[Beowulf] [gregory.brittelle at kirtland.af.mil: Re: Intel?]

Robert G. Brown rgb at phy.duke.edu
Thu Jun 9 04:58:39 PDT 2005

On Thu, 9 Jun 2005, Franz Marini wrote:

> On Thu, 2005-06-09 at 03:25, Robert G. Brown wrote:
> > I think it is safer to rant less and refer to the present more -- at the
> > MOMENT I think that AMD has an obviously superior HPC architecture, both
> > multicore and otherwise.  
> Wait, whereas it's (probably) true that AMD has a superior HPC
> architecture right now, it is even more true that Apple's target market
> ain't HPC, for the most part. And, given their target market(s) and
> their experience with IBM (sub-par) ability to provide them with lots of
> cpus at low prices and high speeds (do you remember Apple's statement
> some two years ago about reaching 3 GHz with G5s ina  coupel of years
> ?), and AMD (proved) inferior ability to deliver quantity relative to
> Intel, Apple's move to Intel makes quite some sense.
> And you have not to forget the fact that right now Apple is biting dust
> with their Powerbooks. Intel has got a clear edge here over AMD with
> their P-M (which, btw, it looks like will be the future for Intel's x86
> processors). And P-M will move, soon, to the desktop, too. Intel only
> needs to add SSE2/SSE3, maybe HT, a faster bus and definetely dual cores
> to their P-M line and they've got a killer cpu. Faster than most x86 out
> there, and drawing a lot less current...

I'm not writing Intel off by any means.  I'm just noting that for
several years now AMD has blown the pants off of Intel on the HPC scene,
and that I have some doubts about their CURRENT picture of multicore
(after having seen roadmap presentations from Intel, AMD and Sun, BTW,
so I've actually looked a bit at what they claim they'll being doing
ahead).  All I can say is that thermal issues and throttling and the
like are still very much on everybody's mind as they try to cram ever
more stuff on ever less real estate.  But Intel is a BIG company with
SMART people and LOTS of money on the table.  I'd be insane to claim
that they won't work things out and remain competitive, even if I see a
window over which they won't be the price/performance leader and where
they will lag AMD in the practical engineering.

> > When they're both running the same CPU at the same
> > speeds, when one can get a Dell with linux or windows or an Apple with
> > MacOS (both basically unix plus a windowing system) and the hardware --
> > especially that all-important "CPU clock" is basically the same -- how
> > can Apple charge that premium that pays the rent on a warehouse sized
> > store in the hottest shopping center around?
> Support, a better OS (compared to Windows), a way better and much more
> consistant windowing system (compared to Linux). 
> See, you have to consider Apple's target market. Most Mac users out
> there are in the image/photography industry (well, maybe not most, but
> quite a lot of them ;)). Macs have quite an edge there over anything
> else. Not because of the hardware (in fact, they're lagging quite a bit
> behind x86s in that compartment), but because of a superior OS/software
> layer.
> Just to give you a quick example, Linux is out of the equation when it
> comes to graphic, simply because color management support is nil.
> Windows' one is much better, but it's still much worse than OS/X's one
> (for example, using a multi-monitor, calibrated and color managed setu

All software.  Yes, yes, yes.  Apple is a software company.  With a
hardware arm that has been an anchor around their corporate neck
forever, largely the legacy from Steven Jobs, who (after all) cofounded
a COMPUTER company not a SOFTWARE company.

> under windows is almost impossible, whereas it is a snap under OS/X). :)
> So I think it all comes down to what is Apple's (perceived or real)
> market. 
> > For Apple the real
> > challenge is to transition to some sort of sustainable model, hardware
> > and/or software, before the remnants of their "branding" is spent and
> > nobody cares anymore.
> > 
> > Really interesting, actually.
> It is, indeed. 
> > As I understood it, Apple was still not being clear about whether or not
> > they were going to do a CELL computer.  Does the latest announcement
> > clarify that?  Are they going to ONLY Intel, or is CELL still a
> > possibility?
> Dunno know for sure, but I think they haven't completely excluded the
> possibility of developing something around CELL. 

That's what I've read, but lack in hard data.  Not excluded isn't the
same as "are devoting resources to".

> But, I right now foresee a problem (and not an Apple's one, but with
> regard to all possible CELL uses), again with IBM's production capacity.
> Right now IBM is committed to provide the CELL and a not-so-different
> (as far as i can tell from the [little] docs available about the
> xbox360) cpu for the future PS3 and XBOX360 consoles. We're talking
> about (really) high numbers here. It looks like NVidia will be switching
> to TMSC for their new gpu, because of limitations in the IBM fab that's
> producing most of their chips right now, both from a technical and a
> quantity point of view. 
> All these things make me wonder if IBM will be effectively able to
> deliver...

Look at IBM's CELL partners.  Hmmm, BIG japanese chip foundries, hmmm.

Not so good for the USA's industrial sector, but I have faith that they
can deliver numbers of demand for numbers materializes, and Japan can
likely deliver a good chunk of the Asian market pushed by IBM.  IBM
>>is<< a hardware company (unlike Apple:-).  They "invented" the only PC
that really counts.

> Anyway, time will tell. 
> Meanwhile, I'm looking forward to see IBM's presentation at LinuxTag
> 2005 of their CELL workstation with linux. It could shed some light on
> the real capabilities of the CELL processor for general computing...

Please share, if it does.  I at least would love to hear it.


> Franz
> ---------------------------------------------------------
> Franz Marini
> Sys Admin and Software Analyst,
> Dept. of Physics, University of Milan, Italy.
> email : franz.marini at mi.infn.it
> phone : +39 02 50317221
> ---------------------------------------------------------

Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu

More information about the Beowulf mailing list