On Tue, Mar 18, 2008, Henrik Nordstrom wrote: > On Mon, 2008-03-17 at 13:37 +1100, Robert Collins wrote: > > it matter if process all the request on one CPU, or part on one part on > > another? Given NUMA it clearly does matter, but how many folk run > > squid/want to run squid on a NUMA machines? > > NUMA comes in many forms. SMP Opteron is a quote common NUMA server, and > a quite likely Squid server hardware candidate when we venture into > multithread/multiprocess land..
Well, with the advent of multicore boxes with hierarchical cache and memory constructs you will probably find a well-tuned application + aware operating system can get quite a benefit from abusing local caching. (A cache being cache aware; how delightful! :) Adrian -- - Xenion - http://www.xenion.com.au/ - VPS Hosting - Commercial Squid Support - - $25/pm entry-level VPSes w/ capped bandwidth charges available in WA -
