Re: an idea about dual cpus

In my experience overclocking has a very adverse effect on *NIX. I'm using a
single PIII 500 with 256MB RAM on scsi hd's and my load peaks at around .7
with ~=600 clients. I have previously ran squid on a lesser machine, with
less clients, with ide hds and determined that ide is the bottleneck on
performance. As to the idea of celerons, I wouldn't use them because of the
reduced cache on the chip/board(you'll be lucky to get 512k cache). Just my
opinion, I could be wrong.

--Jeff

On Tue, Nov 16, 1999 at 01:43:09AM -0700, Dustin Byford <dustin@nmsu.edu> wrote:
>
>
> Alex Rousskov wrote:
> >
> > On Mon, 15 Nov 1999, Dustin Byford wrote:
> >
> > > Our setup is currently consuming 80% of the cpu (average) with regular
> > > peaks to 100%. The machine is not io or network bound. So in an effort
> > > to fix the situation I was thinking about using smp. What does everyone
> > > think about this.
> > >
> > > Run 2 squid processes under different config files/log files/etc. Let
> > > them use icp to talk to each other through loopback. Give each process
> > > 2 of the 4 disks. This way each squid process will get scheduled on 1
> > > cpu. Of course the overhead of icp may be to inefficient. What do you
> > > think?
> >
> > a) Are you already using dual CPUs? I would not be surprised if a
> > single Squid installation benefits from using smp.
>
> No this is a single PII-400 over clocked to 500 MHz, 125 MHz bus, 41 MHz
> PCI, 1G RAM, 4 18G u2w drives.
>
> RAM, network, and disk are all doing very well. It's just the CPU
> that's holding us back. Currently it looks like we will get around 3M
> hits a day. Soon this should go up as we double our outgoing
> connection. First of all does this sound normal? Should I track down
> what is using all the CPU and simply quit doing it?
>
> Async-io is an option but previously when we turned it on squid would
> crash every few hours.
>
> If this is normal it seems like instead of splitting up our hardware and
> having 2 boxes connected through a "slow" 100 Mbit ethernet link with
> 512 MB RAM, 2 18G drives. It would be a better utilization of our
> hardware to use a smp box with each squid process talking through the
> system bus. We do have memory to spare for digests and the such.
> Perhaps using 2 100 Mbit ethernet cards if network becomes a problem,
> maybe we could run 2 squids as 2 wccp caches and let the router take
> care of splitting up requests.
>
> If this continues to be a problem I think we will try a dual celeron 500
> with the above idea. Of course I'll let everyone know how it goes.
>
> >
> > b) If you have to run two Squids and are afraid of ICP overheads,
> > give Cache Digests a try.
>
> >
> > c) Please share your SMP experience with others when you find a good
> > working solution.
> >
> > $0.02,
> >
> > Alex.
>
> Thanks everyone for all the replies.
>
> --Dustin
Received on Tue Nov 16 1999 - 07:23:33 MST