Intel "insults INQUIRER readers' intelligence"

Page 5 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
That's the key, AMD knew that they could explain the thing 'till they were blue in the face and people would still take the PR as MHz. The average person has an attention span of about 3 seconds when it comes to that stuff. AMD knew it, and took advantage of it.
Click to expand...
Okay, let's say that's true (although I don't think it is). Some customer who doesn't know his ass from a hole in the ground walks in and buys an Athlon system thinking that the PR is the same as MHz. How is that any worse than customers buying P4 systems based solely on the MHz, without considering real performance? You say AMD is counting on confusing customers, but Intel is doing the same thing. Intel just isn't getting dinged for it because in the past MHz was a generally reliable measure of performance among x86 chips. Now that it's not, Intel still wants to use it even though there are other factors to consider, such as IPC. God forbid that AMD should do something to change the holy MHz rating. I'm sure Intel is getting a good laugh at all the flak AMD is taking for the PR system, while it sells its Pentium 4 2GHz/PR1600+ chips.

<i>There are two theories on arguing with women. Neither one works.</i>
 
The most logical thing I can say is that if they want to rate a CPU by another standard, they should use a quantitave one (like Gigaflops) so that everyone is playing by the same rules.

What's the frequency, Kenneth?
 
quantitave one (like Gigaflops)
Click to expand...

Then intel would complain because it requires sse2 to shine doing flops, etc etc etc. I feel that amds performance based system is pretty good, a suite of benchmarks and the performance is averaged, just because the actual numbers were converted from relative tbird scores dosent invalidate their testing.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
 
I agree, everyone should be playing by the same rules. There should be some way of measuring processor performance that's determined by an impartial body not controlled by Intel or AMD. It's not fair to judge solely on MHz and the AMD PR system is imperfect at best.

<i>There are two theories on arguing with women. Neither one works.</i>
 
"The most logical thing I can say is that if they want to rate a CPU by another standard, they should use a quantitave one (like Gigaflops) so that everyone is playing by the same rules."

Funny, that's basically wat AMD said several times. They went on to state the PR rating is only in place until performance quantification across the board can be settled upon (a standardized performance rating) In fact, I think AMD is one of the major companies promoting this initiative to standardize performance ratings among chips.


When all else fails, throw your computer out the window!!!
 
Well, I disagree with your reply to my post. I think my version of your analogy is more appropriate. I don't think anyone's mind will be changed though.

Anyway saying that AMD is lying by giving their 1.66 Ghz a PR rating of 2000 is a bit of an overstatement. If they labeled it at 2000 mhz then it would be an out and lie.
The PR rating scheme is flawed because there is no comparable rating for an intel CPU.


"It's like this...a 350 Chevy makes more power than a 351W Ford. But you don't see Chevrolet calling it a 400, they instead give you real horsepower numbers.
Or compair a Porche to a Chevy Lumina, same sized engine, the Porche makes about 3 times the power. But they don't lie about engine size either."



Well if AMD listed their 1.667 Ghz CPU as 2.0 Ghz then I would agree with your analogy. But they give it a PR rating. The PR rating can be seen as their version of a horsepower rating. The degree to which that will confuse customers is a worth debating. I don't, however, agree that they are blatantly lying.


"I like AMD's products, but think their marketing shenanagens will come back to bite them BIG time. Especially when you consider this: under there PR scheme, a processor gets 100 PR's for every 66MHz. At that scale, an XP1000+ would have REALLY been 1000MHz, but an XP2000+ is only 1666MHz. That means when the PR goes up 100%, the performance only goes up 66%! A 4000+ would only be 3000MHz!"



There seems to be some miscomunication.
We agreed that CPU speed doesn't reflect the performance anymore but in your above statement you're saying that the CPU cycles of the Athlon XP equals Performance. "As the PR rating goes up 100% the performance only goes up 66%"? As the PR rating goes up 100% the clock cycles go up 66% not the performance. How much the perfomance goes up isn't measured by CPU speed in Mhz/Ghz anymore. Right?
AS I said the PR rating as an indicator of performance is debatable.

Rating a CPU by how many FLOPs it performs would be a better as you said but there has already been a reply to that statement.
 
Well, I can garuntee that the XP2000+ would NOT perform TWICE as many operations per second as the theoretical XP1000+, it would only perform 66% more operations per second, so the performance would only be 66% greater if you measure performance as work over time.

What's the frequency, Kenneth?
 
Well a 2.0 Ghz P4 isn't twice as fast as a 1.0 Ghz either.
Anyway isn't the 1.66 Athlon XP rated at 2000+ PR supposed to be the equivelant performance that you would get from a Thunderbird at 2.0 Ghz? That's what some people have said the PR rating is.

"At that scale, an XP1000+ would have REALLY been 1000MHz, but an XP2000+ is only 1666MHz. That means when the PR goes up 100%, the performance only goes up 66%! A 4000+ would only be 3000MHz!"


I wouldn't expect a 1.666 Ghz processor so perform more than a 1.0 Ghz Processor anyway. I just said that you seem to be freely interchanging performance with Mhz in your reply to my post as in your statement above. *shrugs*
 
Can't use the AXP model numbers to predict AXP performance when scaling the frequency. The AXP model numbers indicates performance levels relative to the Athlon Thunderbirds, not against itself.
 
I just said that you seem to be freely interchanging performance with Mhz in your reply to my post as in your statement above. *shrugs*
Click to expand...

So you're admitting that the PR rating is less acurate (read, less HONEST) than the MHz rating when comparing a certain processor to the same model at a different speed? That's what I thought!


What's the frequency, Kenneth?
 
ahhhh. the contundrum of rating a CPU.

mhz used to be the way to go... when all were around the same, but now that amd and intel are diverging where does one go?

i read some whacky article about how the us government might regulate things if the differences get too different.
kinda like a industry standards thinggy.

wierd, but possibly useful.

happened elsewhere... lightbulb wattage is rated using specialised standards for brightness, heat output, current draw etc
car engines use CC, but how one rates horsepower can vary on the method...
unless standards are inplace.

and its gonna be increasingly important once AMD and intel diverge more with the introduction of hammer

The lack of thermal protection on Athlon's is cunning way to stop morons from using AMD. :)
 
Your claim that AMD is lying is invalid. The rating reflects performance, and it did deliver the promised performance. If a XP 2000+ only performs like a P4 1600 or something, then it is lying. The customer only cares about performance. They ask for Mhz because they thought Mhz = performance. And there is no good way of saying how much performance other than the Mhz. So if they ask for a 2Ghz machine, they are really asking for something at the level of 2Ghz's performance. They are skeptical of lower Mhz processor performing at same level because Intel's advertisements constantly 'educate' them that high Ghz equals high performance. So as long as a customer's new XP 2000+ computer performs just as well if not better than his neighbor's P4 2000Ghz, then he is happy, and he won't feel cheated. If the customer only want Ghz, do you think he will be happy to have a million Ghz machine that takes an hour to do a simple addition? The customer gets what he really want, this is the only thing that matters. As long as this statement is true. There is no cheating.
 
You fail to understand that you can deceive people without lying. You'd make a poor lawyer, AMD's marketing department is probably made up of the slick lawyer types.

As for whether they really are lying, why don't you overclock a T-Bird to 2GHz and find out if it's really the same speed as an XP2000+?

What's the frequency, Kenneth?
 
Maybe the government SHOULD come up with a standard, but then one company would offer additional features not measured by the standard, throwing more crap in the pile.

Yes, lightbulbs are rated by wattage and lumens.

Horsepower used to be measured at the crankshaft, but there were so many transmissions and differentials with such a great degree of efficiency that the government stepped in and made companies go with drive wheel horsepower. Before that companies would go so far as to remove the WATER PUMP before the test to boost their HP rating, running water through the block from tap!

What's the frequency, Kenneth?
 
"... Your really beating a dead horse if you insist that <b>two different speeds of the same processor can't be directly compared against each other.</b> ..."

I never said that. Model numbers don't indicate absolute peformance of the AXP itself when scaling the frequency. They indicate performance relative to the TBird.
 
You fail to understand that you can deceive people without lying. You'd make a poor lawyer, <b>AMD's marketing department is probably made up of the slick lawyer types.</b>
Click to expand...
You are making things up which probably has little basis in reality.
 
MHZ alone is not good - its as usless as made up numbers becouse preformance is:

Clock Speed * IPC = Preformance

Clock Speed would just hang thier without IPC and sence we dont use IPC to rate a processors MHZ alone is USLESS.
we can say that a given processor will do twice the instructions at twice the clock speed.
but instructions is IPC.
 
After rereading some of the sentences in your replies to my post it seems pretty obvious that you have to have the last word.

"So you're admitting that the PR rating is less acurate (read, less HONEST) than the MHz rating when comparing a certain processor to the same model at a different speed? That's what I thought!"

Actually, yeah it is flawed, but it's not really less honest than a direct Mhz to Mhz comparison because an XP rated at 1700 PR compares pretty favorably against a P4 at ~1700 Mhz from the benchmarks I've look at. Now a Mhz to Mhz of the the those two processors would lead people to believe that the Intel CPU runs applications a good deal faster than the AMD CPU. However we know that an Athlon XP will run pretty much neck and neck in most applications with a P4 that runs 150 to 200 Mhz faster. So a direct Mhz to Mhz comparison of AMD's and Intel's latest chip designs aren't really all that accurate either...



You're right; it is possible to deceive people without lying.
If AMD is "deceiving" people by using the PR rating, then Intel is doing so by using a layperson's assumption that Mhz is final word in how fast a CPU is as a marketing ploy and pumping up the clock cycles of their CPU while it performs less IPC's than the AMD's CPU.



"Which is exactly why it is NOT a flawed comparison!"
"That's what I thought!"

Frankly those statements are rather close in tone to an adolescent saying something along the lines of "Nyah Nyah!" or "In your face!" or "Bite Me!"

<edited for spelling>
<P ID="edit"><FONT SIZE=-1><EM>Edited by Skull_Angel on 01/27/02 01:20 PM.</EM></FONT></P>
 
Of course Intel is deceiving people. I've ripped into them several times about the P4 1.4GHz being about the same speed as the PIII 1000EB. So youre take on this is that one companies dishonesty excuses the other's?

What's the frequency, Kenneth?
 
I believe so, for the benchmark baseline that AMD uses to calculate a PR rating.

<A HREF="http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/model_number_methodology_MP_v6.pdf" target="_new">AMD Athlon™ MP Processor
Benchmarking and Model
Numbering Methodology</A>
 
Not at all.

AMD is trying a simple PR rating to give the consumer an idea of how well their CPU performs overall. Now it wasn't concieved by an outside party and there are better ways to measure the XP's performance. Is AMD using a rating that can be called misleading? Yes. Is it dishonest? As much as Intel using Mhz to give an arguably overinflated idea of the P4's performance; to that degree the PR rating can be called dishonest. It also seems that AMD has tried to be a little more conservative with the PR rating, So I wouldn't call the company out and out liars anymore than I would call Intel a liar.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom