Re: [squid-users] Tweaking Squid for speed, not max requests

On Thursday 30 August 2001 02:24 pm, Brian wrote:
> On Thursday 30 August 2001 10:01 am, Steve Snyder wrote:
> > There have been several threads on this list recently discussing the
> > tweaking that is required for improving max requests/second in Squid.
> > How about raw performance?
> >
> > I've already done the obvious: sufficient RAM, dual P3 CPUs, fast
> > (single) SCSI disk, and ReiserFS ("noatime,notail") filesystem. All my
> > NICs are 100Mbps, running at full-duplex. I've configured Squid to use
> > async-io. I'm using the default LRU replacement algorithm because I
> > found that LFUDA didn't really provide any overall benefit (in Squid
> > v2.3S4).
>
> Going for raw speed, I'd probably load the box with memory and run the
> cache_dir on tmpfs (a dynamically sized RAM disk). Since 'disk' reads
> would be instant, ufs would be the best choice. Of course, like you
> mentioned in another branch, that doesn't leave the other CPU with much
> to do. Maybe you could lock the NIC's interrupt to the second processor.

Why is ufs the best choice for accessing a RAM disk?

Thanks for the response.
Received on Thu Aug 30 2001 - 14:44:43 MDT