Re: [ILUG] [OT] Celerons

From: Vincent Cunniffe (vcunniff at domain arbgroup.com)
Date: Wed 16 Feb 2000 - 11:23:07 GMT


ilug-admin at domain linux.ie wrote:
>
> On Wed, Feb 16, 2000 at 10:32:23AM +0000, Vincent Cunniffe mentioned:
> > > Sorry for the Off Topic post but, as I know that some people on this list
> > > are running Dual Celeron's on the BP-6 motherboards. I just thought I ask a
> > > few questions before ordering my one.
>
> I'd subscribe to the Linux abit BP6 mailing list, (majordomo at domain geek.net,
> with a body of 'subscribe linux-abit'), and see what sort of problems
> people are having with it. It's a good, fast board, that's flaky. I get a
> hard hang/crash about every three days. With my old Dell, I ran it for
> about ten months without a crash (well, without one that wasn't caused by
> a bug in my filesystem module ;)
>
> > > How good are they ? Will I really see 2 * the performance of a celeron ? Is
> > > it hard to configure under Linux ? How do they perform when compared with a
> > > PIII or Athlon for example ?
> > They're pretty good : not as powerful as a P-III of the same speed for some
> > things, but pretty powerful all the same. An Athlon would kick their asses ;-)
>
> And, to be honest a dual 533 Celeron will severly kick the ass off a
> 750Mhz Athlon for most stuff that uses the SMP.

Which is why I'm probably going to go for an Athlon 850+ from overclockers ;-)

The Celerons do lose quite a lot of power when the task doesn't fit neatly
into their caches... compared on Seti times, for example, they're a lot
slower than a P-III of the same speed.

So an Athlon 850 should pretty much equal the dual Celerons... I'll tell you
when I find out anyway.

> > For some apps, you can get a pretty linear speedup. For single-threaded
> > stuff you won't notice a difference, obviously. Kernel compilations are a
> > lot faster if you enable more than one concurrent job (I use make -j4),
> > and you'll really notice it if you're running things like VMWare, which can
> > take a whole CPU to itself without impacting the rest of the system.
>
> I'd also add that it's nice to have a CPU running X, and the other one
> running an application. X feels very snappy on an SMP box.

Yep, another advantage.

Vin



This archive was generated by hypermail 2.1.6 : Thu 06 Feb 2003 - 13:05:26 GMT