggy for some time so there hasn't been very many meaningful results.

I think the situation has improved considerably during the spring, even if the list of known bugs is still farily long.

Regards
Henrik

Reply via email to