My lack of enthusiasm for AMD

[Robert Myers]

I now have a machine with AMD processor: 4600+, dual core, 2.4GHz. No
smoke coming out of the back yet.

I bought it because it had the features I needed at a price that was
too attractive to resist. No other reason.

When I turn this into a Linux box, I'll have to figure out what Intel
has done to make life miserable for people who use its compiler with
AMD processors. Until then, I doubt I'll notice any difference, except
on csiphc, where no one will be able to accuse me of never having
bought AMD.

This article

http://www.marketwatch.com/news/sto...x?guid={B6EB10FD-4578-4B54-96FE-5FCF10DCBDDC}

does a reasonably good job of explaining my chronic lack of enthusiasm
for AMD:

<quote>

Still, AMD's warning is not necessarily good news for Intel.

Both companies have been ramping up production of PC chips, which has
led to record-high inventory levels in the PC-chip sector, according to
iSuppli, a technology research firm.
Intel has been pushing its new Core 2 Duo chips into the market in
large volumes, hoping to reclaim market share it lost to AMD.

In 2006, AMD boosted factory production between 30% to 35% from 2005
levels, company executives indicated at a Wall Street analyst meeting
on Dec. 14.

High inventory levels and falling chip prices will hurt both companies,
said Hickey, who had been a longtime backer of AMD until March of last
year.

"It's a war no one wins," Hickey said.

</quote>

AMD pulled a trick play and changed the direction of the industry. The
very nature of trick plays means that they are a one shot deal. You
use them to win a game, because you'll never get another shot at the
same trick again. AMD wounded Intel. That's all. It didn't gain
permanent market share (as I predict, and as the logic of the article I
cite confirms).

I'm not interested in more of the same at lower prices, which is what
the current competitive standoff has created. I want something new,
and companies have to be financially healthy to do that (even if AMD
got a one-time exception with its trick play). I'd rather see IBM,
Microsoft, and Intel battling one another. If AMD has weakened Intel,
that only leaves the real monopolist, Microsoft, in an even stronger
position, something that *still* doesn't seem to have sunk in on
management at IBM.

Robert.

 

[Yousuf Khan]

This article

http://www.marketwatch.com/news/sto...x?guid={B6EB10FD-4578-4B54-96FE-5FCF10DCBDDC}

does a reasonably good job of explaining my chronic lack of enthusiasm
for AMD:

<quote>

Still, AMD's warning is not necessarily good news for Intel.

Both companies have been ramping up production of PC chips, which has
led to record-high inventory levels in the PC-chip sector, according to
iSuppli, a technology research firm.
Intel has been pushing its new Core 2 Duo chips into the market in
large volumes, hoping to reclaim market share it lost to AMD.

In 2006, AMD boosted factory production between 30% to 35% from 2005
levels, company executives indicated at a Wall Street analyst meeting
on Dec. 14.

High inventory levels and falling chip prices will hurt both companies,
said Hickey, who had been a longtime backer of AMD until March of last
year.

"It's a war no one wins," Hickey said.

</quote>

And this article seems to indicate that Intel was hit by the exact same
problems:

ARNnet | Hit by reorganisation, Intel profit falls 39 per cent
"Intel had lost significant market share to rival, AMD over the year,
but many observers felt the company had rebounded in recent months with
the launch of many new chips, including the Core 2 Duo and quad-core
Xeon chips."
http://www.arnnet.com.au/index.php/id;1553898472;fp;8;fpid;0

It doesn't look like Core 2 Duo has allowed Intel to gain back any
marketshare.

AMD pulled a trick play and changed the direction of the industry. The
very nature of trick plays means that they are a one shot deal. You
use them to win a game, because you'll never get another shot at the
same trick again. AMD wounded Intel. That's all. It didn't gain
permanent market share (as I predict, and as the logic of the article I
cite confirms).

You seem to have read a different article than the one you quoted.

I'm not interested in more of the same at lower prices, which is what
the current competitive standoff has created. I want something new,
and companies have to be financially healthy to do that (even if AMD
got a one-time exception with its trick play). I'd rather see IBM,
Microsoft, and Intel battling one another. If AMD has weakened Intel,
that only leaves the real monopolist, Microsoft, in an even stronger
position, something that *still* doesn't seem to have sunk in on
management at IBM.

Intel has simply executed its own little trick shot here. So far it
hasn't gained one iota of marketshare from Core 2.

Yousuf Khan

 

[Robert Myers]

You seem to have read a different article than the one you quoted.

<quote>

"ThinkEquity Partners analyst Eric Ross said major PC makers are
"disillusioned" with AMD's technology roadmap and are likely to give
longer-term design wins back to Intel."

</quote>

<quote>

This has put AMD in a precarious - and familiar - position.

"Intel has always bombed them in price. The only time it didn't work is
when AMD had a big lead" in technology, Hickey added.

Intel has simply executed its own little trick shot here. So far it
hasn't gained one iota of marketshare from Core 2.

Not a trick play. A two minute drill to regain the lead, with the odds
at this moment being that it will probably for good. Or maybe AMD will
be back to trying another bet-the-franchise onsides kick.

As I've commented before, the commercial airline business went through
this same kind of ruinous competition, and now there is just *one*
commercially-viable maker of long-haul commercial airplanes. The
competition is Europe, Inc., not Lockheed or McDonnell Douglas. There
is competition for military aircraft, but that's because the Pentagon
wants it that way, not because a true market that would support the
level of competition actually exists.

Robert.

 

[gaffo]

<quote>

"ThinkEquity Partners analyst Eric Ross said major PC makers are
"disillusioned" with AMD's technology roadmap and are likely to give
longer-term design wins back to Intel."

</quote>

<quote>

This has put AMD in a precarious - and familiar - position.

"Intel has always bombed them in price. The only time it didn't work
is when AMD had a big lead" in technology, Hickey added.


Not a trick play. A two minute drill to regain the lead, with the
odds at this moment being that it will probably for good. Or maybe
AMD will be back to trying another bet-the-franchise onsides kick.

As I've commented before, the commercial airline business went through
this same kind of ruinous competition, and now there is just one
commercially-viable maker of long-haul commercial airplanes. The
competition is Europe, Inc., not Lockheed or McDonnell Douglas.

McDonnell Douglas ceased to exist nearly 8 yrs ago. It "merged" with
Boeing.

There
is competition for military aircraft, but that's because the Pentagon
wants it that way, not because a true market that would support the
level of competition actually exists.

Robert.

I think the days of massive new speeds and architectures are over (at
least using Si).

AMD will come out with their K8L (or whatever it is called) - which is
basically a souped up k8 (like the core is a souped up Pentium-3) - and
that chip will be on par with Intel's core.

i think at that point we will get a .45 die shrink and maybe one more.

I doubt that we will see any new architectures for many years. just
more cores for the core/k8l and smaller sized chips.

--

 

[Robert Myers]

McDonnell Douglas ceased to exist nearly 8 yrs ago. It "merged" with
Boeing.

Just as DEC "merged" with Compaq, which "Merged" with HP. Lockheed,
McDonnell Douglas (MacDAC), and Boeing were the three players in the
jumbo jet wars of the 80's.

Robert.

 

[Yousuf Khan]

Yousuf Khan wrote:


"ThinkEquity Partners analyst Eric Ross said major PC makers are
"disillusioned" with AMD's technology roadmap and are likely to give
longer-term design wins back to Intel."

</quote>

Great, now try to answer the actual question I was asking you, which was:

You seem to have read a different article than the one you quoted.

So where in that article did it say AMD pulled a "trick play"?

Not a trick play. A two minute drill to regain the lead, with the odds
at this moment being that it will probably for good. Or maybe AMD will
be back to trying another bet-the-franchise onsides kick.

Unfortunately, that's an interesting theory, but that's all it will
remain, a theory. That's because Intel has once again lost some
marketshare this past quarter, despite their "longer-term design wins".

http://img71.imageshack.us/my.php?image=q4mste6.jpg

Intel is now down to 74% marketshare (from 76% last Q), and AMD is upto
25%, with almost nothing left for VIA or Transmeta now. It's a true
two-horse race now.

BTW, if it took Intel a two-minute drill to regain the lead, why did it
take them 3 years to start the drill?

Yousuf Khan

 

[Robert Myers]

The logic of the article agrees with my prediction that AMD didn't
gain permanent market market share. They got a one time boost. I
chose to call it a trick play. What you choose to call it is a matter
of your writing skills. The words "trick play" were not enclosed in
quotes, and I nowhere implied that it was a quotation. I assumed that
you wouldn't set up such a silly straw man only to knock it down.
Apparently, you did. I'm not interested in csiphc AMD/Intel pissing
matches.

remain, a theory. That's because Intel has once again lost some
marketshare this past quarter, despite their "longer-term design wins".

http://img71.imageshack.us/my.php?image=q4mste6.jpg

Intel is now down to 74% marketshare (from 76% last Q), and AMD is upto
25%, with almost nothing left for VIA or Transmeta now. It's a true
two-horse race now.

Fascinating. Whenever mention what the markets are saying, you find
some piece of data that fits your preconceptions, even if it's a one-
day move in share price. My piece of data is that AMD lost 60% of its
market value over the last year. The pipeline is full of AMD
designs. The *design wins*, according to the article, are going to
Intel.

BTW, if it took Intel a two-minute drill to regain the lead, why did it
take them 3 years to start the drill?

Assign that to your MBA case studies class, or take such a class and
do it for yourself. I'm not interested in a fanboy pissing match.

Robert.

 

[Del Cecchi]

The logic of the article agrees with my prediction that AMD didn't
gain permanent market market share. They got a one time boost. I
chose to call it a trick play. What you choose to call it is a matter
of your writing skills. The words "trick play" were not enclosed in
quotes, and I nowhere implied that it was a quotation. I assumed that
you wouldn't set up such a silly straw man only to knock it down.
Apparently, you did. I'm not interested in csiphc AMD/Intel pissing
matches.



Fascinating. Whenever mention what the markets are saying, you find
some piece of data that fits your preconceptions, even if it's a one-
day move in share price. My piece of data is that AMD lost 60% of its
market value over the last year. The pipeline is full of AMD
designs. The *design wins*, according to the article, are going to
Intel.



Assign that to your MBA case studies class, or take such a class and
do it for yourself. I'm not interested in a fanboy pissing match.

Robert.

I would wonder if either of you have been involved on the inside of the
industry.
And in the Windows compatible business it doesn't seem that market share
is in any way a permanant thing, contrary to the beliefs of each side.
There is no real reason for customers to care much which company's
processor is in the box, right?

 

[Robert Myers]

I would wonder if either of you have been involved on the inside of the
industry.
And in the Windows compatible business it doesn't seem that market share
is in any way a permanant thing, contrary to the beliefs of each side.
There is no real reason for customers to care much which company's
processor is in the box, right?

It's actually not a belief of mine, and I don't know where you got
the idea that it is. I was citing the opinions expressed in the
article. The logic is that, when AMD has a big technical lead, it can
hold its own. When it doesn't, Intel "bombs them on price." You can
disagree with that if you like, but whether it's what the article says
or not or whether it's correct or not has nothing to do with whether
I've been inside the industry. The people whose opinions are cited
bet their careers and the money of investors on their opinions.

Following the logic of the article, the only question now is whether
there is another way for AMD to gain a big lead. The "Itanium-buster"
strategy was out there for a long time. Intel has taken its lumps,
and AMD's big lead has evaporated. What's AMD going to do for an
encore, call it a trick play or not?

As to whether customers care what's inside the box or not, Intel has
poured an awful lot of money into "Intel Inside" and "Centrino"
branding. Everything that I've read gives Intel high marks for the
money it has spent on those efforts. It didn't do any good though,
when it started to be obvious that Opteron was the clearly superior
processor. Now it isn't. Given two products that are
indistinguishable, customers buy on price and perception, a battle
that Intel has historically won.

Robert.

Robert.

 

[Yousuf Khan]

The logic of the article agrees with my prediction that AMD didn't
gain permanent market market share. They got a one time boost. I
chose to call it a trick play. What you choose to call it is a matter
of your writing skills. The words "trick play" were not enclosed in
quotes, and I nowhere implied that it was a quotation. I assumed that
you wouldn't set up such a silly straw man only to knock it down.
Apparently, you did. I'm not interested in csiphc AMD/Intel pissing
matches.

No, of course you're not interested in a pissing match, that's entirely
evident from your choice of subject title, "My lack of enthusiasm for AMD".

Fascinating. Whenever mention what the markets are saying, you find
some piece of data that fits your preconceptions, even if it's a one-
day move in share price. My piece of data is that AMD lost 60% of its
market value over the last year. The pipeline is full of AMD
designs. The *design wins*, according to the article, are going to
Intel.

Yeah, I'm sorry that facts and data are getting in the way of your
hypothesis. Do you have any data points on how much of *design wins* are
going Intel's way, to back up your case?

Assign that to your MBA case studies class, or take such a class and
do it for yourself. I'm not interested in a fanboy pissing match.

What's an MBA gotta do with it? That's a technical question. It took
Intel 3 years to come up with an answer to AMD's products, not two-minutes.

Yousuf Khan

 

[Yousuf Khan]

It's actually not a belief of mine, and I don't know where you got
the idea that it is.

Where do we get the idea that you were looking for a permanent market
share increase? How about when you said:

AMD pulled a trick play and changed the direction of the industry. The
very nature of trick plays means that they are a one shot deal. You
use them to win a game, because you'll never get another shot at the
same trick again. AMD wounded Intel. That's all. It didn't gain
permanent market share (as I predict, and as the logic of the article I
cite confirms).

It sure sounds like you were looking to see whether AMD gained
"permanent marketshare" in that quote above.

As to whether customers care what's inside the box or not, Intel has
poured an awful lot of money into "Intel Inside" and "Centrino"
branding. Everything that I've read gives Intel high marks for the
money it has spent on those efforts. It didn't do any good though,
when it started to be obvious that Opteron was the clearly superior
processor. Now it isn't. Given two products that are
indistinguishable, customers buy on price and perception, a battle
that Intel has historically won.

The "Intel Inside" strategy is precisely what the anti-trust lawsuit is
about, and precisely what Intel won't be allowed to use anymore
afterwards. In fact, Intel has already abandoned it. And without "Intel
Inside" threats and coercion, we've seen Intel go from a 90%+
marketshare company to a 75% marketshare company (so far). And that's
happening even with some of their most advanced and innovative
technologies in years.

Yousuf Khan

 

[Robert Myers]

Where do we get the idea that you were looking for a permanent market
share increase? How about when you said:


It sure sounds like you were looking to see whether AMD gained
"permanent marketshare" in that quote above.

In that sense, yes. What position will AMD find itself in (and,
indeed, is in already)? Pretty much the same position it found itself
in pre-Opteron: ruinous price competition with Intel. Whether the
analysts quoted in the article got it right or not, they agree with my
perception of the playing field: if AMD can't pull another ace out of
its sleeve, Intel will simply grind it into the ground. Nobody has
"permanent" market share, but companies and products have positions
that are hard to attack. AMD had such a position. The temporary (and
significant) advantage that AMD got from its Itanium-killer strategy
has evaporated, thus putting the game back on rules by which Intel
will win.

The "Intel Inside" strategy is precisely what the anti-trust lawsuit is
about, and precisely what Intel won't be allowed to use anymore
afterwards. In fact, Intel has already abandoned it. And without "Intel
Inside" threats and coercion, we've seen Intel go from a 90%+
marketshare company to a 75% marketshare company (so far). And that's
happening even with some of their most advanced and innovative
technologies in years.

Maybe one of these days we'll know the real story: how much of AMD's
market share was Intel being more cautious about sales tactics and how
much was because AMD simply had the better product. I tend to believe
that it is the latter.

I didn't state my own opinion here so much as find someone else whose
opinion roughly agreed with mine. Feel free to state your own opinion
and, should you care to, supporting opinion from elsewhere. As I've
stated, the fact that Intel's profits are under pressure because it is
in a price war is no surprise and doesn't invalidate what the article
I cited says.

As to Intel marketing, I continue to be impressed by its
aggressiveness. Core 2 Duo is everywhere you look right now. AMD
just doesn't have the resources.

For the IBM'ers here who don't think that marketing and perception are
important, consider IBM falling all over itself to rush out news of
its own high-K dielectric on exactly the same day as Intel. Marketing
and perception are important, independent of technical details. That
is to say: name brands *do* matter. Every computer I've seen
advertised recently is very explicit about what kind of processor is
inside.

Robert.

 

[Robert Myers]

No, of course you're not interested in a pissing match, that's entirely
evident from your choice of subject title, "My lack of enthusiasm for AMD".

I found an article that succinctly described why I have a position
that has been described as being a shill for intel. I don't claim
that it should be everyone's position, and I don't think it will ever
be yours, but I don't want to spend forever arguing about it. If you
feel that your position vis a vis Intel or AMD needs clarification,
please to tell us.
..

Yeah, I'm sorry that facts and data are getting in the way of your
hypothesis. Do you have any data points on how much of *design wins* are
going Intel's way, to back up your case?

I don't claim to be an industry insider. I only know what I read in
the newspapers, to coin a phrase. I read something and reported it
and my reaction to it. That's not an invitation to you or anyone else
to get into a pissing match. If you don't like the evidence I
presented, find some of your own. Telling me that Intel's profits are
down doesn't tell me a thing.

What's an MBA gotta do with it? That's a technical question. It took
Intel 3 years to come up with an answer to AMD's products, not two-minutes.

Intel's problems have been the result of bad strategy, so far as I can
tell. Intel was counting on pushing the edge of the frequency
envelope. AMD didn't have to. For intel, hitting the frequency wall
first, because of its high frequency strategy, was a killer. Intel
also thought it could force Itanium into the market. To the extent
that Itanium survives, it will never be what Intel expected it to be.

What AMD did was, in a sense obvious, at least for someone in a
competitive position: put the memory controller on the die, focus on
performance and not frequency, and use the extension to 64-bits to
provide some relief to the architectural register starved x86
architecture.

Intel stubbornly stuck to it's NetBurst/Itanium strategy, and it has
paid a price. The decision to pursue such a course is a *business*
decision, not a technical decision. That intel was taking a huge risk
in banking on always being able to operate at a higher frequency is
obvious in retrospect. How intel made those decisions and how it
should have made those decisions is studied at places like the Sloan
School, not in EE/CS departments.

Intel's decision to make a quad core out of two dual core chips is a
business decision, as is Intel's decision to keep the memory
controller off the die. The pluses and minuses are pretty well
known. The details and the dollars and cents are not. I'd love it if
someone knowledgeable would talk about how such decisions are made.
The only way I know to engage in such a discussion is to go to a first
rate school of management.

Robert.

 

[Del Cecchi]

In that sense, yes. What position will AMD find itself in (and,
indeed, is in already)? Pretty much the same position it found itself
in pre-Opteron: ruinous price competition with Intel. Whether the
analysts quoted in the article got it right or not, they agree with my
perception of the playing field: if AMD can't pull another ace out of
its sleeve, Intel will simply grind it into the ground. Nobody has
"permanent" market share, but companies and products have positions
that are hard to attack. AMD had such a position. The temporary (and
significant) advantage that AMD got from its Itanium-killer strategy
has evaporated, thus putting the game back on rules by which Intel
will win.

Maybe one of these days we'll know the real story: how much of AMD's
market share was Intel being more cautious about sales tactics and how
much was because AMD simply had the better product. I tend to believe
that it is the latter.

I didn't state my own opinion here so much as find someone else whose
opinion roughly agreed with mine. Feel free to state your own opinion
and, should you care to, supporting opinion from elsewhere. As I've
stated, the fact that Intel's profits are under pressure because it is
in a price war is no surprise and doesn't invalidate what the article
I cited says.

As to Intel marketing, I continue to be impressed by its
aggressiveness. Core 2 Duo is everywhere you look right now. AMD
just doesn't have the resources.

For the IBM'ers here who don't think that marketing and perception are
important, consider IBM falling all over itself to rush out news of
its own high-K dielectric on exactly the same day as Intel. Marketing
and perception are important, independent of technical details. That
is to say: name brands *do* matter. Every computer I've seen
advertised recently is very explicit about what kind of processor is
inside.

Robert.

Actually I believe the timing has to do with the ISSCC and its rules as
to prerelease of papers.

So both had papers and both became news on the day the embargo is lifted?

del

 

[Del Cecchi]

I found an article that succinctly described why I have a position
that has been described as being a shill for intel. I don't claim
that it should be everyone's position, and I don't think it will ever
be yours, but I don't want to spend forever arguing about it. If you
feel that your position vis a vis Intel or AMD needs clarification,
please to tell us.
.
I don't claim to be an industry insider. I only know what I read in
the newspapers, to coin a phrase. I read something and reported it
and my reaction to it. That's not an invitation to you or anyone else
to get into a pissing match. If you don't like the evidence I
presented, find some of your own. Telling me that Intel's profits are
down doesn't tell me a thing.

Intel's problems have been the result of bad strategy, so far as I can
tell. Intel was counting on pushing the edge of the frequency
envelope. AMD didn't have to. For intel, hitting the frequency wall
first, because of its high frequency strategy, was a killer. Intel
also thought it could force Itanium into the market. To the extent
that Itanium survives, it will never be what Intel expected it to be.

What AMD did was, in a sense obvious, at least for someone in a
competitive position: put the memory controller on the die, focus on
performance and not frequency, and use the extension to 64-bits to
provide some relief to the architectural register starved x86
architecture.

Intel stubbornly stuck to it's NetBurst/Itanium strategy, and it has
paid a price. The decision to pursue such a course is a *business*
decision, not a technical decision. That intel was taking a huge risk
in banking on always being able to operate at a higher frequency is
obvious in retrospect. How intel made those decisions and how it
should have made those decisions is studied at places like the Sloan
School, not in EE/CS departments.

Intel's decision to make a quad core out of two dual core chips is a
business decision, as is Intel's decision to keep the memory
controller off the die. The pluses and minuses are pretty well
known. The details and the dollars and cents are not. I'd love it if
someone knowledgeable would talk about how such decisions are made.
The only way I know to engage in such a discussion is to go to a first
rate school of management.

Robert.

Actually it is also a technical decision. And the way Intel makes such
decisions probably has not much to do with Harvard MBAs. Those are the
guys that came up with the Itanium plan.

del

 

[Yousuf Khan]

I don't claim to be an industry insider. I only know what I read in
the newspapers, to coin a phrase. I read something and reported it
and my reaction to it. That's not an invitation to you or anyone else
to get into a pissing match. If you don't like the evidence I
presented, find some of your own. Telling me that Intel's profits are
down doesn't tell me a thing.

I didn't give you a table that showed Intel's profits are down (though
they are, quite obviously), I gave you a table that showed Intel's
market share is down, while AMD's is up. I thought that's what your
whole argument was against AMD? As for fighting against evidence
presented, we're not the ones fighting the market share numbers.

Intel's problems have been the result of bad strategy, so far as I can
tell. Intel was counting on pushing the edge of the frequency
envelope. AMD didn't have to. For intel, hitting the frequency wall
first, because of its high frequency strategy, was a killer. Intel
also thought it could force Itanium into the market. To the extent
that Itanium survives, it will never be what Intel expected it to be.

The only problem Intel had was that Core 2's predecessors (Pentium M,
followed by Core 1) weren't ready for the non-laptop market, so Intel
had no choice but to push the Pentium 4 for the last 3 years. It needed
the last 3 years to design Core 2 to be more than just a laptop chip.

What AMD did was, in a sense obvious, at least for someone in a
competitive position: put the memory controller on the die, focus on
performance and not frequency, and use the extension to 64-bits to
provide some relief to the architectural register starved x86
architecture.

How was any of that an obvious path? They've only been the obvious path
since AMD did it that way. But even if they implemented only half of
those features, it would still be considered "next-generation".

Yousuf Khan

 

[chrisv]

The "Intel Inside" strategy is precisely what the anti-trust lawsuit is
about, and precisely what Intel won't be allowed to use anymore
afterwards. In fact, Intel has already abandoned it.

Eh? The "Intel Inside" marketing campaign is somehow
anti-competitive? Seems like quite-fair brand-marketing, to me...

And without "Intel Inside" threats and coercion,

Whatever the legalities are, I'll still take exception to your
description of a marketing campaign as "threats and coercion".

we've seen Intel go from a 90%+
marketshare company to a 75% marketshare company (so far).

Are you saying that this alleged market-share decline coincided with
the cessation of the "Intel Inside" marketing campaign? Or that it
coincided with the instigation of the anti-trust lawsuits? Because I
believe that to be entirely false, in either case. Has not AMD been
around 20% for some years, now?

 

[Robert Myers]

I didn't give you a table that showed Intel's profits are down (though
they are, quite obviously), I gave you a table that showed Intel's
market share is down, while AMD's is up. I thought that's what your
whole argument was against AMD? As for fighting against evidence
presented, we're not the ones fighting the market share numbers.

You assume that market share numbers reflect acceptance of current
designs. That's like assuming that the current through a capacitor or
inductor is proportional to the voltage across it. The article cites
*design wins*, not sales to Best Buy and CompUSA.

The only problem Intel had was that Core 2's predecessors (Pentium M,
followed by Core 1) weren't ready for the non-laptop market, so Intel
had no choice but to push the Pentium 4 for the last 3 years. It needed
the last 3 years to design Core 2 to be more than just a laptop chip.

Who knows when Intel threw in the towel on Pentium 4, or knew that it
would have to. Your assumption is that Intel decided, say around 2003
Fall IDF, that P4 was a lame duck. Many of us might have agreed, but
there is just no way to know what Intel thought or planned
internally. I sure wish I knew, but I don't.

How was any of that an obvious path? They've only been the obvious path
since AMD did it that way. But even if they implemented only half of
those features, it would still be considered "next-generation".

Alpha had the controller on the die. Somewhere I have an email from
an alpha advocate about alpha's low latency...something like 75 ns, if
I recall correctly. That email is at least six years old. The
downside of the low IPC of NetBurst was known, as it is known that
frequency as the overriding consideration was a marketing ploy. Intel
could have changed the x86 architecture any time and knew the
advantages but didn't because it wanted Itanium to take over the 64-
bit market. Itanium, not x86, was intel's next-gen processor. If you
want to argue that Itanium was a big mistake, you'd find many, even in
the corner offices of intel, who'd agree with you.

Robert.

 

[Robert Myers]

Actually it is also a technical decision. And the way Intel makes such
decisions probably has not much to do with Harvard MBAs. Those are the
guys that came up with the Itanium plan.

By all appearances, Intel's operations are driven by margin. The only
sound way to decide between an MCM and a single chip quad-core design
that I know of is the difference in margin. If you know of arguments
to the contrary, I'd love to hear them, if you can say.

I'll agree that the memory controller on the die is a balance of
technical and business considerations. One question is: do the watts
you free up by moving the memory controller off the die compensate for
the added latency you incur in doing so?

One the other hand, there are powerful business reasons for keeping
the CPU and memory controller separate, one of them being that Intel
has said it really doesn't want to be in the memory controller
business, another being that memory controllers can be engineered,
changed, and specialized separately from the CPU. I've always
assumed, correctly or not, that flexibility in operations has been the
powerful driver for Intel to keep the memory controller off the die.
That Intel has gotten away with it (sort of) constantly astonishes me.

In any case, I'd rather be talking about supercomputer designs based
on Cell.

Robert.

 

[willbill]

meaning that he is in fact wanting to
do a pissing match?

I found an article that succinctly described why I have a position
that has been described as being a shill for intel. I don't claim
that it should be everyone's position, and I don't think it will ever
be yours, but I don't want to spend forever arguing about it. If you
feel that your position vis a vis Intel or AMD needs clarification,
please to tell us.

curious comment given your opening post in this thread:

<"I now have a machine with AMD processor: 4600+, dual core,
2.4GHz. No smoke coming out of the back yet. ...

This article

www.marketwatch.com/news/story/amd-may-wilting-heat-applied/story.aspx?guid=%7BB6EB10FD-4578-4B54-96FE-5FCF10DCBDDC%7D

does a reasonably good job of explaining my chronic lack
of enthusiasm for AMD:"> <snip>

given that you finally own an AMD CPU/mobo (after
decades of Intel?), i find it deeply curious that
you seem to have taken the side of Intel in this
thread that *you* started.

i mean... LOL!

to my mind, the consumer is the one that wins when
competition like this takes place (so long as at
least 2 of the players survive (preferably more)

correct me if i'm wrong.

i mean, the AMD setup that you got...

(put together by yourself?)

*is/was* a better value than what you could
find with Intel when you did this at some
point prior to your opening Jan.12 post, right?

if not, then why did you buy AMD??

surely not just because you could say
you're now an owner of an AMD CPU/mobo?

all ears

also, why did you buy a single dual AMD CPU?

i mean, why not buy a true dualie mobo
with two singles on it?

bill