AMD Doomed?? PR Raiting Misleading??

[SSXeon5]

Not jumping to conclusions but it may be a possibility ...

http://www.sharkyforums.com/showthread.php?s=2cef116b0bf42ff5a5875a573dce0854&threadid=130217

good topic about the upcoming amd processors and how misleading the PR raiting is ... Arcadian had this point that was interesting:




<< Actually, watch this:

MHz -> Model number
-------------------
1.73GHz -> 2100+
1.67GHz -> 2000+
1.60GHz -> 1900+
1.53GHz -> 1800+
1.47GHz -> 1700+
1.40GHz -> 1600+
1.33GHz -> 1500+
-------------------
(Extrapolation using current method)
-------------------
1.26GHz -> 1400+
1.20GHz -> 1300+
1.13GHz -> 1200+
1.06GHz -> 1100+
1.00GHz -> 1000+
-------------------
(The performance rating breaks even, and now starts going backward)
-------------------
933MHz -> 900+
866MHz -> 800+
800MHz -> 700+
733MHz -> 600+
667MHz -> 500+
600MHz -> 400+
533MHz -> 300+
467MHz -> 200+
400MHz -> 100+
-------------------
(Now, the performance of the chip is break-even or *negative* at a given frequency)
-------------------
333MHz -> 0
266MHz -> -100
200MHz -> -200
133MHz -> -300
67MHz -> -400
0MHz -> -500
-------------------

See? The ratings don't make sense. As you say, it is a linear graph when performance does not scale linearly with frequency, and the line certainly does not cross at the origin, but at 1GHz, and after that, it's clear that the credibility of the rating sinks into oblivion.

Arcadian >>



Its been said that the hammer will be launched to OEM's in Q4-2002 but will not ship till Q1-2003. That said intel will have there 3.2Ghz Northwood out and the hammer might launch at just 2Ghz. TBred is delayed again and I havent heard anything on it ..... It isnt looking good for amd.

SSXeon

 

[SSXeon5]

another point Arcadian made .... BTW i got banned from sharky because the mods really didnt like me .... but i still scope it out and look through threads. As much as I dont like amd i still want them to be competitive with Intel and i wana see the hammer in action




<< Some people are making the assumption that the model 3400+ number would be assigned to the Hammer launch prediction of 2GHz. I think there is a little inconsistancy with that. For one thing, AMD started putting 3400+ on their roadmaps back last November.

http://home.attbi.com/~ptanner/roadmaps/index.htm

But since then, their desktop roadmap has slipped quite a bit, mostly because they have not been able to get the frequencies at .13u that they thought they would. Sanders recent announcement that Hammer may launch as low as 2GHz might be a low estimate, given newer circumstances. Therefore, 2GHz may actually have a much lower model number than 3400+.

Another way to look at it is this way:

AMD's current model numbering scheme has a 1.73GHz processor rated at 2100+. If we continue to scale as AMD has with 100 model numbers for every 67MHz, we get 2.06GHz equaling a model 2500+. Many sources have aleady suggested that Hammer will get a benefit of 15-20% on the applications that it is currently being tested on. If AMD were to be more conservative, they would label a 2.0GHz Hammer at model 2800+, and if they were less conservative, they would label it at 3000+. But 3400+ would be crazy, since that would suggest a >35% IPC improvement from the Athlon, and that doesn't nearly coincide with current data.

My theory is that AMD originally intended Hammer to debut at 2.2GHz or higher. An Athlon at this speed would have a model number of 2800+. If we take the less conservative improvement of 20%, we get a model number much closer to 3400+. If AMD gets the clock speed up to 2.6GHz, then an equivalent Athlon would be labeled 3400+, but Hammer would be 4000+, which is the top speed that the roadmap indicated could be available by the end of the first half of 2003. .09u manufacturing would take it up to 2.8GHz, which again suggests an Athlon of 3700+, but a 20% faster Hammer would be 4400+.

These are all straight-forward assumptions, and if AMD could get the frequency, and kept the same formula, then they could boast those model numbers. The only problem is that they are already scaling less than Northwood. A model 2100+ Athlon performs awefully close to a 2.2GHz Northwood, but if you look backward, a 1600+ Athlon outperformed even a 1.8GHz Pentium 4 with the same Northwood core. It's because Athlons are only increasing by 67MHz every time they add 100 model numbers to their "performance rating", so while the Athlon has increased 333MHz by going from 1400+ to 2100+, Intel has increased 400MHz by going from 1.8GHz to 2.2GHz (even though marketing would have you think that AMD made the larger jump). If things continue on this course, then Northwood will move faster, and eventually reach a point where it can outperform a similarly modeled Athlon - especially with the FSB enhancement that will soon be offered.

That's why I'm skeptical that AMD can reach a 3400+ processor in January next year, given their current forecasts, and still have it outperform a 3.0GHz Pentium 4 by more than 20%. Things just don't add up, and AMD may have to adjust their model number forecasts soon to align with newer developments.

Arcadian >>



What do you all think?

SSXeon

Keep it cool people. We don't want this to turn into a flame war, and we are watching this one closely.

AnandTech Moderator

 

[SteelCityFan]

I think this is just going to start another Intel Vs AMD dispute.. at least at some point.

Why don't we all just wait and see? Debating processors and release dates listed on paper is pretty meaningless IMO. Dates are missed, things change.

I don't worry too much about who is faster, better for me, etc until I am close to upgrading (once every 2 years roughly)...

 

[EXman]

I think Intel is misleading with their Mhz. At least the pre northwoods.

Regardless of clockspeed there are other factors in crunching 1's and 0's And Intels superior marketing team knows that Mhz sells to most of the flock of computer illiterates out there. While the AMD clock for clock smokes the Intel. why IPC is the answer.

SSXeon5 I remember you at sharky's you're a intel die hard SE forums suck anyhow...

My 2 cents

 

[mee987]

"arcadian" is doing nothing but speculating on the little information that we know. he thinks he has it all mathematically figured out, but it means nothing. why the hell is he counting backwards to see what a "100+" processor would be clocked at? WHAT DOES IT MEAN?



<< My theory is that AMD originally intended Hammer to debut at 2.2GHz or higher. An Athlon at this speed would have a model number of 2800+. If we take the less conservative improvement of 20%, we get a model number much closer to 3400+. If AMD gets the clock speed up to 2.6GHz, then an equivalent Athlon would be labeled 3400+, but Hammer would be 4000+, which is the top speed that the roadmap indicated could be available by the end of the first half of 2003. .09u manufacturing would take it up to 2.8GHz, which again suggests an Athlon of 3700+, but a 20% faster Hammer would be 4400+. >>

he REALLY wants to sound like he knows what he is talking about.



<< That's why I'm skeptical that AMD can reach a 3400+ processor in January next year, given their current forecasts, and still have it outperform a 3.0GHz Pentium 4 by more than 20%. Things just don't add up, and AMD may have to adjust their model number forecasts soon to align with newer developments. >>

what exactly does he know about "newer developments"? jack sh#t.



<< If things continue on this course, then Northwood will move faster, and eventually reach a point where it can outperform a similarly modeled Athlon - especially with the FSB enhancement that will soon be offered. >>

what does this even mean?


sorry to go on a rant like that. I usually just keep my mouth shut, but this guy is blowing so much hot air I just couldnt stay silent.

 

[Rainsford]

I'm not sure I understand what Arcadian is even talking about. It is moronic to expect that AMD wouldn't change their numbering system when faster processors come out. Just taking a formula of "this many mhz equals this many pr ratings" with different CPUs wouldn't make sense.

This Arcadian person is falling victim to a common logical fallacy. He or she assumes facts that have no basis and then makes an argument against those "facts". In fact there is no indication that the AMD PR ratings will remain true to any supposed "pattern" and extrapolating either forwards or backwards is a waste of time.

Arcadian obviously has no idea what he or she is talking about but wants to sound smart. Here's a hint, people that really know what they're talking about always adopt a "wait and see" attitude about things like this. No one can know what will happen in the future, and random predictions just make you look foolish.

And SSXeon5, jeez, can't you at least think up your own arguments against AMD? Come on, posting other people's ideas from other message boards.

 

[Bozo Galora]

the tables HAVE turned

i havent gone intel in roughly two years, but I am now, in a big way

 

[Degenerate]

.... By giving then TURN TABLES.

hehe though i might add that from Ali G
seriously though. If the PR is missleading, i would not be surprised. It is a marketing strategy and marketing is about lieing in away.

 

[SteelCityFan]

<< I think Intel is misleading with their Mhz. At least the pre northwoods.

Regardless of clockspeed there are other factors in crunching 1's and 0's And Intels superior marketing team knows that Mhz sells to most of the flock of computer illiterates out there. While the AMD clock for clock smokes the Intel. why IPC is the answer.

SSXeon5 I remember you at sharky's you're a intel die hard SE forums suck anyhow...

My 2 cents >>



Who cares about clock for clock? It would only matter is AMD could match them in clock speed. They can't come close. Performance is what matters, and AMD is lagging behind.. especially this morning.

 

[BlvdKing]

AMD is not 'doomed'. The PR ratings don't bother me one bit. However, I consider AMD out of the speed race until Barton or Hammer is released, whichever is first. I have a feeling TBred will be a dud in terms of ramping up speed and performance over the Palimino.

 

[Mingon]

Lets not forget that the PR rating is athlon xp vs athlon i.e tbird vs palimino.

 

[Nate420]

I'm not really an AMD fan, but AMD fan or not, you can't deny that the PR rating was a good idea on their part. The Athlon deserved the PR ratings it has when the only CPU for comparison is a P3/P4(as long as the PR is accurate). AMD would have looked really bad had they kept the old real Mhz model numbers. They don't come close to the actual frequencies of a P4, but their performance is close.

One thing that bugs the hell out of me is when I see AMD fans bitching about the Mhz difference between an Athlon and a P4 when the Athlon gets beat, with the PR equaling the P4. AMD created the PR thing, trying to dispel the Mhz-only mentality...In the end, performance is all that matters. They are totally different architectures, if you need 3Ghz to do it... Fine, if you can do it with 2Ghz... Fine.

 

[CoDerEd]

I don't care about the clock or PR rating
all i see is the benchmark and the price and make
a reasonable Price Performance Ratio=PPR.

meaning: i don't want to spend $50 more on INTEL
if it's just has an equal or slower performance (once again... not clock but performance)
than a $50 cheaper AMD and VICE VERSA.

also I don't want to stuck on a same mobo/cpu setup for more than 6mos(well a year top)
so I always see which one has the better PPR and get that one.
I had Athlon XP last year and happy with my nortwood today, i don't know what
will i have next summer. I don't believe their roadmaps... just see it when it arrives.

Just my $.02


<<Keep it cool people. We don't want this to turn into a flame war, and we are watching this one closely.

AnandTech Moderator
>>

Agreed

 

[Mingon]

<< One thing that bugs the hell out of me is when I see AMD fans bitching about the Mhz difference between an Athlon and a P4 when the Athlon gets beat, with the PR equaling the P4. AMD created the PR thing, trying to dispel the Mhz-only mentality...In the end, performance is all that matters. They are totally different architectures, if you need 3Ghz to do it... Fine, if you can do it with 2Ghz... Fine. >>



It seems too many people dont understands the pr rating a 1800xp is equal to a athlon (thunderbird model) running at 1.8ghz and NOT a 1.8 pentium

 

[Necrolezbeast]

<< It seems too many people dont understands the pr rating a 1800xp is equal to a athlon (thunderbird model) running at 1.8ghz and NOT a 1.8 pentium >>




uhhh, I'm not sure what a 1.8ghz tbird would run at cause there is no way to get one to the speed...how can they compare xp to something that doesn't/can't exist? that might be what it says under their explanation, but they meant it to be compared to the P4...

 

[Bateluer]

Just one thing you need to remember with the model numbers. They were not designed to relate to the P4s, or the P3s, or the P2s, or any other Intel proc. They were designed to relate only to the Tbird cores. Comparing them to other procs is pointless.

 

[Emo]

<< Just one thing you need to remember with the model numbers. They were not designed to relate to the P4s, or the P3s, or the P2s, or any other Intel proc. They were designed to relate only to the Tbird cores. Comparing them to other procs is pointless. >>



Sure, and how is that not misleading the average Joe SixPack at BestBuy? They see an AMD XP 2000+ next to a P4 2.0Ghz and what do you think it's more likely to think: Gee, this AMD system is faster than the P4, or Gee, this AMD system is faster than a 2.0Ghz T-bird if it ever existed? AMD is walking on a thin line, IMHO, and I wouldn't be surprised if it gets sued for false advertising.

 

[HendrixFan]

Emo, thats not really the case. Looking at Anand's review of the Athlon XP 2100+ you will find that the XP 2000+ is probably a little better than the P4 2.0A. They go back and forth, but the XP seems to win the most benches.

Of course your own argument about about AMD can be applied to the P4 and the 2.0 versus the 2.0A, and we all know the performance difference. With the new 133fsb P4s, named with "B" at the end it gets even more confusing for Joe sixpack. Shouldnt they be sued for false advertising as well?

Truth is, at any price point AMD XP's beat out the P4 in performance, not considering clockspeed or rating. AMD is the better value, Intel has the highest performance.

 

[jiffylube1024]

The "PR" rating does not have any credibility below 1500+ (the first Athlon XP). The PR rating was not designed to accomodate old chips, it was merely a way of comparing the Athlon XP's performance to the Pentium 4 of similar rated MHz. There is no point talking about the fact that below 1 GHz the PR rating becomes warped; there is no Athlon XP below 1.33 GHz.

AMD is currently backed into a corner (in terms of keeping up with Intel's very high clock frequencies), so they designed a PR rating system with a generous amount of headroom; there is not a single Athlon XP whose PR rating (ie 2000+) runs slower than an actual Intel P4 at the same speed (ie 2.000 GHz).

This PR rating cannot keep up indefinitely; with only a 66.6 MHz increase per 100+ points in the PR system, it is bound to hit a wall where after a certain point, the palomino core's performance does not increase (per 66.6 MHz) as much as a P4 does (per 100 MHz).

To the person who said that Intel's actual MHz rating is misleading; Intel is the one keeping true to their word: a P4 2.2 runs at just that, 2.2 GHz. You could say it is misleading in the sense that it is not as fast as, say, a Pentium 3 2.2 GHz (which does not exist) but Intel is not lying to you whatsoever on the clock speed of the P4, so the argument is moot.

I think that anyone claiming AMD is doomed on the reason that extrapolating the PR rating to speeds it was never designed for is ridiculous. Just my $0.02

 

[jiffylube1024]

Oh yeah one more thing; it seems that other forum has a real Intel bias or something: the last post was talking about if Intel made up a "PR rating" of their own, like calling the P4 2.5 GHz "Performance 5000" or something, and the guy said we would cry bloody murder if they did. Damn right we would! Giving it a stupid name like that is totally misleading! My point is this: with Intel being basically the market leader, AMD designed its PR rating realative to Intel's speed. It is a direct performance comparaison to the Pentium 4; the name "Performance 5000" is not relative to anything.

These guys want to make themselves sound so informed (the guy talking about AMD getting good/bad yields on 2.2 GHz Hammers was one of the funniest!); they are just pulling this stuff out of thier @$$es. Let AMD worry about coming out with the Hammer CPU's; I'm sure they will thoroughly test them and come up with a PR rating that is more than fair in comparaison to Intel's flagship at that time. And if AMD skimps out and overshoots it with thier PR rating, we will always have Anandtech, Tom's Hardware, Sharky Extreme, and many other hardware sites to give us the skinny on just how well all of these next-gen platforms will perform.

 

[SSXeon5]

<< Oh yeah one more thing; it seems that other forum has a real Intel bias or something: the last post was talking about if Intel made up a "PR rating" of their own, like calling the P4 2.5 GHz "Performance 5000" or something, and the guy said we would cry bloody murder if they did. Damn right we would! Giving it a stupid name like that is totally misleading! My point is this: with Intel being basically the market leader, AMD designed its PR rating realative to Intel's speed. It is a direct performance comparaison to the Pentium 4; the name "Performance 5000" is not relative to anything.

These guys want to make themselves sound so informed (the guy talking about AMD getting good/bad yields on 2.2 GHz Hammers was one of the funniest!); they are just pulling this stuff out of thier @$$es. Let AMD worry about coming out with the Hammer CPU's; I'm sure they will thoroughly test them and come up with a PR rating that is more than fair in comparaison to Intel's flagship at that time. And if AMD skimps out and overshoots it with thier PR rating, we will always have Anandtech, Tom's Hardware, Sharky Extreme, and many other hardware sites to give us the skinny on just how well all of these next-gen platforms will perform. >>



They arnt bias thats you

They are right .... MHz does mean MHz .... like i said before this is the greatest thing to happen to Intel ... a core that can clock to 10Ghz and stays at 6 IPC .... I want to see amd do that!

Can people stop that "It seems too many people dont understands the pr rating a 1800xp is equal to a athlon (thunderbird model) running at 1.8ghz and NOT a 1.8 pentium" CRAP go get a 1.33Ghz XP (1500+) and a oced athlon 1.4@1.5 and the Tbird will rape it .... there is no way a 1.53Ghz XP can match a 1.8GHz Tbird

And where is the TBred .... o delayed again ..... and hammer .... 2003 .... ah ok

SSXeon

 

[sandorski]

Well you see, that's what we geeks are for. If AMD's PR rating begins to slag off, meaning for example: A PR 2800 only performs = a P4 2.4ghz(this is only an example, not necessarily what will happen, aka I'm pulling numbers out of my butt ), then it is the job of us who know the truth to slag AMD. If AMD adjusts the PR to better reflect reality(or perceived reality), then there is no problem.

Let me clarify something though. AMD's PR rating *is*based on the TBird's, or would be TBird's, performance. However, it seems that AMD did want to make some connection to the P4, hence their pricing of PR rated cpus at about P4 prices of the actual ghz: ie P4 1.8ghz price is about the same as PR1800 Athlon price. Since this dual nature of AMD's PR rating exists, AMD will eventually need to adjust it's numbers if it doesn't want to begin deceiving the consumer.

For various reasons, I'm an AMD fan and have been since the k6-2 300, I'll likely continue being an AMD fan even if Intel begins to pull away in the performance race. That said, AMD will need to adjust their PR rating eventually to be solely P4 based and not TBird based. This is because, eventually the P4 will likely perform mhz-mhz with the TBird core.

 

[Rainsford]

<< They arnt bias thats you

They are right .... MHz does mean MHz .... like i said before this is the greatest thing to happen to Intel ... a core that can clock to 10Ghz and stays at 6 IPC .... I want to see amd do that! >>



You've just hit on exactly why AMD needed the PR numbers in the first place. Everyone looks at P4s (even techies apparently) and drools over the high clock speeds. What many people don't seem to understand is that clock speed is a stupid way of measuring performance of a CPU. Faster does not always mean better, anyone who's ever worked with Sun or SGI boxes knows this. Unfortunitly, this is the real world and marketing plays a big roll, and most consumers look at one thing, how fast does the CPU run. Intel figured this out and made a CPU that plays to consumers. AMD was left with (at the time the P4 came out) far superior chips that no one would by because they were "slower". So they needed a bigger number to attach to their CPUs.

 

[Rainsford]

<<
Sure, and how is that not misleading the average Joe SixPack at BestBuy? They see an AMD XP 2000+ next to a P4 2.0Ghz and what do you think it's more likely to think: Gee, this AMD system is faster than the P4, or Gee, this AMD system is faster than a 2.0Ghz T-bird if it ever existed? AMD is walking on a thin line, IMHO, and I wouldn't be surprised if it gets sued for false advertising. >>



PR ratings are just a relevant of a speed measuring tool as clock speed. Again, Sun and SGI anyone?

And in your example, The XP would beat the P4 in most apps, so having the consumer think they are equal is certainly not something AMD is doing wrong. Look at almost every product out there, they all have flashy names and misleading descriptions. Look at Kia and their "V6 power" cars. A consumer might think that a V6 Kia is faster than a 4 cylinder WRX. Should Kia be sued?

All the data is available to anyone that looks for more than 10 seconds. If a consumer isn't smart enough to find out all the facts before spending all that money, they deserve whatever they get.

 

[jiffylube1024]

<<

They arnt bias thats you

They are right .... MHz does mean MHz .... like i said before this is the greatest thing to happen to Intel ... a core that can clock to 10Ghz and stays at 6 IPC .... I want to see amd do that!

>>



I'm not biased against Intel; I own and love my P4! I'm just trying to get a point across. The PR rating system for AMD is necessary for non-techies like us.



<< Can people stop that "It seems too many people dont understands the pr rating a 1800xp is equal to a athlon (thunderbird model) running at 1.8ghz and NOT a 1.8 pentium" CRAP go get a 1.33Ghz XP (1500+) and a oced athlon 1.4@1.5 and the Tbird will rape it .... there is no way a 1.53Ghz XP can match a 1.8GHz Tbird

And where is the TBred .... o delayed again ..... and hammer .... 2003 .... ah ok >>



Your anti-AMD bias is evident though: a 1.8 GHz Tbird destroying a 1.53 GHz XP; let's compare actual CPU's, not made up ones please!