Re: Why is SUN falling so far behind IBM?

• Previous message: Arne S: "Re: min pp sizes for harddisk"
• In reply to: Arthur Corliss: "Re: Why is SUN falling so far behind IBM?"
• Next in thread: Arthur Corliss: "Re: Why is SUN falling so far behind IBM?"
• Reply: Arthur Corliss: "Re: Why is SUN falling so far behind IBM?"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Date: Tue, 24 Aug 2004 12:44:50 +0200

Arthur Corliss wrote:

>> Maybe. But at least our software angineers have no problem with
>> wasted bandwith on Itanium. And from what I know most other ISV
>> don't have that problem, too.
>
> Have they even done statistcal modeling to see what they should be
> getting compared to what they are getting?

Yes, of course. And comparisons with other platforms as well...

> Most engineers seem to be
> just ducky out-performing the last code base, without really
> understanding just what they should be getting out of the hardware.

Well, our engineers are quite experienced in HPC programming, most of them
already had a long-term experience when they came to us. We don't recruit
some Visual Studio Windows programmers...

> I would put forth that it's reasons like this why customers like
> Wright-Patterson AFB still put out the bucks for Origins.

Well, military often buys for reasons other than performance (I know that as
we supply a lot of military customers)...

> The
> pipeline to the processor is kept a hell of a lot fuller in them.

Maybe, but only because the MIPS processor is much slower to finish the
tasks ;-)

Altix and Origin for example share the exactly same architecture (with the
Altix gotting some additional improvements). I really doubt that the Origin
can even come somewhat close to throughput...

> Yes and no. I agree that MIPS hasn't pursued the "easy"
> optimisations like they could have (when SGI had direct control), but
> by the same token, as I mentioned above, they did compensate by
> trying to ensure nothing external
> to the processor was going to slow things down as well. That may
> have not compensated fully for the lack of clock rate, but it does
> make a huge difference.

Well, no. The _only_ advantage of Origin was their scalability to 2048 CPUs,
but with the new Itanium products from SGI that's nothing special any more.
And for the vast majority of HPC installations the scalability of Sun, IBM
and HP was more than sufficient because the number of customers requiring
that much CPUs is really really low. Origin had a very fast I/O architecture
but still isn't top-mark any more, and that for some time now. And fast I/O
can't compensate for low CPU performance, which is the reason why SGI
changed to Itanium while keeping the basic architecture...

> And at the same time I will put forth my assertion that if MIPS had
> been pushed like the other RISC architectures that it could be a
> superior computing architecture, even today. The processors would
> use a fraction
> of the die size and power for equivalent performance.

No. There are lots of factors in the MIPS architecture which makes it
different to achieve a better performance or a higher clock rate. SGI
itself, which had a very good development team, failed because of these
limitations. You can't just shrink the die and expect to be able to raise
clock rates at infinitum...

> Don't get me wrong: my shop is primarily an IBM shop (p & iSeries),
> so I like IBM's hardware. But let's face it: the Power-n processor
> is to hardware what modern applications (dare I say it? C++ code ;-)
> are to software -- extremely fat and bloated. The only saving grace
> is that processors, for all their bloat, do appear to gain us some
> performance improvements, while the latest software merely serves to
> make those new processors feel as snappy as our old Z80s.

We're getting more and more an HP shop, so I have no current insights into
the pSeries. But my impression is that IBM is somewhat the opposite of SGI:
very strong CPUs with a relative low bandwith. Can't say if this is correct
for the current machines, though...

> Cray is much better off without SGI's mishandling (completely two
> different philosophies of scaled computing), but selling off non-core
> business units
> is hardly the only way SGI has stayed in business (though it has
> helped).

It was more than just help. Look at the financial situation SGI has been at
that times. Selling their assets was the only way to stay in business long
enough to get Altix started...

> To refute your assertions that Altix/Linux is the only thing bringing
> in
> new revenue I would point out only one of their publically listed
> "customer successes" in servers mentions the Altix, each of the
> others mention the Origin. And look at successes under visualization
> systems, that, too, is dominated by Onyxs and Origins. If the Altix
> was that vital to their revenue stream I'd think they'd do a hell of
> a lot more press releases on it, wouldn't you?

No. I know from a lot of sites that bought Altix systems. It seems sell
quite well. On the other side there aren't many MIPS/IRIX sales, and the few
sales are all upgrades or expansions from traditional IRIX customers (like
the said military agencies). SGI hasn't aquired new customers with IRIX for
a long time, and even the traditional customers that have bought now know
that this will be the last sale of MIPS/IRIX, and that it's time thinking of
moving on to something different.

Press Releases are fine, but if You read between the lines You'll see that
the MIPS/IRIX line hasn't a very hig priority at SGI any more. They are
working on moving their visualisation centers to Itanium, because today SGI
has some real competition in this area, and with Sun and HP offering more
performance than You can squeeze out of the old MIPS systems...

> Did you not read my post? I think I was pretty clear in saying that
> in the real world clock rate means nothing if you're just sitting,
> waiting for data to work on. Which is why the Origin is so
> effective, even at the slower rates.

I/O is nothing if the CPU is slow like hell. A system is only as good as the
weakest part. I/O on Origin is good but not as good as on Altix or
Superdome...

> <G> Compare the original 6.5 to 6.5.13, and that to 6.5.23. The
> versions lie, my friend, there's been major enhancements within the
> 6.5 stream. SGI just doesn't feel the need to pull a Sun-like move
> to pretend to advance the platform. How big a jump was 2.7 to
> Solaris 8? I don't believe for a minute that they've improved
> Solaris from 2.7 to 9 as must as IRIX has improved within the 6.5
> series.

What major enhancements? I have 6.5.24 here and besides from a lot of error
fixes (and introduction of a lot of new errors and incompatibilities) there
are lots of smaller improvements, but IMHO nothing special when thinking
that SGI is carrying IRIX 6.5 for over 6 years now.

>> Well, at least DEC had technology that was performing very good.
>> That's not the case for SGI MIPS...
>
> I disagree. The Alpha is a great chip, but I wouldn't rate MIPS too
> far behind.

Well, the experience we had when we had SGI and Alpha tells me otherwise...

> I'd take a MIPS box over a Sparc box anyday, that's for
> sure. And there's a good reason why my Octane sees more use than the
> Athlon-based box next to it, even though the Athlon is several
> multiples faster than my R12k...

Well, for hobby use the criteria might be different, and of course I can't
say anything about Your home use or Your Athlon (besides this, I find
comparisons between a cheap standard PC and a RISC workstation rather
useless). But I never would go back to work from my PC workstations to my
Octane, it's just much too slow and too limited for the work I have to do...

Benjamin

• Previous message: Arne S: "Re: min pp sizes for harddisk"
• In reply to: Arthur Corliss: "Re: Why is SUN falling so far behind IBM?"
• Next in thread: Arthur Corliss: "Re: Why is SUN falling so far behind IBM?"
• Reply: Arthur Corliss: "Re: Why is SUN falling so far behind IBM?"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]