Interesting read about upcoming K9 processors

[Ketil Malde]

If your first sentence is true, HP's IA64 lines are in worse trouble
than even I thought they were.

I guess I should emphasize that I'm just guessing - I know about a few
Superdomes that do number crunching (which anyway is what IA64 is good
at). Are there any numbers; anywhere?

OTOH, I think the commercial segment generally chooses vendor first,
and architecture later. So the customers who used to buy HPPA boxes
now buy HP IA64 boxes, and will (happily?) buy HP x86_64 boxes, if
that's what it takes. They'll need some reassurance that the new
stuff is in some way better, but will tend to stick with the vendor as
long as they get the service they want. IMHO.

-kzm

 

[Nick Maclaren]

|>
|> >> I think most of HP's IA64 sales also are going to the scientific
|> >> computing segment.
|>
|> > If your first sentence is true, HP's IA64 lines are in worse trouble
|> > than even I thought they were.
|>
|> I guess I should emphasize that I'm just guessing - I know about a few
|> Superdomes that do number crunching (which anyway is what IA64 is good
|> at). Are there any numbers; anywhere?

Almost certainly, but HP are unlikely to disclose them

HP effectively consigned the academic and technical marketplaces
to limbo over a decade ago, and are a bit player in HPC. The
number crunching SuperDomes I know of were special deals to try to
promote them in those areas. I have not heard of many 'normal'
sales.


Regards,
Nick Maclaren.

 

[Tom Gardner]

(e-mail address removed) (Nick Maclaren) wrote in @pegasus.csx.cam.ac.uk:

Er, no. Look at [SGI's] financials and customers. The former are
now balanced, and the latter have retrenched to the 'technical'
marketplace. And we are a damn sight less worried about which
CPUs we use than the 'commercial' market, as we are accustomed to
rebuilding codes on a few year timescale. Provided that Intel
still produces IA64 systems and delivers boring enhancements
(shrinks, clock rate, caches etc.)

Interestingly, in the commercial arena there is a move to
commoditise the CPU as well. The core point is that much
*new* commercial development is being done in Java or .NET,
both of which have machine-independent "bytecode" plus
interpreters at their core. The choice is thus becoming
between Java or .NET, with the added twist of a choice
of application server.

It will, of course, be a long time before the *old* codes
can be consigned to the dustbin


--
Tom Gardner
Mobile Payments,
LogicaCMG Wireless Networks
_____________________________
LogicaCMG
2420 The Quadrant
Aztec West
Bristol BS32 4LD
United Kingdom

T: +44 (0) 117 901 7896 (direct)

 

[chrisv]

It's not really their problem and they don't really have a solution.
There are lots of old, badly written applications that just won't run
at all in WinXP... and MS doesn't really care.

Can't say I blame them for that. Sometimes it's just time to move on.
They seem to have made a very reasonable attempt at retaining
backwards compatibility (not that they had much choice)...

 

[George Macdonald]

So, it would appear that even in 1996 the concept was not to eliminate x86
entirely, in 1997 it was publicly stated that IA32 would be around for some
time after Y2K, and in 1998 it was publicly stated that IA64 would not be on
the desktop for at least another 3.5 years from *today*. This despite the
recollections of a few who are certain that Intel had more nefarious plans
early on...

Uhh, it's called spin. The fact you or I can't find the docs/pages to
confirm the "recollections" does not mean they did not exist. Nefarious is
your term but I did see a roadmap which showed x86 relegated to mostly STBs
by 2005... after a steady year by year decline in PCs.

Rgds, George Macdonald

"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??

 

[Keith]

Humbug. They don't want to be held back by _accidental_ backward
compatibility, which is a big difference.

Ah so... If an application I bought yesterday doesn't work today, it's
*my* fault? I don't think IBM became a giant with that attitude.

MS's marketing? No. But I can read about the effects of their recent
patches, not only to the base operating system.

You really believe they care? Why don't they patch Win2K without having
to sign up for a XPish license agreement? Come on! What's with the XP
license anyway? Yes, and next year you'll rent the OS. ...good plan
this "security" is.

 

[Keith]

Uhh, it's called spin. The fact you or I can't find the docs/pages to
confirm the "recollections" does not mean they did not exist. Nefarious is
your term but I did see a roadmap which showed x86 relegated to mostly STBs
by 2005... after a steady year by year decline in PCs.

The charts I saw were Itanic to exceed x86 sales by 2003 and by '05 x86
was relegated to the dust-bin of embedded losers. ...and that was in
1997ish. There was some discussion about this in AFC a few weeks ago (I'm
not the only one remembering such).

 

[Keith]

Of course you do. Just as Corse did. Makes perfect sense.

Except that we *were* right. RamBus is dead, *dead*, *DEAD*! So far our
predictions about Itanic are pretty much right on too. It's exceedingly
(and embarrasingly) late and equally as unimpressive. Sorta like the
promise of RISC over CISC (though all processors *did* go RISC).

I'll say no more on the subject, because it is just as emotional and just as
stupid as that argument was.

Somehow I doubt you.

I can see that the .chips crowd hasn't changed a bit - how sad.

What a condescending little shit you've turned into.

 

[Robert Redelmeier]

That's the whole point: that backward compatibility wasn't
promised by Intel, but Microsoft's market position and the
market's de facto demand for backward compatibility at all
costs was so strong that a minor implementation detail of
a new chip needed Microsoft's go-ahead.

Actually, not. Intel, AMD, etc can add whatever new they like to
their chips so long as it doesn't break any current software.
viz MMX, SSE[2], 3DNow, etc. That's how progress is made.
If they break current software, they'll probably not sell many.

You seem to have a bit of an odd view of backward compatibility.
Maybe a hazard of technophilia. The market doesn't know what
backward compatibility is, much less demand it _per se_.

The market buys machines to do work and solve problems. Always
has. Always will. The first step is application software --
what app will do what needs doing? Then what OS does that app
need, and what hardware is needed. Lotus 1-2-3 & WordStar sold
lots of original IMB PCs, XTs and ATs. DTP sold lots of Macs.
Many people have put up with inferior OSes and hardware to get
needed apps.

Backwards compatibility only comes into it since good apps
usually have been around a while and hence run on older hw.

If Microsoft is good at anything, it's marketing. And that's
why they now care about security.

They now _appear_ to care about security. More marketing,
I'm afraid. I haven't heard of them undertaking measures
like OpenBSD's line-by-line walkthrus.

-- Robert

 

[Dean Kent]

On Wed, 04 Aug 2004 08:55:56 +0200, Jan Vorbrüggen wrote:

Ah so... If an application I bought yesterday doesn't work today, it's
*my* fault? I don't think IBM became a giant with that attitude.

Well, that isn't the only reason IBM became a giant. There were a bunch of
salespeople, and some rather draconian contracts before the anti-trust
lawsuits. I can recall when I worked at ADP in 1978/79 that IBM threatened
to cancel the lease and yank the 370/155 if the 3203 printers were replaced
with a competitors. I left there in 1980, and when I went back they had
replaced the entire system with one from a company called Magnuson (or
something to that effect). Needless to say, they were *not* happy with IBM
tactics. I also recall some interesting interactions when Amdahl sold them
a half-meg memory upgrade. IBM refused to allow Amdahl to touch the box,
and IBM would not touch the Amdahl unit - so our system programmer had to
hook everything up.

Yeah, IBM has always been your friend. ;-).

You really believe they care? Why don't they patch Win2K without having
to sign up for a XPish license agreement? Come on! What's with the XP
license anyway? Yes, and next year you'll rent the OS. ...good plan
this "security" is.

Hmm. At my job we 'rent' the OS we run today from IBM - I think it is
called zOS. Perhaps that is why IBM became a giant, too? ;-).

Regards,
Dean

 

[Dean Kent]

The charts I saw were Itanic to exceed x86 sales by 2003 and by '05 x86
was relegated to the dust-bin of embedded losers. ...and that was in
1997ish. There was some discussion about this in AFC a few weeks ago (I'm
not the only one remembering such).

I saw charts in 1997 also - and none of them showed Itanium moving out of
the high end server segment. Though I was only given reseller roadmaps
directly, I was given OEM roadmaps from a motherboard maker associated with
a very, very large Asian OEM. I don't recall seeing any of those things -
and I still have those roadmaps. I wonder if anyone 'recalling' such things
does?

As for George's argument, as usual it is a fallacy. It is called
"argumentum ad ignorantium". Just because it cannot be proven to be false,
does not mean that it is true. The burden of proof is upon those trying to
make the claim. Public information says Intel did not intend to replace
x86 anytime soon, so it will take a bit more than 'recollections' to make
the case that they did. Sorry.

Regards,
Dean

 

[Yousuf Khan]

The charts I saw were Itanic to exceed x86 sales by 2003 and by '05
x86 was relegated to the dust-bin of embedded losers. ...and that was
in 1997ish. There was some discussion about this in AFC a few weeks
ago (I'm not the only one remembering such).

I think a lot of us can remember Intel's predictions about IA64 eventually
replacing IA32 by some point in time. You don't need some archival webpage
in order to prove it. Just the fact that so many of us who have been in this
business for so long can recall these statements is more than enough.

Yousuf Khan

 

[Dean Kent]

I think a lot of us can remember Intel's predictions about IA64 eventually
replacing IA32 by some point in time. You don't need some archival webpage
in order to prove it. Just the fact that so many of us who have been in this
business for so long can recall these statements is more than enough.

Mass recollections of events is no guarantee of the accuracy of those
recollections. The repeated telling of events can 'shape' those memories,
though they may seem to be vivid and accurate. Human recollection is a
very unreliable determinant of actual events.
http://www.memory-key.com/ResearchReports/tversky2000.htm

What *is* reliable is factual evidence, which has yet to be presented.
While the lack of evidence is no proof that Intel did *not* intend to
replace x86 in the near future, it is most certainly a reason to doubt the
accuracy of the recollections of those who most vehemently dislike
Itanium...

And, as I said - I have official Intel roadmaps from 1996 thru 2000. On
paper and in electronic form. You have recollections. Excuse me if I
doubt your powers of recall unless and until some better evidence is
presented. Sorry.

Regards,
Dean

 

[Yousuf Khan]

So, it would appear that even in 1996 the concept was not to
eliminate x86 entirely, in 1997 it was publicly stated that IA32
would be around for some time after Y2K, and in 1998 it was publicly
stated that IA64 would not be on the desktop for at least another 3.5
years from *today*. This despite the recollections of a few who are
certain that Intel had more nefarious plans early on...

Regards,
Dean

Hey, I dug up an old Register article by Mike McGee from April 1999:

http://www.theregister.co.uk/1999/04/28/secrets_of_intels_ia64_roadmap/

<quote>
Secrets of Intel's IA-64 roadmap revealed
By Mike Magee
Published Wednesday 28th April 1999 11:48 GMT
Updated Reliable sources said yesterday that a future Intel IA-64 chip
called Northwood would hit 3000MHz at its release. At the same time, it
emerged that McKinley is likely to launch using P858 aluminium technology.
The source who requested anonymity, works at Intel's R&D centre in Israel.
He said that all generations of microprocessors following Deschutes are
developed in pairs: Katmai-Tanner, Coppermine-Cascades and
Willamette-Foster. Northwood, like Madison and Deerfield will be X60
compactions of the IA-64 but for the Willamette architecture. Northwood,
further, is missing a Xeon counterpart and that suggests that Merced,
McKinley and Madison are likely to replace IA-32 server chips. Deerfield is
likely to be the first IA-64 chip aimed at the consumer market with a launch
date in the 2003 timeframe. Meanwhile, the source said there is "practically
no way" that Willamette and Foster will use copper technology. According to
another source at Intel Germany, the Merced platform was originally laid out
for .35 micron technology...
</quote>

So it looks like at one time Northwood was supposed to be an IA-64 chip, but
it actually turned out to be an IA-32 Pentium 4. Now it also looks like they
were saying that Northwood would have "no Xeon counterparts", so it would
seem that Northwood was meant to be a desktop chip -- which it actually
turned out to be. Of course Northwood did actually end up having Xeon
counterparts (what were they? Gallatin, or something?), but still 32-bit
Xeons.

It's fun digging up archives. It's almost like being a paleontologist.

Yousuf Khan

 

[Dean Kent]

So it looks like at one time Northwood was supposed to be an IA-64 chip, but
it actually turned out to be an IA-32 Pentium 4. Now it also looks like they
were saying that Northwood would have "no Xeon counterparts", so it would
seem that Northwood was meant to be a desktop chip -- which it actually
turned out to be. Of course Northwood did actually end up having Xeon
counterparts (what were they? Gallatin, or something?), but still 32-bit
Xeons.

It's fun digging up archives. It's almost like being a paleontologist.

Is it possible that Mike was wrong, or perhaps his source? Isn't the Israel
lab the one that developed Banias? Would someone there have direct
knowledge of IA-64 plans, or would it most likely be rumors? I wonder if
there has been any other time Intel has re-used a codename for two different
processors? So, if the Northwood was being planned as IA64 in mid-1999, but
by early 2001 it was released as a P4, that seems like a very short amount
of time to do such a redesign. I'm no EE, but it seems that if Intel has
such release cycles it would be pretty difficult to just change them on the
fly like this.

Regards,
Dean

 

[Dean Kent]

Is it possible that Mike was wrong, or perhaps his source?

Check this link to MDR. This is the 2nd half 2000 Intel forecast. That
means it was likely compiled in the first half, or perhaps even late 1999.
It shows Northwood as a P4 part, not IA-64. Can we presume that Mike or
his source were out in left field?

http://www.mdronline.com/publications/tl/intel_2h2000/toc2.html

Regards,
Dean

 

[Yousuf Khan]

Mass recollections of events is no guarantee of the accuracy of those
recollections. The repeated telling of events can 'shape' those
memories, though they may seem to be vivid and accurate. Human
recollection is a very unreliable determinant of actual events.
http://www.memory-key.com/ResearchReports/tversky2000.htm

Ah, I see, so we who have been in this business for so long are just
suffering from collective hallucinations?

What *is* reliable is factual evidence, which has yet to be presented.
While the lack of evidence is no proof that Intel did *not* intend to
replace x86 in the near future, it is most certainly a reason to
doubt the accuracy of the recollections of those who most vehemently
dislike Itanium...

And, as I said - I have official Intel roadmaps from 1996 thru 2000.
On paper and in electronic form. You have recollections. Excuse
me if I doubt your powers of recall unless and until some better
evidence is presented. Sorry.

Thus you shall have it. If the only proof you shall accept is documentary
proof (which would include webpages), then we shall do Google searches on
your behalf to find those precious archival webpages.

This article from late 1998, it was thought that the IA-32 line of
processors would end in 2003 with the Foster, and from that point afterwards
IA-64 would take over starting with Deerfield. It's also interesting to note
that back then, Intel thought 64-bit for the masses would take off starting
in 2003. It turned out that they were absolutely right, but it just wasn't
one of their chips, it was the Athlon 64 and Opteron.

http://www.theregister.co.uk/1998/10/22/madison_and_deerfield_to_split/

http://www.theregister.co.uk/1998/10/15/intel_doctors_foster_to_extend/

Now, it's obvious that Intel's plans didn't actually live upto its original
roadmaps. That's not surprising or unexpected. However, it's also not
important, it was their intention that is being discussed here only. It was
obvious that in 1998, Intel was hinting at replacing IA-32 at least by 2003.

Yousuf Khan

 

[Dean Kent]

Check this link to MDR. This is the 2nd half 2000 Intel forecast. That
means it was likely compiled in the first half, or perhaps even late 1999.
It shows Northwood as a P4 part, not IA-64. Can we presume that Mike or
his source were out in left field?

http://www.mdronline.com/publications/tl/intel_2h2000/toc2.html

Just found this post on comp.sys.intel back in Jan 1996. It says that Intel
delayed Merced until late 1998, and that there would be a performance
discrepancy when running x86 code.

http://www.google.com/groups?q=IA-6..._maxy=1996&[email protected]&rnum=2

I am finding it harder and harder to believe that Intel said anything about
replacing x86 with IA-64 in the 1996 and beyond timeframe (might have said
something in 1994/95 about it - but searching Google Groups turns up nothing
from anyone about this - so I am doubting that anyone actually heard such a
thing or it probably would have been posted as a question, at least).
Official information indicates that they knew there would be performance
issues, and that x86 would not be easily replaced.

David Wang posted in Dec 1996, that MDR claimed IA-64 would likely be
'phased in' over a long period of time (he estimates 5 to 10 years), but
also states that he believed Intel would be developing x86 processors "well
into the 21st century".

http://www.google.com/[email protected]&rnum=1

Searching Google Groups from 1981 thru 1997, I see nobody posting anything
about Intel claiming IA-64 would replace x86 in any official or unofficial
statement. I *do* see several people trying to make the argument that it
would, which picked up steam in 1997 (only a couple of posts in 1996). If
so many people saw charts and heard Intel statements, how come nobody posted
a single comment or question about it anywhere on Usenet?

I really am trying to find the evidence. It just seems so damned hard to
find that I must question whether it ever existed...

Regards,
Dean

 

[Dean Kent]

Thus you shall have it. If the only proof you shall accept is documentary
proof (which would include webpages), then we shall do Google searches on
your behalf to find those precious archival webpages.

This article from late 1998, it was thought that the IA-32 line of
processors would end in 2003 with the Foster,

Foster, Cascades and Tanner are Xeon parts, not desktop x86. You are either
being intentionally disingenuous, or your shades are letting through only
what will support your argument. The article specifically states:

"Graylish meanwhile confirms that Intel doesn't want people to get too
excited about IA-64 too soon. As reported here earlier, Intel's plans for
the continuation of IA-32 make it clear that it anticipates this being the
volume platform for some time"

and from that point afterwards
IA-64 would take over starting with Deerfield. It's also interesting to note
that back then, Intel thought 64-bit for the masses would take off starting
in 2003. It turned out that they were absolutely right, but it just wasn't
one of their chips, it was the Athlon 64 and Opteron.

Um. Exactly which 'masses' are you referring to? Opteron has 6-7% of the
market. A64 has even less.

http://www.theregister.co.uk/1998/10/22/madison_and_deerfield_to_split/

http://www.theregister.co.uk/1998/10/15/intel_doctors_foster_to_extend/

Now, it's obvious that Intel's plans didn't actually live upto its original
roadmaps. That's not surprising or unexpected. However, it's also not
important, it was their intention that is being discussed here only. It was
obvious that in 1998, Intel was hinting at replacing IA-32 at least by

2003.

Read it again. It did not say that. The entire article is discussing
server chips, and specifically states that IA-32 chips would be volume for a
long time after 2003...

Regards,
Dean

 

[Ketil Malde]

Just found this post on comp.sys.intel back in Jan 1996. It says that Intel
delayed Merced until late 1998, and that there would be a performance
discrepancy when running x86 code.

Well, the fact that they spent an effort on x86 compatibility seems to
make it pretty clear that they wanted it to target at least some of
the same market. At least, it's hardly something HP would insist on
in a HPPA replacement.

I am finding it harder and harder to believe that Intel said anything about
replacing x86 with IA-64 in the 1996 and beyond timeframe (might have said

IA-64 was supposed to be the Next Thing, much faster and better than
x86, and backwards compatible. Even if Intel didn't explicitly
sentence x86 to death, the market perception was that it was going to
happen.

Analogously, I think one would be hard pressed to find an AMD
statement giving a timeframe for ending x86 in favor of x86-64, but I
think it's fair to say that this is the plan and likely outcome.

Given the ('fin de siecle'? perception that IA64 would outperform
x86, have a 64bit address space, and perhaps most importantly: be
proprietary to Intel; why wouldn't Intel do all in their power to
terminate x86 as the CPU of choice? (Including newspeak about IA64
being "open", as opposed to "closed" architectures like SPARC, but I
digress.)

IMHO, they kept x86 available 'until IA64 could take over', and
only relatively recently (with the development of their own x86-64
line) changed to 'indefinitely'. Good for them they didn't terminate
x86 development early on (unlike some competing RISC chips).

-kzm