Itanium sales hit $14bn (w/ -$13.4bn adjustment)! Uh, Opte..

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Yousuf Khan wrote:

>
> Yeah, gotta wonder about that. I thought the highest end Itaniums were
> supposed to be those SGI's? What with all of that supercomputer stuff they
> keep selling to NASA, etc. And who the hell are NEC's customers that they
> command such huge avg sales prices?
>
>
I'm not exactly sure where Japanese Altix servers would be accounted -- SGI
Japan is owned by NEC, and Altix servers sold there do not count as
revenue for Silicon Graphics, Inc. anymore...

Anyway -- the average there is only a matter of how you count installations.

Is the Dutch National Super's 400+ CPU installation one, two, four or
eight servers?

How many kernels are shepherding the installation depends on the whim of the
administrators...

--
Alexis Cousein Senior Systems Engineer
alexis@sgi.com SGI/Silicon Graphics Brussels
<opinions expressed here are my own, not those of my employer>
If I have seen further, it is by standing on reference manuals.
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Tony Hill wrote:
>> That was the entire point of Opteron -- bringing 64-bit computing to
>> the commodity market. Oh, and taking market share away from Xeon,
>> and showing IT managers what a stupid idea it is to lock themselves
>> into proprietary IA64 when they can run open AMD64 systems.
>
> Well, on the latter case they seemed to have done pretty well (though
> AMD64 was definitely not the only reason for IA64's rather limited
> success), but they aren't exactly taking a huge amount of market share
> away from Xeon. There was something like 1.4M Xeon servers sold in Q2
> vs. 60,000 Opteron servers. This gives the Opteron only about 4%
> market share. I guess this is a lot better than 0%, though at it's
> height the AthlonMP managed something like 5 or 6% of the global
> server market, so the Opteron hasn't even reached that stage yet,
> despite signing up some big OEMs.

I found this new article which gives the actual number of server chips sold:

http://www.thestreet.com/_yahoo/tech/semis/10181117.html?cm_ven=YAHOO&cm_cat=FREE&cm_ite=NA

http://tinyurl.com/3mfo4

In the second quarter of 2003, AMD shipped a mere 110,000 server chips
compared with Intel's 4.6 million shipments, according to Gartner. Since
then Opteron scored some major design wins, helping it nearly double
shipments to 205,000 as of the second quarter this year.

But it still lagged far behind Intel's 5.4 million shipments.



Now 205,000 chips into the 60,000 servers (previously stated) equates to
about on average 3.4 processors per server. Considering that the vast
majority of Opteron servers are usually either 2P or 4P, that makes complete
sense. And since the number is closer to 4P than to 2P, that would indicate
that more 4P Opteron servers got sold than 2P ones.

So it would seem, that Opteron's multiprocessing capacities are being
exploited to their utmost. Once 8P Opterons come into more common usage, it
would be interesting to see if corporations are utilizing their capacity
will be utilized too?

Wonder how many Xeon servers were sold that same quarter? That way we can do
the same math and find out what the average number of processors there are
in a Xeon.

Yousuf Khan
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Wed, 01 Sep 2004 20:56:08 GMT, "Yousuf Khan" <bbbl67@ezrs.com> wrote:

>
>I found this new article which gives the actual number of server chips sold:
>
>http://www.thestreet.com/_yahoo/tech/semis/10181117.html?cm_ven=YAHOO&cm_cat=FREE&cm_ite=NA
>
>http://tinyurl.com/3mfo4

Heres another,
Linux Server Shipments Grew 55 Percent in the Quarter
Wednesday, August 25 2004 @ 08:33 AM
http://www.linuxelectrons.com/article.php/20040825083301801
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Wed, 01 Sep 2004 16:11:25 -0500, Ed <nosay@home.com> wrote:

>Linux Server Shipments Grew 55 Percent in the Quarter
>Wednesday, August 25 2004 @ 08:33 AM
>http://www.linuxelectrons.com/article.php/20040825083301801

While Itanium still grew at strong rates, the industry saw the emergence
of the x86-64 CPU space, which had a year-over-year growth rate of 2,183
percent.

2,183%, what?
Ed
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Wed, 01 Sep 2004 21:23:46 GMT, "Yousuf Khan" <bbbl67@ezrs.com> wrote:

>Ed wrote:
>> Heres another,
>> Linux Server Shipments Grew 55 Percent in the Quarter
>> Wednesday, August 25 2004 @ 08:33 AM
>> http://www.linuxelectrons.com/article.php/20040825083301801
>
>But does that tell which were AMD and which were Intel.
>
> Yousuf Khan
>

? just more numbers to look at. ;p
Ed
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

"spinlock" <NullVoid@att.net> writes:

> vs 6000 Itaniums and 600 million in revenue?!?!?!

That is $100,000 per unit!

Where did you say that bridge was Nick...

--
Paul Repacholi 1 Crescent Rd.,
+61 (08) 9257-1001 Kalamunda.
West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

josmala@cc.hut.fi (Jouni Osmala) writes:

>> > Stephen Sprunk wrote:
>> >
>> >> That was the entire point of Opteron -- bringing 64-bit computing to
>> >> the commodity market. Oh, and taking market share away from Xeon, and
>> >> showing IT managers what a stupid idea it is to lock themselves into
>> >> proprietary IA64 when they can run open AMD64 systems.
>> >
>> >
>> > What do you mean by proprietary versus open?
>> >
>> > Would AMD let VIA or Transmeta implement AMD64 in their CPUs?
>> >
>> > For a fee or gratis?
>> >
>> > I suppose Intel would refuse to let another company produce
>> > IA-64 compatible chips?
>>
>> Transmeta has indeed licensed AMD64 from AMD, i don't know about Via.
>> Intel obviously is making AMD64 compatible chips also.
>>
>> I don't think Intel alone has the authority to let someone make a IA-64
>> compatible chip, apparently the patents are tied up in a company owned
>> by both Intel and HP.
>
> Purpose of that company is to keep IA-64 intellectual property for
> BOTH Intel and HP and exclude others from the fact. So you would need
> both of em to agree for letting anyone else make IA-64 processors. So
> if either of them says you cannot do that.

Not quite. The main reason is to sever the itanic from the two way licence
deals untel has done over the years. With lots of people. This way they can
just say, "Sorry, not our camel..."

--
Paul Repacholi 1 Crescent Rd.,
+61 (08) 9257-1001 Kalamunda.
West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Wed, 01 Sep 2004 22:57:56 +0800, Paul Repacholi
<prep@prep.synonet.com> wrote:
>
>"spinlock" <NullVoid@att.net> writes:
>
>> vs 6000 Itaniums and 600 million in revenue?!?!?!
>
>That is $100,000 per unit!
>
>Where did you say that bridge was Nick...

They actually only sold $319M in revenue on a bit shy of 6,000 units.
the $600M figure was for the year to date, so actually their
per-server average is somewhere a bit over $50,000. MUCH higher than
the ~$3000 average per Opteron server, but not $100,000.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Wed, 01 Sep 2004 20:56:08 GMT, "Yousuf Khan" <bbbl67@ezrs.com>
wrote:
>Tony Hill wrote:
>>> That was the entire point of Opteron -- bringing 64-bit computing to
>>> the commodity market. Oh, and taking market share away from Xeon,
>>> and showing IT managers what a stupid idea it is to lock themselves
>>> into proprietary IA64 when they can run open AMD64 systems.
>>
>> Well, on the latter case they seemed to have done pretty well (though
>> AMD64 was definitely not the only reason for IA64's rather limited
>> success), but they aren't exactly taking a huge amount of market share
>> away from Xeon. There was something like 1.4M Xeon servers sold in Q2
>> vs. 60,000 Opteron servers. This gives the Opteron only about 4%
>> market share. I guess this is a lot better than 0%, though at it's
>> height the AthlonMP managed something like 5 or 6% of the global
>> server market, so the Opteron hasn't even reached that stage yet,
>> despite signing up some big OEMs.
>
>I found this new article which gives the actual number of server chips sold:
>
>http://www.thestreet.com/_yahoo/tech/semis/10181117.html?cm_ven=YAHOO&cm_cat=FREE&cm_ite=NA
>
>http://tinyurl.com/3mfo4
>
>
>In the second quarter of 2003, AMD shipped a mere 110,000 server chips
>compared with Intel's 4.6 million shipments, according to Gartner. Since
>then Opteron scored some major design wins, helping it nearly double
>shipments to 205,000 as of the second quarter this year.
>
>But it still lagged far behind Intel's 5.4 million shipments.
>
>

Err, uhh, Yousuf, either you're cut 'n paste is a little wonky or they
changed the article since you read it. Now it reads:

"In the second quarter of 2003, AMD shipped a mere $110 million in
server chips compared with Intel's $4.6 billion in shipments,
according to Gartner."


Note the dollar values instead of units shipped. Lat though they do
seem to be switching back to units shipped:

"Since then, Opteron scored some major design wins, helping it nearly
double shipments to 205,000 as of the second quarter this year.

But it still lagged far behind Intel's 5.4 million shipments."


Hmm... strange.

>Now 205,000 chips into the 60,000 servers (previously stated) equates to
>about on average 3.4 processors per server. Considering that the vast
>majority of Opteron servers are usually either 2P or 4P, that makes complete
>sense. And since the number is closer to 4P than to 2P, that would indicate
>that more 4P Opteron servers got sold than 2P ones.

I'm not sure that this is accurate as it might also include some
AthlonMP chips where the 60,000 server number might just be for
Opterons.

Also the two numbers came from two different companies, so I wouldn't
be surprised if they are not measuring quite the same thing, it
certainly would not be the first time that Gartner and IDC came up
with conflicting reports.

>So it would seem, that Opteron's multiprocessing capacities are being
>exploited to their utmost. Once 8P Opterons come into more common usage, it
>would be interesting to see if corporations are utilizing their capacity
>will be utilized too?

I really doubt that the 3.4 processors/server number is accurate, it
seems just way too high considering that Sun only just recently
started selling 4P Opterons, IBM never sold them and many small OEMs
also stick to only 1 and 2P Opteron servers. I would be VERY
surprised if AMD is really selling more 4P Opteron servers than 2P
ones, it just doesn't fit the market dynamics at all.

Of course, part of the confusion might be related to dollar value vs.
unit shipment confusion mentioned above.

>Wonder how many Xeon servers were sold that same quarter? That way we can do
>the same math and find out what the average number of processors there are
>in a Xeon.

Roughly 1.4M Xeon servers were sold in Q2 of 2004. I don't know the
exact number, but it was somewhere around 1.6M total servers and about
90% of them are x86.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Wed, 01 Sep 2004 16:17:25 -0500, Ed <nosay@home.com> wrote:
>On Wed, 01 Sep 2004 16:11:25 -0500, Ed <nosay@home.com> wrote:
>
>>Linux Server Shipments Grew 55 Percent in the Quarter
>>Wednesday, August 25 2004 @ 08:33 AM
>>http://www.linuxelectrons.com/article.php/20040825083301801
>
>While Itanium still grew at strong rates, the industry saw the emergence
>of the x86-64 CPU space, which had a year-over-year growth rate of 2,183
>percent.
>
>2,183%, what?

There were hardly any x86-64 servers shipped in Q2 of 2003, so the
fact that 20 times as many servers shipped in Q2 of 2004 isn't all
that big of a surprise. They're just talking of growth from ~3,000
servers to ~60,000 servers.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

In article <5hgdj0t585h0n47907jnfb48s5hs8ua7j3@4ax.com>,
Tony Hill <hilla_nospam_20@yahoo.ca> wrote:
>
>There were hardly any x86-64 servers shipped in Q2 of 2003, so the
>fact that 20 times as many servers shipped in Q2 of 2004 isn't all
>that big of a surprise. They're just talking of growth from ~3,000
>servers to ~60,000 servers.

That's the point. IA64 CPUs dribbled onto the market and, in the
first couple of quarters that they were sold widely, pretty well
all sales were of workstations and small servers for development
and testing. The Opteron is less radical, so should ramp faster,
but the same is occurring.

The current figures are so confused as to tell us no more than the
Opteron is at least being a qualified success, and the Itanium has
neither crashed and burned nor taken off. No more than that.


Regards,
Nick Maclaren.
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Nick Maclaren wrote:

> That's the point. IA64 CPUs dribbled onto the market and, in the
> first couple of quarters that they were sold widely, pretty well
> all sales were of workstations and small servers for development
> and testing. The Opteron is less radical, so should ramp faster,
> but the same is occurring.

The thing is : It's not just Opteron, it's Athlon64 too. Athlon64
is where the volume will be (if anywhere), perhaps we're looking
in the wrong place ?

Cheers,
Rupert
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Tony Hill wrote:
>> I found this new article which gives the actual number of server
>> chips sold:
>>
>>
http://www.thestreet.com/_yahoo/tech/semis/10181117.html?cm_ven=YAHOO&cm_cat=FREE&cm_ite=NA
>>
>> http://tinyurl.com/3mfo4
>>
>>
>> In the second quarter of 2003, AMD shipped a mere 110,000 server
>> chips compared with Intel's 4.6 million shipments, according to
>> Gartner. Since then Opteron scored some major design wins, helping
>> it nearly double shipments to 205,000 as of the second quarter this
>> year.
>>
>> But it still lagged far behind Intel's 5.4 million shipments.
>>
>>
>
> Err, uhh, Yousuf, either you're cut 'n paste is a little wonky or they
> changed the article since you read it. Now it reads:
>
> "In the second quarter of 2003, AMD shipped a mere $110 million in
> server chips compared with Intel's $4.6 billion in shipments,
> according to Gartner."

Hmm, it looks like they re-edited the article since I originally read it. I
copy'n'pasted straight from the article up there. If they were talking about
dollars instead of units, then I wouldn't have even found it necessary to
quote it at all. It's a good thing I decided to quote excerpts from it,
otherwise people wouldn't have known what I was talking about.

> Note the dollar values instead of units shipped. Lat though they do
> seem to be switching back to units shipped:
>
> "Since then, Opteron scored some major design wins, helping it nearly
> double shipments to 205,000 as of the second quarter this year.
>
> But it still lagged far behind Intel's 5.4 million shipments."
>
>
> Hmm... strange.

Looks like there might still be some re-editing of the article left to do.
🙂

>> Now 205,000 chips into the 60,000 servers (previously stated)
>> equates to about on average 3.4 processors per server. Considering
>> that the vast majority of Opteron servers are usually either 2P or
>> 4P, that makes complete sense. And since the number is closer to 4P
>> than to 2P, that would indicate that more 4P Opteron servers got
>> sold than 2P ones.
>
> I'm not sure that this is accurate as it might also include some
> AthlonMP chips where the 60,000 server number might just be for
> Opterons.

I think that was simply 60,000 Opterons, from the Register article that I
originally posted to start off this thread. I doubt there's much AthlonMP
sales left.

In fact, I think people with AthlonMP mobos are probably going to need to
replace their Athlon MPs with Socket A Semprons from now on.

>> So it would seem, that Opteron's multiprocessing capacities are being
>> exploited to their utmost. Once 8P Opterons come into more common
>> usage, it would be interesting to see if corporations are utilizing
>> their capacity will be utilized too?
>
> I really doubt that the 3.4 processors/server number is accurate, it
> seems just way too high considering that Sun only just recently
> started selling 4P Opterons, IBM never sold them and many small OEMs
> also stick to only 1 and 2P Opteron servers. I would be VERY
> surprised if AMD is really selling more 4P Opteron servers than 2P
> ones, it just doesn't fit the market dynamics at all.

Well, they did say that the white boxers overwhelmingly outnumber the OEMs
in Opteron sales. Some of those whiteboxers include such brands as Verrari
Systems (formerly Racksaver), and others, which do have a large server brand
presense. So it may not have mattered if IBM, Sun or HP had their 4-way
boxes in place yet.

> Of course, part of the confusion might be related to dollar value vs.
> unit shipment confusion mentioned above.

We'll await the final re-edit. 🙂

>> Wonder how many Xeon servers were sold that same quarter? That way
>> we can do the same math and find out what the average number of
>> processors there are in a Xeon.
>
> Roughly 1.4M Xeon servers were sold in Q2 of 2004. I don't know the
> exact number, but it was somewhere around 1.6M total servers and about
> 90% of them are x86.

Well at that number, if 5.4 million Xeon chips were sold into 1.4 million
servers then that would come out to 3.8 chips/server average. So it would
mean 4-way Xeons outnumber 2-way Xeons, which doesn't make too much sense I
guess.

Yousuf Khan
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

In comp.arch Yousuf Khan <bbbl67@ezrs.com> wrote:
> Grumble wrote:
> > What do you mean by proprietary versus open?
> >
> > Would AMD let VIA or Transmeta implement AMD64 in their CPUs?
>
> As a matter of fact, yes.
>
> http://news.zdnet.co.uk/hardware/chips/0,39020354,2087519,00.htm
>
> Will let them use Hypertransport too.

Hypertransport licence is rather cheap i thought? On the order of
something that even a quite small company (10+ people) could afford?

>
> Yousuf Khan
>

--
Sander

+++ Out of cheese error +++
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Fri, 3 Sep 2004 22:40:01 +0000 (UTC), Sander Vesik
<sander@haldjas.folklore.ee> wrote:

>In comp.arch Yousuf Khan <bbbl67@ezrs.com> wrote:
>> Grumble wrote:
>> > What do you mean by proprietary versus open?
>> >
>> > Would AMD let VIA or Transmeta implement AMD64 in their CPUs?
>>
>> As a matter of fact, yes.
>>
>> http://news.zdnet.co.uk/hardware/chips/0,39020354,2087519,00.htm
>>
>> Will let them use Hypertransport too.
>
>Hypertransport licence is rather cheap i thought? On the order of


http://www.hypertransport.org/

FAQ #3. What does it cost to join the consortium?

The HyperTransport Consortium is based on five membership classes:
Promoter, Contributor, Advisor, Adopter, Academic. Major differences
between membership classes are the type of rights and free services the
member is entitled. Adopter memberships are $5,000 annually, Advisor
and Contributor membership are $15,000 annually, Promoter memberships
are $40,000 annually. Additional membership class information can be
found at the membership benefits information page.

Ed
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Sander Vesik wrote:
>>> Would AMD let VIA or Transmeta implement AMD64 in their CPUs?
>>
>> As a matter of fact, yes.
>>
>> http://news.zdnet.co.uk/hardware/chips/0,39020354,2087519,00.htm
>>
>> Will let them use Hypertransport too.
>
> Hypertransport licence is rather cheap i thought? On the order of
> something that even a quite small company (10+ people) could afford?

Yep, it is -- as a matter of fact. 🙂

Yousuf Khan
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

"Yousuf Khan" <bbbl67@ezrs.com> wrote in message
news:tcQYc.48$%ER1.39@news04.bloor.is.net.cable.rogers.com...
> ***Big News*** Intel's Itanium chips have hit the $14 billion in revenue
> mark!! However there was a small one-time over-optimism charge of $13.4bn.
> BUT THIS STUFF IS INCREDIBLE, IT'S EXACTLY AS IDC HAD PREDICTED ALL
ALONG!!
> That's an amazing 5,665 server units, this past quarter!!!
>
> PS- Oh, and btw, if you're interested (and frankly, I can't see why anyone
> would be), Opterons sold 60,000 server units, or something or another,
> blah-blah-blah.

One of the interesting things about numbers is that people can get
completely lost in them. For example, the numbers published by The
Register show that Opteron systems sold for an average of ~$3000 each,
while the Itanium systems sold for a mere ~$53,000 each. IOW, one Itanium
system is not necessarily equivalent (in either revenue, number of
processors, or market segment) to one Opteron system.

FWIW, there are only about 11,000 z/Series systems (or equivalent) in the
world, yet they run essentially all of the mission critical apps of Fortune
1000 companies. Number of systems sold is not an indication of
failure/success, nor of importance in a market. There are much more
important metrics one might want to focus on to paint a true picture of the
value of a product.

Regards,
Dean

>
> Now back to Itanium! HULK SMASH! HULK SMASH! Yeah!
>
> http://www.theregister.co.uk/2004/08/30/opteron_itanium_sales_q2/
>
> Yousuf Khan
>
> --
> Humans: contact me at ykhan at rogers dot com
> Spambots: just reply to this email address ;-)
>
>
>
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

In article <1N__c.16808$xy3.1399@newssvr27.news.prodigy.com>,
Dean Kent <dkent@realworldtech.com> wrote:
>
>One of the interesting things about numbers is that people can get
>completely lost in them. For example, the numbers published by The
>Register show that Opteron systems sold for an average of ~$3000 each,
>while the Itanium systems sold for a mere ~$53,000 each. IOW, one Itanium
>system is not necessarily equivalent (in either revenue, number of
>processors, or market segment) to one Opteron system.

If I recall, the first figure published for the average selling
price of Itanium systems was c. $15,000 - which was the price of
a high-end workstation. The initial buyers bought - surprise,
surprise - workstations for testing and development.

What will be interesting is to see how the average price of the
Opteron systems changes. If it goes up significantly, we have
evidence of more sales in the server and MPP/cluster market; if
it doesn't, then it is stuck in the workstation market.


Regards,
Nick Maclaren.
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Mon, 06 Sep 2004 18:36:41 +0000, Nick Maclaren wrote:

> In article <1N__c.16808$xy3.1399@newssvr27.news.prodigy.com>,
> Dean Kent <dkent@realworldtech.com> wrote:
>>
>>One of the interesting things about numbers is that people can get
>>completely lost in them. For example, the numbers published by The
>>Register show that Opteron systems sold for an average of ~$3000 each,
>>while the Itanium systems sold for a mere ~$53,000 each. IOW, one Itanium
>>system is not necessarily equivalent (in either revenue, number of
>>processors, or market segment) to one Opteron system.
>
> If I recall, the first figure published for the average selling
> price of Itanium systems was c. $15,000 - which was the price of
> a high-end workstation. The initial buyers bought - surprise,
> surprise - workstations for testing and development.

And do you suppose that this is happening with Opterons too?

> What will be interesting is to see how the average price of the
> Opteron systems changes. If it goes up significantly, we have
> evidence of more sales in the server and MPP/cluster market; if
> it doesn't, then it is stuck in the workstation market.

I'm not sure the average system price matters much here. If the UP or SMP
Opteron (1xx and 2xx) servers/workstations sell tremendously well and
the >4P servers sell tremendously well (for their segment), the average
system price will still be far lower than any Itanic (or Z, for that
matter).

I'm not sure how one compares chip ASP, on one hand, to system price on
the other.

--
Keith
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

"keith" <krw@att.bizzzz> wrote in message
news😛an.2004.09.07.02.37.41.937282@att.bizzzz...
> On Mon, 06 Sep 2004 18:36:41 +0000, Nick Maclaren wrote:
>
> > If I recall, the first figure published for the average selling
> > price of Itanium systems was c. $15,000 - which was the price of
> > a high-end workstation. The initial buyers bought - surprise,
> > surprise - workstations for testing and development.
>
> And do you suppose that this is happening with Opterons too?

It is my recollection that this is exactly the market that AMD originally
had in mind for Opteron. While many were making comparisons to Itanium
(which AMD skillfully has never denied), all of their marketing material was
about displacing Xeon systems. The relative pricing of the chips should
also be an indication of this.

>
> > What will be interesting is to see how the average price of the
> > Opteron systems changes. If it goes up significantly, we have
> > evidence of more sales in the server and MPP/cluster market; if
> > it doesn't, then it is stuck in the workstation market.
>
> I'm not sure the average system price matters much here. If the UP or SMP
> Opteron (1xx and 2xx) servers/workstations sell tremendously well and
> the >4P servers sell tremendously well (for their segment), the average
> system price will still be far lower than any Itanic (or Z, for that
> matter).
>
> I'm not sure how one compares chip ASP, on one hand, to system price on
> the other.

As you know, you can't. Companies such as Stratus will make
fault-tolerant, fully redundant systems selling for hundreds of thousands of
dollars using $1K/$2K Xeons. Thus far, these companies have not used
Opterons. Likely it has nothing at all to do with whether Opterons can or
cannot be used in such systems, but is due to the fact that such systems
take a *long time* to design, build and validate... and that their customers
expect certain attributes, including various name-brand components.

The real point here is that you cannot simply count systems sold across all
market segments, then make some general statement about the relative
success/failure of a component used within a fraction of them. This is not
an apples/apples comparison, and I suspect that many people know this - even
those who report/repeat such numbers. Of course, there is always the
problem of properly identifying each market segment as well as the intended
target of the component. Once you can do that, you have a better chance of
determining what 'success' and 'failure' really means.

On a somewhat related note, it was reported that Bob Evans passed away very
recently. He led the S/360 architecture team, which apparently cost $5B
back in the '60s - at at time when IBMs annual revenues were just North of
$3B. This would be the equivalent of Intel spending multiple tens of
millions of dollars on a new architecture (no, I am not trying to equate
S/360 with Itanium in anything except corporate investment terms). That
one turned out spectacularly successful, and I doubt that Itanium achieve
even a small fraction of that success - but if they can carve out, and hold
onto, a niche in the lucrative high-end space, it may not be as unsuccessful
as many would like it to be (or, it might be - but time will tell).

Regards,
Dean
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

In article <pan.2004.09.07.02.37.41.937282@att.bizzzz>,
keith <krw@att.bizzzz> writes:
|> >
|> > If I recall, the first figure published for the average selling
|> > price of Itanium systems was c. $15,000 - which was the price of
|> > a high-end workstation. The initial buyers bought - surprise,
|> > surprise - workstations for testing and development.
|>
|> And do you suppose that this is happening with Opterons too?

I know that it was, for the period for which we have figures.
Whether it is continuing is another matter.

|> > What will be interesting is to see how the average price of the
|> > Opteron systems changes. If it goes up significantly, we have
|> > evidence of more sales in the server and MPP/cluster market; if
|> > it doesn't, then it is stuck in the workstation market.
|>
|> I'm not sure the average system price matters much here. If the UP or SMP
|> Opteron (1xx and 2xx) servers/workstations sell tremendously well and
|> the >4P servers sell tremendously well (for their segment), the average
|> system price will still be far lower than any Itanic (or Z, for that
|> matter).

Sigh. Yes. That is largely because the Itanic has completely lost
out in the workstation and probably even small server market. As
other people say, that wasn't the intent. It certainly wasn't the
intent that it would be an HP and SGI only chip.

|> I'm not sure how one compares chip ASP, on one hand, to system price on
|> the other.

One doesn't. Or, at least, I don't.


Regards,
Nick Maclaren.
 
Archived from groups: alt.comp.hardware.amd.x86-64,comp.arch,comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

In article <mp9%c.17001$xd1.10147@newssvr27.news.prodigy.com>,
"Dean Kent" <dkent@realworldtech.com> writes:
|>
|> On a somewhat related note, it was reported that Bob Evans passed away very
|> recently. He led the S/360 architecture team, which apparently cost $5B
|> back in the '60s - at at time when IBMs annual revenues were just North of
|> $3B. This would be the equivalent of Intel spending multiple tens of
|> millions of dollars on a new architecture (no, I am not trying to equate
|> S/360 with Itanium in anything except corporate investment terms). That
|> one turned out spectacularly successful, and I doubt that Itanium achieve
|> even a small fraction of that success - but if they can carve out, and hold
|> onto, a niche in the lucrative high-end space, it may not be as unsuccessful
|> as many would like it to be (or, it might be - but time will tell).

I posted that analogy nearly a year back, and I was rhetorically
asked by a Itanic flag waver whether I meant that it would dominate
the whole industry for a decade. I replied, no, that I meant it
would take the company to the brink of bankruptcy (I believe that
IBM was within 6 months of filing) and it be a matter of chance
whether it went over the edge.

[ Note that, as Brooks says, it wasn't the hardware that had the
trouble, but the software. The hardware had its problems, but not
on the same scale. ]

Well, as it happened, the Intel bean-counters and senior executives
had enough sense to refuse to bet the farm on the Itanic (which
was believed NOT to be the case at the time I posted). So Intel
avoided the the crisis that IBM had with the System/360.


Regards,
Nick Maclaren.