Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Sparc Mhz vs. Intel Mhz for Oracle

Re: Sparc Mhz vs. Intel Mhz for Oracle

From: Burt Peltier <burttemp1ReMoVeThIs_at_bellsouth.net>
Date: Wed, 3 Dec 2003 00:25:49 -0600
Message-ID: <Ckfzb.15233$5d.2120@bignews4.bellsouth.net>


Didn't M$ buy tpc.org ? :)

Anyway, definitely take the "clustered" results at tpc.org with a grain (or molecule) of salt, especially when M$ can use a clustering process for benchmarks that no one would ever probably use in the real world.

It's so misleading I don't understand how it is allowed.

-- 
"Mike" <michaeljnugent_at_NOSPAMhotmail.com> wrote in message
news:vsqlpp49ompt43_at_corp.supernews.com...

> You might also take a peek at http://www.tpc.org/
>
> Grain of salt when checking all benchmarks...
>
> I like cross checking INT performance from SPEC for
> something like apples-apples for raw CPU (FP doesn't
> matter much for most RDBMS implementations), and
> then taking a look at TPC. Note that TPC tests more
> complex functions so is easier to manipulate by the
> vendor. Also note that some SPEC benchmarks handle
> parallelism for multiCPUs and others don't... keep an
> eye out for that.
>
> Don't forget about about 32 bit vs. 64 bit if you need
> the memory address capability.
>
> Don't you hate it when the RDBMS vendors drive you
> to a possibly unnatural system config, just because of
> their current per CPU pricing policies...!
>
> "Hans Forbrich" <forbrich_at_yahoo.net> wrote in message
> news:3FCCF735.A4BA6576_at_yahoo.net...
> > NetComrade wrote:
> > >
> > > What's windows?
> > >
> > > Suse or RedHat.
> > > We are CPU bound, that's why I am concerned. I know it's not an
> > > apples-to-applies comparison, hence asking if anyone has done it :).
> > >
> >
> > Ah!
> >
> > Do you run X? Using a remote X-server? With our without local
> > X-server? If local, with an X Desktop or just X services?
> >
> > I generally don't like benchmarks (too artificial .... vendors can and
> > do twist it). However I believe the SpecIntMark stuff
> > provides a relatively decent comparison of a variety of situations.
> > check http://www.spec.org
> >
> > /Hans
>
>
Received on Wed Dec 03 2003 - 00:25:49 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US