[SQU] Accelerator performance tuning?

I'm using squid as accelerator on the same box as the apache backends
and I wonder if I can tune some parameters.

Currently I have a bottleneck in the number of mod_perl processes I
can run simultaneously (memory restriction) and I have a load of > 5
but quite some *idle* CPU. It may be that disk IO is slowing me down,
but I do not believe so as procinfo does not show extraordinary high
numbers.

So I wonder if I could tune some squid parameter to get the 12
mod_perl processes to have less wait states than they currently seem
to have.

Thanks!

-- 
andreas
--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html