Intel follows lead of AMD, introduces model numbers

T

Tony Hill


I, for one, actually think this is a good idea... or at least
potentially good. MHz is a totally meaningless measure of CPU
performance, but somehow the marketing departments have managed to
convince people that MHz is all that really matters. I've even met
people studying computer science and working in IT departments who
were convinced that two chips running at the same clock speed would
always perform the same.

If Intel can come up with some sort of decent model numbering scheme
that does not simply copy clock speeds (ie something along the lines
of AMD's numbering scheme for their Opteron or Athlon64 FX line rather
than what AMD uses for their AthlonXP and regular Athlon64 lines), I
think it will be beneficial for consumers. This will help retailers
emphasize that it's actual performance, and not MHz that matters.
Then maybe people will buy chips that are good for what they want
rather than wasting their money on trash like Intel's Celeron chips
just because of the high clock speed (Intel's Celeron 2.8Ghz currently
sells for ~$150 but would be EASILY beaten by chips like AMD's
AthlonXP 2200+ selling for ~$75).
 
B

Black Jack

Tony Hill said:
I, for one, actually think this is a good idea... or at least
potentially good. MHz is a totally meaningless measure of CPU
performance, but somehow the marketing departments have managed to
convince people that MHz is all that really matters. I've even met
people studying computer science and working in IT departments who
were convinced that two chips running at the same clock speed would
always perform the same.

Intel were having trouble selling the Pentium-M processors, because
people would look at their Mhz and scoff.

In the desktop and laptop realm, people will continue to require
Mhz-simulating model numbers for next few years. Of course by that
time the model numbers and the Mhz will be completely askew of each
other, but it will still sort of look like Mhz to the regular public.
The Opteron & FX style model numbers are only for the cogniscenti. The
AMD Quantispeed marquecitecture really seems to have worked fine for
AMD, in the consumer space so I think we'll continue to see that even
with Intel. However, I can imagine that Intel will skew their numbers
a little bit higher than AMD's to make the AMD parts look slower.

BTW, does anyone remember Intel's benchmarking scheme from a few years
ago called iSpeed or something like that? It was designed specifically
for this purpose as well, during the time of the transition from the
486 to the Pentium, where the Pentium was a little bit slower in Mhz
than 486's from AMD at the time.

Yousuf Khan
 
N

Nick Maclaren

Intel were having trouble selling the Pentium-M processors, because
people would look at their Mhz and scoff.

Really? In the hobbyist market, perhaps. In the commercial sector
(where the profit is), things were and are very different. Quite a
few OEMs were building Pentium-M based servers despite Intel's
discouragement, because you can get double the number in a rack
compared to Pentium 4s (let alone Itania!)

A few years ago, there were only a few of us saying that Intel had
lost the plot with the Pentium 4 - there are now quite a lot of such
people, and the Inquirer is saying that Intel has actually taken the
decision to base its future ranges on the Pentium-M and not the
Pentium 4. I have no such inside information, but I can say that it
is the sensible strategy.


Regards,
Nick Maclaren.
 
R

Robert Myers

I, for one, actually think this is a good idea... or at least
potentially good. MHz is a totally meaningless measure of CPU
performance, but somehow the marketing departments have managed to
convince people that MHz is all that really matters. I've even met
people studying computer science and working in IT departments who
were convinced that two chips running at the same clock speed would
always perform the same.
No wonder the industry is in trouble. I don't think MHz is a
meaningless number and I don't particularly cotton to the idea of
model numbers. MHz only has meaning only in the context of the
particular processor you are considering, but, in that context, it
conveys more useful meaning than a model number.

Everybody has their own way of interpreting signs and portents. I
take this one as a sign that Intel is losing it. Copying AMD...again?

... This will help retailers
emphasize that it's actual performance, and not MHz that matters.
Then maybe people will buy chips that are good for what they want
rather than wasting their money on trash like Intel's Celeron chips
just because of the high clock speed (Intel's Celeron 2.8Ghz currently
sells for ~$150 but would be EASILY beaten by chips like AMD's
AthlonXP 2200+ selling for ~$75).

The people who buy those chips will never notice. Think of it as a
tax on stupidity. If Intel comes up with a numbering scheme that in
any way reveals what a scam the P4 Celeron's are, I'll _know_ they're
losing it.

On the other hand, maybe they're getting ready to dump NetBurst, and
they're afraid people will remember what a sweet deal Tualatin
Celeron's were. Is there a "value" Pentium-M out there yet?

I dropped comp.arch from the cross-post. These marketing issues are
not insignificant, and I don't mind reading and yammering about them,
but I don't think they belong on comp.arch.

RM
 
R

Rob Stow

Black said:
Intel were having trouble selling the Pentium-M processors, because
people would look at their Mhz and scoff.

Balderdash. Everyone seems well aware that a P-M will perform
on a par with a P4 clocked about 40% higher. What is stopping
them is the high prices for the P-M and tremendous difficulties
in getting a decent motherboard for the damned things.

Put P-M prices in line with their performance and put
motherboards for them on the market, and sales will take off.

Currently a 1.6 GHz P-M costs more that a 3.2 GHz P4 when
performance wise it should be priced more like a much
cheaper 2.4 GHz or 2.6 GHz P4. And the only motherboards
available for them are crap with no AGP slots and usually
only one DIMM slot.
 
R

Robert Myers

Put P-M prices in line with their performance and put
motherboards for them on the market, and sales will take off.

At least as big a threat to NetBurst as anything from AMD, that's for
sure. Intel's problem with MHz is not with selling Pentium-M's. It's
with reliving all the drama of: "Can you explain the rationale of the
NetBurst architecture again? I don't think I got it the first time."

Now that it looks like it might not scale as well as Intel had hoped,
that argument is gone. What's left? Hyperthreading? To get back to
a performance level they could have gotten with a slower clock but
without the ridiculously deep stack that keeps getting deeper.
Currently a 1.6 GHz P-M costs more that a 3.2 GHz P4 when
performance wise it should be priced more like a much
cheaper 2.4 GHz or 2.6 GHz P4. And the only motherboards
available for them are crap with no AGP slots and usually
only one DIMM slot.

Intel will not license Pentium-M for desktop applications, as I'm sure
you know. They are protecting P4/Xeon. Nothing more.

As to its price, the Pentium-M sells at that price where the lower
power consumption justifies the price.

RM
 
M

Mark Hahn

BTW, does anyone remember Intel's benchmarking scheme from a few years
ago called iSpeed or something like that? It was designed specifically

ispec.
the problem is that a single number is necessarily wrong to almost everyone.
 
T

Tony Hill

No wonder the industry is in trouble. I don't think MHz is a
meaningless number and I don't particularly cotton to the idea of
model numbers. MHz only has meaning only in the context of the
particular processor you are considering, but, in that context, it
conveys more useful meaning than a model number.

The thing I like about model numbers is that everyone knows that they
are meaningless. Nobody would go out to compare an ATI Radeon 9800 to
an nVidia 5900 and think "Obviously the ATI is a better video chip
because it's number is so much bigger!". However this is exactly how
a lot of people think about processors.
Everybody has their own way of interpreting signs and portents. I
take this one as a sign that Intel is losing it. Copying AMD...again?

Losing it or not, I think it's a good move. CPU's being sold
according to MHz is really a bit of an oddity in terms of marketing.
It's extremely rare for a mass-market product to be sold based
entirely on one rather unimportant factor of the products performance.
On the odd occasion you might see it in cars, eg. Lexus' GS300 vs. the
GS430, where the numbers directly relate to the engine displacement,
but even here this is the exception rather than the norm.

Generally speaking, people have grown to expect model numbers that
mean nothing unless the quality/performance of a part can be
accurately shown with a single number/letter/mark/whatever. I might
go out and buy Large eggs as opposed to Small eggs and have a pretty
good idea what I'm getting, but I certainly would expect to buy a
microwave marketed as an "LG 1000W".
The people who buy those chips will never notice. Think of it as a
tax on stupidity. If Intel comes up with a numbering scheme that in
any way reveals what a scam the P4 Celeron's are, I'll _know_ they're
losing it.

They definitely have a tricky proposition here. They do want to
market their P4 chips well and especially their Pentium-M chips
better, but they can't go entirely based on performance or it will
become painfully obvious that the Celerons are trash.
On the other hand, maybe they're getting ready to dump NetBurst, and
they're afraid people will remember what a sweet deal Tualatin
Celeron's were. Is there a "value" Pentium-M out there yet?

Celeron-M, released a few months ago. Nice little chip, though Intel
disabled most/all of the dynamic power saving features found on the
Pentium-M.
I dropped comp.arch from the cross-post. These marketing issues are
not insignificant, and I don't mind reading and yammering about them,
but I don't think they belong on comp.arch.

Come now, this is Usenet, the home of off-topic discussions! :>
 
R

Robert Myers

On Sun, 14 Mar 2004 20:50:53 -0500, Tony Hill

Come now, this is Usenet, the home of off-topic discussions! :>

I could have been more candid. I do a fair bit of posting to
comp.arch that I'm sure the hardware types regard as off-topic. There
are people on comp.arch who will complain about any kind of discussion
of marketability of a particular technology. Or rather, they seem to
want to discuss marketability as if it could always be related to some
sensible figure of merit.

Discussions of market strategy, positioning, and segmentation make
people squirm. I think there are people there who actually believe
that products are designed for performance and not for profit. Since
I have no other similarly reliable way of finding out, for example,
why a z-Series mainframe is still a marketable product, I'd prefer not
to waste my off-topic posts on numbering schemes for x86 chips.

RM
 
S

Sander Vesik

In comp.arch Rob Stow said:
Balderdash. Everyone seems well aware that a P-M will perform
on a par with a P4 clocked about 40% higher. What is stopping
them is the high prices for the P-M and tremendous difficulties
in getting a decent motherboard for the damned things.

Considering they are primarily a laptop cpu, it suprising you can
get any at all. This also affects the price.
 
R

Rob Stow

Sander said:
Considering they are primarily a laptop cpu, it suprising you can
get any at all. This also affects the price.

They are far and away Intel's best *desktop* cpu. I am
amazed that their refusal to market it accordingly does
not run afoul of any industrial laws, particularly in
places like California where energy shortages are a huge
concern and getting worse every year.

I keep expecting places like California to start putting
caps on the amount of power a desktop cpu can draw - much
the same way they have minimum fuel efficiency standards.
 
B

Black Jack

Rob Stow said:
Balderdash. Everyone seems well aware that a P-M will perform
on a par with a P4 clocked about 40% higher. What is stopping
them is the high prices for the P-M and tremendous difficulties
in getting a decent motherboard for the damned things.

The cogniscenti (us) know about it. The general public is a little
confused about it. But the general public barely knows what Wi-Fi is
either. Centrinos seem to be bought by enthusiasts and IT departments
for their bosses.

Yousuf Khan
 
R

Robert Myers

They are far and away Intel's best *desktop* cpu.

The only available published SpecCFP2000 numbers I am aware of, 547
base, 552 peak, Hewlett-Packard Company, ProLiant BL10e G2 (1.0GHz,
ULV Intel Pentium M), do not support that assertion, even if you
assumed that results would scale linearly with frequency without
changing the Front Side Bus, which they would not. Compare that to
Intel Corporation Intel D875PBZ motherboard (3.4 GHz, Pentium 4
processor with HT Technology Extreme Edition): 1548 base, 1561 peak.

I suspect that media benchmarks would look even more unfavorable for
the Pentium-M. SpecInt2000 numbers for the systems already cited 673
base, 687 peak for the Pentium-M and 1342 base, 1393 peak for the P4
systems already cited, respectively, make it more of a horse race, but
let's not get carried away.

RM



I am
 
G

George Macdonald

Balderdash. Everyone seems well aware that a P-M will perform
on a par with a P4 clocked about 40% higher. What is stopping
them is the high prices for the P-M and tremendous difficulties
in getting a decent motherboard for the damned things.

"Everyone"? That's not what I get as feedback from users when they see
their new laptop is "only 1.5GHz". Explanation of IPC and approx.
equivalence only invoke a "yeah...yeah" response.
Put P-M prices in line with their performance and put
motherboards for them on the market, and sales will take off.

It seems to me that there's something "funny" going on with P-M just now.
Up to 1.9GHz has been promised for >6months and yet current product is till
pegged at 1.7GHz max with many premium laptops still using 1.4 & 1.5GHz.
It would seem that either Intel is holding back on higher clock speeds or
they are having trouble making them. I do know that when you run a 1.6GHz
at full tilt it gets bloody hot.

Rgds, George Macdonald

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

Stephen Sprunk

Robert Myers said:
The only available published SpecCFP2000 numbers I am aware of, 547
base, 552 peak, Hewlett-Packard Company, ProLiant BL10e G2 (1.0GHz,
ULV Intel Pentium M), do not support that assertion, even if you
assumed that results would scale linearly with frequency without
changing the Front Side Bus, which they would not. Compare that to
Intel Corporation Intel D875PBZ motherboard (3.4 GHz, Pentium 4
processor with HT Technology Extreme Edition): 1548 base, 1561 peak.

This depends on a common definition of "best". If Rob meant _per Watt_
performance, I think the PM outperforms the P4 by a large margin...

S
 
N

Nick Maclaren

|> |> > On Sun, 14 Mar 2004 21:57:29 -0600, Rob Stow <[email protected]>
|> > wrote:
|> > >They are far and away Intel's best *desktop* cpu.
|> >
|> > The only available published SpecCFP2000 numbers I am aware of, ...
|>
|> This depends on a common definition of "best". If Rob meant _per Watt_
|> performance, I think the PM outperforms the P4 by a large margin...

And per cubic metre, per square metre, per kilogram, per rack,
per desk, ....

Currently, it doesn't do it per dollar, but Intel could change
THAT at the stroke of a pen[*]!

[*] A logical pen.


Regards,
Nick Maclaren.
 
T

Tony Hill

It seems to me that there's something "funny" going on with P-M just now.
Up to 1.9GHz has been promised for >6months and yet current product is till
pegged at 1.7GHz max with many premium laptops still using 1.4 & 1.5GHz.
It would seem that either Intel is holding back on higher clock speeds or
they are having trouble making them. I do know that when you run a 1.6GHz
at full tilt it gets bloody hot.

I think that this goes back to the problems Intel has been having with
getting their 90nm fab process up to steam. Intel originally planned
to start shipping their "Dothan" Pentium-M chips, made on a 90nm
process, a good 6+ months ago, but now it looks like they won't be
here for another 2-4 months at least. All the > 1.7GHz parts were
originally planned to be these "Dothan" chips, so until they start
shipping, 1.7GHz is the max we're going to see. In fact, it looks
like even reaching 1.7GHz is a bit much for the "Banias" Pentium-M
chips, hence the reason why most run at 1.4 to 1.6GHz.

I don't know quite what the problem is, but Intel definitely seems to
be having some SERIOUS troubles making 90nm chips. What Intel claimed
was going to be their fastest ramp of a new process technology ever is
turning out to be a VERY slow process.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top