Sign in with
Sign up | Sign in
Your question

AMD's admission, Matisaro, are you listening?

Last response: in CPUs
Share
a b à CPUs
August 21, 2002 10:13:50 PM

Quote:
In any case, our recent test comparisons made it clear that the model numbering for the top AMD CPUs did not correspond to the performance of comparable P4s. Because of this, AMD adjusted its virtual values to become lower: previously, a clock increase of 66 MHz meant the equivalent of 100 additional points; now, the 100 points are related to the absolute values. This is AMD's admission that the previous performance scale was set too high, especially when it came to the higher clock speeds

I had this argument with Matisario like 4 months ago, that AMD's XP rating sytem could not scale indefinately when compared to the P4, he said that the 66MHz=100points system would always hold true. Guess who was right?

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
August 21, 2002 10:28:22 PM

Quote:
I had this argument with Matisario like 4 months ago, that AMD's XP rating sytem could not scale indefinately when compared to the P4, he said that the 66MHz=100points system would always hold true. Guess who was right?

Lol sounds like that argument's been stewing under your skin for some time. I'm sure you're the only one who remembers it. The PR system had to be revised eventually anyway because it started at a non-zero value and was hence non-linear compared to the P4.

*However*, IMO the reason it's been revised so early is because the new T-Bred core doesn't perform per clock as well as the old core, so they had to change the PR system. AMD obviously did something to the core to squeeze out a few more MHz at the expense of IPC. The FPU seems as strong as ever but many of the benchmarks indicate that other parts of the core were compromised.

Ritesh
a b à CPUs
August 21, 2002 10:36:17 PM

Quote:
The PR system had to be revised eventually anyway because it started at a non-zero value and was hence non-linear compared to the P4.


Yes! That was the basis for MY argument!

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
Related resources
August 21, 2002 10:57:59 PM

</away>

Did anybody else think it ironic that AMD's auditing firm and PR rating stance changed at the same time?

<away>

<font color=blue>Hi mom!</font color=blue>
August 22, 2002 12:03:24 AM

AMD's new PR system will incorporate fuzzy math 2i, using advanced mathematic calculations and simple marketing manipulations from 100 supercomputers and a car salesman.

my prediction of the athlon XP 8000abM4ggx2.158a6 versus the P4 5GHz looks like it will come true after all!

What I don't understand is that: if amd marketing thinks that people like a higher number (if not, then why can't they just publish MHz and note equal performance to p4 XGHz), why not just use number in the trillions already.

New! AMD Athlon XP 50,000,000,000,000,000+ (at 2.8GHz) is better than XP 3000+, don't you think?

Ok let me be more realistic, cos I know there's alot of thickheaded people here. Why couldn't there be a AMD Athlon XP 8000+ at 2.8GHz. 8000+ is bigger than 3000+, it must be better (lol) and since their PR is going to make no sense whatsoever very soon, why not just blow it out of the water. (i'm sure some fool will buy it if it was 8000+ instead of 3000+ tho)

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 22, 2002 12:08:15 AM

maybe they might want to keep some remote bit of honesty involved :smile: whoknows.


<b>MegaHertz Matters! ... But not without Cache our a decent chipset!!! :cool: </b>
August 22, 2002 12:34:44 AM

Quote:
The PR system had to be revised eventually anyway because it started at a non-zero value and was hence non-linear compared to the P4.

1) The PR system was never meant to be compared to the P4 it was based on a Tbird comparison. Like it or not this was AMD's stance from the get go. Sure, you can say that it was intented to REALLY be used against the P4, but isn't that basically the same silly argument some fools are using to try to sue Intel with?

2) Even if you still want to argue that, when first introduced, wasn't the Willy the p4? How well would this 2600+ look against a Willy at 2.53 ghz?

3) I've never sean a CPU that scaled linearly (?) throughout its lifespan...have you?

This being said, what do you think the Barton with 512 cache and a 333 fsb bring to the table? Perhaps coupled with a better chipset, say Nvidia nforce2?
Suffice it to say AMD is far from dead and has at its disposal a few tricks up its sleave until they are ready with the Hammer. Good for them I say, rather then release a casterated POS Hammer like Intel did with the Willy.

It's not what they tell you, its what they don't tell you!
August 22, 2002 2:55:35 AM

Quote:
1) The PR system was never meant to be compared to the P4 it was based on a Tbird comparison. Like it or not this was AMD's stance from the get go.

Do you really believe that? Honestly? Trident says its PX4 graphics chip will perform 70% of a Radeon9700.....You should get one, because Trident said it's good. Tell Pavlov I said hi.

This sig runs too hot.
August 22, 2002 3:07:30 AM

As much as I believe I need a PEE4 to enhance my internet expeirance.

Quote:
Trident says its PX4 graphics chip will perform 70% of a Radeon9700.....You should get one, because Trident said it's good.

Crazier things have happened. SIS is making good chipsets. At least maybe they will have better driver support. Oh, and you can tell Homer I said hi as well.

It's not what they tell you, its what they don't tell you!
August 22, 2002 3:27:47 AM

"As much as I believe I need a PEE4 to enhance my internet expeirance."

Then why make a point about AXPs PRs being in comparison to the birds, when we all know its to market against Intel? If you don't even believe it yourself...

"Crazier things have happened. SIS is making good chipsets. At least maybe they will have better driver support."

Trident cheapo boards competing performance wise to Geforce4 Ti's sure is crazy.

This sig runs too hot.
August 22, 2002 3:46:11 AM

Dude, Anand tested the XP4 T2, which was not even finished, with Beta drivers, was about 80% the performance of a Ti4200, which is expected. The T3 was not tested, but coupled with finished drivers, I doubt their 80% is way off, in fact I think it will eventually be true.

I beleive the fact that a 2.6GHZ Tbird would not beat the 2.53GHZ, wouldn't you?
It would in FPU, no doubt, but since it has less IPC than the Palomino, this means that the P4's added cache and bandwidth obviously CAN beat the Tbird per clock. So I don't think the XP2600 is THAT much exaggerated.

To Crash, doesn't 'AMD's admission' mean that AMD admitted they were way off? If so, aren't you the one who respected companies that ADMIT their mistakes?

--
Is the opportunity to earn money by working, free?
August 22, 2002 3:56:03 AM

Quote:
The PR system was never meant to be compared to the P4 it was based on a Tbird comparison.

That's bulls_h_i_t but AMD likes the fact that you believe their marketing. No matter what AMD says on their web site, everyone knows that the PR was developed purely as an arbitrary formula to try to boost the image of the lower clocked Athlons against the P4, because megahertz sells. The vast majority of the market uses Intel so they never knew or cared how the T-bird performed. Therefore the PR figure would be completely meaningless to them if it was based on the T-bird.

Let's put it another way. Assume that the P4 had been clocked twice as fast but had half the IPC. You can bet your right arm that AMD would have made their PR twice as high: AXP 4000+ for a 2 GHz processor. And they would have fed you some crap about the PR being based on their 386 clone. And you would have believed it wouldn't you...

Ritesh
August 22, 2002 4:41:48 AM

Exactly. But something I don't understand is why will they continue it with the Hammer since it's supposed to be able to compete with the P4 in the numbers department.

Knowledge is the key to understanding
a b à CPUs
August 22, 2002 4:50:58 AM

1) Matisario made the arguement in favor of the viability of XP numbers vs P4 speeds
2) The Northwood was already out when HE argued the XP rating system would remain the same and remain viable for all future speeds.
3) I tried to argue that, but HE denied it.

So you THINK you're arguing AGAINST me, but in reality you're simply VARIFYING WHAT I SAID.
3)

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
August 22, 2002 5:03:41 AM

i wasn't around for that argument, but i like to add this:
(and maybe this is exactly the same as everyone is pointing out.. but anyways..):

The current line of chips is based on a core codenamed "Thoroughbred." An older version of the Athlon-branded chips was based on a core codenamed "Thunderbird." The numbers attached to Thoroughbred-based chips officially describe how their performance would compare to a chip made on the Thunderbird core with the same clock speed-that is, if such a chip had ever been made, which it wasn't. So the Athlon XP 2600+ performs like an older Athlon with a clock speed of 2.6GHz would have.
... while that supposedly is the real reason for the PR+ and amd will never admit this, but the PR (2600+) is actually aimed squarely at intel's current flagship, the p4 2.53GHz.

Intel is rumored to have a 2.8GHz P4 launch as soon as next week as a warm-up for its 3.0GHz in the christmas season. goodnite amd, happy hibernation (mode)..we'll see you next spring.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
a b à CPUs
August 22, 2002 5:43:41 AM

LOL, I would LOVE to see an XP2600+ perform twice as many operations per second as the Tbird 1333!

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
August 22, 2002 6:12:59 AM

Quote:
had this argument with Matisario like 4 months ago, that AMD's XP rating sytem could not scale indefinately when compared to the P4, he said that the 66MHz=100points system would always hold true. Guess who was right?


Thats a blatant mischaracterization of my comments.

I dare you to quote me saying it would ALWAYS hold true, in fact I said that they would have to reset them eventually because it isnt perfect.

Im ashamed of this post crash.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 22, 2002 6:15:10 AM

PS: My name is Matisaro, not matisario, weve ben posting together for almost 2 years crash lol.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
August 22, 2002 8:52:26 AM

Quote:
That's bulls_h_i_t but AMD likes the fact that you believe their marketing. No matter what AMD says on their web site, everyone knows that the PR was developed purely as an arbitrary formula to try to boost the image of the lower clocked Athlons against the P4,

I never disputed this fact. Of course it was a marketing scheme by AMD. At the time the Tbird was performing on par with the p4 clock for clock, if not better. If a=b and b=c, then a=c......got it?

Quote:
because megahertz sells

I see you have bought into Intels marketing.

Quote:
Let's put it another way. Assume that the P4 had been clocked twice as fast but had half the IPC. You can bet your right arm that AMD would have made their PR twice as high: AXP 4000+ for a 2 GHz processor. And they would have fed you some crap about the PR being based on their 386 clone. And you would have believed it wouldn't you...

Just exactly what is your point? Are you trying to say that AMD is more guilty of misleading advertising then Intel? You sure you want to make that argument?



It's not what they tell you, its what they don't tell you!
August 22, 2002 9:00:38 AM

Quote:
Then why make a point about AXPs PRs being in comparison to the birds, when we all know its to market against Intel? If you don't even believe it yourself...

You answered your own question. It was a comparison to the tbird for the purposes of marketing against the p4. Never did I state this was not a marketing tool. It really comes down to a question of legality. How Joe Sixpack interprets it is entirely a different matter. The fact that AMD knew if would be percieved as a direct comparison to the P4 is hardly any worse than Intel knowing that the vast majority of its users expect a 1.4 ghz part to outperform a 1.2 ghz part.


It's not what they tell you, its what they don't tell you!<P ID="edit"><FONT SIZE=-1><EM>Edited by ncogneto on 08/22/02 05:11 AM.</EM></FONT></P>
August 22, 2002 9:10:29 AM

Quote:
So you THINK you're arguing AGAINST me, but in reality you're simply VARIFYING WHAT I SAID.

No, just clarifying. I never saw the argument you had with Mat. If that was indeed his argument it was a bit flawed. I do see however he doesn't agree he claimed it would remain linear. I don't see how it could, as the p4 has gone from willy to northwood @400 fsb to northwood @533 fsb.

It's not what they tell you, its what they don't tell you!
a b à CPUs
August 22, 2002 9:27:40 AM

Help me out here, I can't find the argument!

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
a b à CPUs
August 22, 2002 9:28:33 AM

LOL

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
a b à CPUs
August 22, 2002 9:46:56 AM

I looked for 10 minutes and gave up, maybe Matisaro will find it.

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
August 22, 2002 12:19:31 PM

Uhm...even AMD stated when they introduced their PR rating system that it was only a temporary ratings system UNTIL a ratings system agreed upon by the industry as a whole agreed to a set of performance ratings for CPUs. So, with that being the case I doubt Mat would say the PR would ALWAYS hold true.

Mark-

PS: As for a 2600 having double the IPC in benchmarks when compared to a T-bird....since performance isn't linear, and since there are other bottlenecks to system performance, like memory bandwidth, bus speed/bandwidth, hard drive speed and throughput, etc....it's doubtful that a 2600 would perform double the TBird.

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>
August 22, 2002 2:13:29 PM

I just have a couple things to say on this matter:

1. When AMD first introduced the XPxxxx+ rating system, they said it was to compare vs a t-bird. Now why in god's name would they do that, unless they were feeling pressured by their actual clock speed compared to the P4's clock speed, which should be able to scale much better than the AXPs, due to long pipelines and other considerations. Couldn't Intel just say, "Hey, we're gonna base our P4 rating system on a comparison vs a 286, so our P4 2.53 will now be called P4 EX1000000000000000+++++". So, just think for a moment. When AMD says the PR rating is to compare vs the t-bird, why would they do that? It's totally pointless to do that, but in comparison to P4s, it looks much better than the actual clock speed.

2. I'm sure for most consumers, they've either never heard of the reason for the ratings system, and I'm sure many salespeople have said that a XP2000+ is equivalent to a P4 2GHz. With that in mind and through the time that has gone by, most people just think: XP2000+ = P4 2GHz, and not XP2000+ = t-bird 2GHz. Now, looking at the THGC tests, showing an o/ced t-bred to XP3000+ and XP3400+, it clearly shows that they both can't reach the P4 2.8GHz's performance. Now, for a consumer who sees XP3000+/XP3400+, they'd probably automatically think, well that's like a P4 3GHz/3.4GHz. But no. Actually when looking at the charts, the XP3400+, which is a AXP 2.8GHz, actually equals the P4 2.8GHz. Wow, finally, the P4 has caught up clock per clock. But now, AMD's PR rating totally looks overinflated huh, even though by AMD, it's only supposed to equate t-bird performance at 3.4GHz.

So, what's Intel gonna do when this happens? Create another ratings system, and say it's suppposed to recreate the Willamette's performance with SDRAM or something? So really, the AMD rating is gonna be in trouble very soon, when we start seeing a P4 and AXP perform clock per clock, the extra couple hundred PR points are gonna be just pure BS.

So, maybe we'll have lawsuits over AMD's raing, just like we have people sueing Intel for wrong impressions of performance. AMD'll say that it's vs a t-bird, but now I ask, why in the first place did they need this? Think about it.

<i>Past mistakes may make you look stupid, but avoiding future ones will make you look smart!</i>
August 22, 2002 4:35:23 PM

because the Palomino generally performed about 10-15% better clock for clock than a Tbird. Better performance generally commands higher prices. However, if AMD had sold solely by clockspeed they would have had lower ASPs yet better performance. It makes perfect sense for marketing reasons.

Mark-

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>
August 22, 2002 4:50:40 PM

But one question still remains:

<b>Why can a 1400MHz T-Bird outperform an AXP 1500+?</b>

It kind of kills AMD's whole PR rating excuse when it doesn't even hold true to the first AXPs that used the PR rating. I'm actually kind of surprised that lawsuits never occured over it.

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 22, 2002 4:59:10 PM

Quote:
2) Even if you still want to argue that, when first introduced, wasn't the Willy the p4? How well would this 2600+ look against a Willy at 2.53 ghz?

Funny how I used this very logic way back when I expressed how disappointed I was with AMD's PR rating listed for the launch of the Hammers. Now other people are finally backing up my points from back then that AMD's PR rating <b>doesn't</b> scale linearly to a P4's performance, which made the estimated performance of the inital Hammers sound rather pathetic for all the improvements that the Hammers are supposed to have. Where were you then when I was being jumped all over by AMD fanatics? :( 

(And has anyone actually changed their point of view over that particular issue yet? Or am I still alone in that state of disappointment?)

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 22, 2002 5:20:44 PM

why would the typical person in this forum even look at the PR ratings? It's just as bad to throw MHz numbers out there because this is not a true rating of performence either. Just check out some other designs (G4s, Sun chips) and its pretty freakin obvious that MHz aren't the most important factor. I think most people here know to look at benchmarks and judge from that. To have a an unbiased value to rate processors' performence would be nice....but we don't have that yet.

*** :cool: Duff Man says a lot of things, OH YEAH!!! :cool: ***
August 22, 2002 5:32:11 PM

To all of you.
The PR rating was against Willamette.
You should all know that.

Be honest.
August 22, 2002 5:42:40 PM

Quote:
Funny how I used this very logic way back when I expressed how disappointed I was with AMD's PR rating listed for the launch of the Hammers.

Are you saying that intels use of pure mhz numbers are an accurate indication of performance? Do you want to compare a p4 @ 1800 mhz to a pally at 1800 mhz? Its painfully obvious that AMD had to do something. What should they have done?

Quote:
Now other people are finally backing up my points from back then that AMD's PR rating doesn't scale linearly to a P4's performance,

How can it when the p4 has changed several attributes? As another has stated, it is only one part of a system, it can never possibly scale in a linear fashion unless other advances to other attributes are made as well. The p4 has changed both its FSB and its l2 cache since the PR rating system was placed in effect. Both major improvements to the chip. Meanwhile the pally has done neither.

a= tbird
b= p4 willy
c= pally

This was what was in place at the begining of the PR rating Scheme. It was generally accepted that a=b give or take a few points depending on what benchmark was used. Here is where everyone makes thier misunderstanding. Amd stated that the new pally with its PR rating would reflect to a tbird. Thus a 1600+ pally would be equivolent to a 1600 mhz tbird which would be equal to a 1600 mhz p4 (at that time). Here an argument could be made that although AMD stated it was supposed to reflect on the performance compared to the tbird, they knew it would be applied to the p4 as well. I have never disputed this fact. However, when they choose to use the tbird they choose a constant, they could never choose to base it on the p4 for the very reason it was not a constant. B has changed and A no longer equals b.

Quote:
which made the estimated performance of the inital Hammers sound rather pathetic for all the improvements that the Hammers are supposed to have.

Just speculation and not worth argueing about IMHO.

Quote:
Where were you then when I was being jumped all over by AMD fanatics? :( 

Taking a break sick and tired of all the useless AMD vs Intel BS from people ( most, not all mind you) that knew little of what they were talking about.

It's not what they tell you, its what they don't tell you!<P ID="edit"><FONT SIZE=-1><EM>Edited by ncogneto on 08/22/02 01:45 PM.</EM></FONT></P>
August 22, 2002 5:53:43 PM

Now let me ask you this. As the 3 gig p4 struggles to get out the door by the end of the year, and as it apears yet another new motherboard has to be made to support it, isn't the life expectancy for the p4 platform getting a little short? Do you think it will still scale to 10 ghz as a one Mr. Raystonn so addimatly claimed? How much higher than 70 watts can it go? Sure, .09 micron will help a little but there comes a point of dimishing returns as far as die shrinks go. Why don't the zeons have a quad pumped 133 fsb yet?

It's not what they tell you, its what they don't tell you!
August 22, 2002 7:04:13 PM

Quote:
Now let me ask you this. As the 3 gig p4 struggles to get out the door by the end of the year, and as it apears yet another new motherboard has to be made to support it, isn't the life expectancy for the p4 platform getting a little short?

sorry, but 3GHz will kick some major butt! if 2.53 holds its down against amd's flagship the 2600+, imagine what 3GHz can do! ..and i rather intel "struggle" to get it out the door than lay hibernating until spring...prescott will happily take over from 4-5GHz on.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 22, 2002 7:45:09 PM

Quote:
Are you saying that intels use of pure mhz numbers are an accurate indication of performance?

Did I even suggest such? Hell no. However you have to at least ponder that this very innacurate form of measurement is hardly any different than the supposedly superior method that AMD came up with to counter it.

Quote:
Its painfully obvious that AMD had to do something. What should they have done?

In a perfect world? Spend even a teeny-tiny amount of money to actually educate clueless consumers through the use of friendly and informative commercials and become consumer and manufacturer-friendly instead of obfuscate through artificial inflation and gestapo their bizzare antics to manufacturers to hide any part of the truth.

In an imperfect world? Come up with a standardization for measuring CPU performance that <i>doesn't</i> depend upon theoretical (but practically disproven) relations to prior hardware and push this measurement to all hardware reviewers and resellers. After all, industry standards can only become standards when someone both creates and uses them.

Quote:
This was what was in place at the begining of the PR rating Scheme. It was generally accepted that a=b give or take a few points depending on what benchmark was used. Here is where everyone makes thier misunderstanding. Amd stated that the new pally with its PR rating would reflect to a tbird. Thus a 1600+ pally would be equivolent to a 1600 mhz tbird which would be equal to a 1600 mhz p4 (at that time). Here an argument could be made that although AMD stated it was supposed to reflect on the performance compared to the tbird, they knew it would be applied to the p4 as well. I have never disputed this fact. However, when they choose to use the tbird they choose a constant, they could never choose to base it on the p4 for the very reason it was not a constant. B has changed and A no longer equals b.

But it was also very plain from the beginning by comparing a 1.4GHz T-Bird to an AXP 1500+ that AMD was full of XXXX.

Quote:
Just speculation and not worth argueing about IMHO.

While I agree that it is both speculation and not worth arguing about, it none-the-less does make AMD's claims of a Hammer that has the performance of a 3.4GHz T-Bird rather depressing. Obviously it can't really be argued either way. Still, it's something to at least think about. It also leaves one hoping that AMD didn't buy their own marketting BS about their pr-rating being related to the T-Bird any more than anyone else did.

Quote:
Taking a break sick and tired of all the useless AMD vs Intel BS from people ( most, not all mind you) that knew little of what they were talking about.

It seems as though relatively little has changed over time. :( 

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
August 22, 2002 8:28:42 PM

Quote:
Did I even suggest such? Hell no. However you have to at least ponder that this very innacurate form of measurement is hardly any different than the supposedly superior method that AMD came up with to counter it.

I agree here. Both methods are totally screwed. Advantage no-one!

Quote:
In a perfect world? Spend even a teeny-tiny amount of money to actually educate clueless consumers through the use of friendly and informative commercials and become consumer and manufacturer-friendly

Without speculation into what a perfect world would imply, that would only increase the cost of their CPU's. I think with that in mind I would rather stick with what they have however inperfectit may be.

Quote:
instead of obfuscate through artificial inflation and gestapo their bizzare antics to manufacturers to hide any part of the truth.

No more so than Intel however be it by a differnet means.

Quote:
In an imperfect world? Come up with a standardization for measuring CPU performance that doesn't depend upon theoretical (but practically disproven) relations to prior hardware and push this measurement to all hardware reviewers and resellers. After all, industry standards can only become standards when someone both creates and uses them.

An initative they are trying to take. The PR was only meant to be temporary.

Quote:
But it was also very plain from the beginning by comparing a 1.4GHz T-Bird to an AXP 1500+ that AMD was full of XXXX.

Linkage? Ihave never seen any documentation to support your claim although I am not saying it does not exist.

Quote:
it none-the-less does make AMD's claims of a Hammer that has the performance of a 3.4GHz T-Bird rather depressing.

I really don't find an intial release of the hammer performing at rates comparable to a 3.4 gig t-bird depressing. It would all depend on its scalability. Then again its all personal opinion.

It's not what they tell you, its what they don't tell you!
August 22, 2002 8:31:53 PM

the 1500+ and t-bird 1.4 performe similarly. The XP wasn't good enough to warrent the 1500+ though.

The Boogie Knights: Saving beautiful monsters from ravoning princesses since 1983.
August 22, 2002 8:32:45 PM

the 1500+ and t-bird 1.4 performe similarly. The XP wasn't good enough to warrent the 1500+ though.

The Boogie Knights: Saving beautiful monsters from ravoning princesses since 1983.
August 22, 2002 8:32:46 PM

Quote:
Did I even suggest such? Hell no. However you have to at least ponder that this very innacurate form of measurement is hardly any different than the supposedly superior method that AMD came up with to counter it.

I agree here. Both methods are totally screwed. Advantage no-one!

Quote:
In a perfect world? Spend even a teeny-tiny amount of money to actually educate clueless consumers through the use of friendly and informative commercials and become consumer and manufacturer-friendly

Without speculation into what a perfect world would imply, that would only increase the cost of their CPU's. I think with that in mind I would rather stick with what they have however inperfectit may be.

Quote:
instead of obfuscate through artificial inflation and gestapo their bizzare antics to manufacturers to hide any part of the truth.

No more so than Intel however be it by a differnet means.

Quote:
In an imperfect world? Come up with a standardization for measuring CPU performance that doesn't depend upon theoretical (but practically disproven) relations to prior hardware and push this measurement to all hardware reviewers and resellers. After all, industry standards can only become standards when someone both creates and uses them.

An initative they are trying to take. The PR was only meant to be temporary.

Quote:
But it was also very plain from the beginning by comparing a 1.4GHz T-Bird to an AXP 1500+ that AMD was full of XXXX.

Linkage? Ihave never seen any documentation to support your claim although I am not saying it does not exist.

Quote:
it none-the-less does make AMD's claims of a Hammer that has the performance of a 3.4GHz T-Bird rather depressing.

I really don't find an intial release of the hammer performing at rates comparable to a 3.4 gig t-bird depressing. It would all depend on its scalability. Then again its all personal opinion.

It's not what they tell you, its what they don't tell you!
a b à CPUs
August 22, 2002 8:38:28 PM

I think it will go well over 4GHz on the 9nm process, but probably won't make 6GHz on that one (it may, but that's reaching a little). Then they have another die shrink comming if I'm not mistaken, but I don't think 10GHz is likely by any stretch of the imagination without a complete overhaul of the core.

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
August 22, 2002 9:28:55 PM

You're talking about the NetBurst core, right?

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
a b à CPUs
August 22, 2002 10:43:41 PM

I don't like slogans, I just call the core by it's name. According to Intel, even the Willy had "Netburst", and we saw how that performed.

<font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
August 22, 2002 11:45:40 PM

10 GHz is merely a guess assuming there are no limitations from processing technology, heat or power consumption. Since all 3 of these areas can be improved upon with later designs, it is feasible that the core design is capable of such and possibly even beyond. I'm not exactly sure what NextGen originally intended their core design (which is in the K7) to be. I'm guessing with no other limitations, maybe 5-6 GHz. Maybe a little less with future core enhancements.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 23, 2002 12:51:44 AM

can amd simple add deeper pipeline stages, say in the ALU, L1 cache, L2 cache, branch misprediction? i've read that intel can hypothetically run at twice frequency if pipeline in branch misprediction was lengthen from 20 stages to 52 stages... but deeper pipelines possess the problem of extreme complex algorithms.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 23, 2002 12:57:18 AM

Extending pipeline stages isn't something "simple". It took Intel close to 5 years to get a working P7 core design that actually performed worth a crap. And this is Intel, as in $5 billion for a new testing machine Intel. Designing a completely new core design from scratch (which is what would be required) is no small feat. Simply "extending" the pipelines would not be so easy. You have to coordinate each extra stage, not to mention balance out at which point you separate each function, then you'd have to make sure that you can effectively split the function into 2 iteration, not to mention compensate for branch mispredict, signal skew, etc. etc. etc. It's not "simple" at all.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 23, 2002 1:39:28 AM

yes i know imgod2u, i didn't mean extending pipeline is simple.. what i meant to say is, is there anything in amd's future to do so? i hate to see amd being another apple, but maybe 10 years from now.. we'll be fighting over intel vs nvidia cpu's and occasionally snicker about the apple vs amd cpu's.

<font color=green> there's more to life than increasing its speed -Ghandi</font color=green>
August 23, 2002 1:54:34 AM

Well, asside from K8, I've not heard of any future CPU design. It is very possible AMD is waiting to see whether IA-64 will take over and effectively try to liscense it from Intel. Or they could be working on something revolutionary, who knows.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
August 23, 2002 2:20:45 AM

What I was thinking is that if they get the OEM market as stated in recent news, and the network corps use its Opteron for servers, they might gain money and finally invest R&D to something brand new.

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 2:21:53 AM

No I was thinking you were talking about K7.
BTW NetBurst IS the core architecture name. Unless someone can dig P7 out of Intel's website, regarding the P4, it stands as NetBurst and will be until 10GHZ and a new architecture.

--
And now, an advice from your friendly Nike shoes slogan: JUST DO HER!
August 23, 2002 4:46:39 AM

Netburst is the marketing slogan. Like "quantispeed" or "Altivec". The core name is what the engineers use. Because we don't see cores, we see processors based on those cores.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.
!