Sign in with
Sign up | Sign in
Your question

I should stop reading the article conclusions

Last response: in CPUs
Share
May 13, 2003 10:37:03 AM

the author of "High-Flying:
AMD Athlon XP 3200+ Squares Off Against Intel P4 3 GHz" pretty much prooved he needed his head surgically removed from his hide-quarters.

I'm sorry, but while the 3200+ does not perform on par with the P3Ghz (closer to the 2.8Ghz on average), he totally misses the point. They have to be consistant within the architechture, and use a consistant numbering scheme. Yeah, It is misleading, but then again, I bet Intel NEVER mislead anyone in their life. The way THG keeps changing their views and going all out, they are like fanboys that don't stick to one side.

I swear I should only come here for these boards, the articles are just completely out there some times. (well at least the conclusions, I feel like only looking at the benchmarks).

Athlon XP 1600+, MSI K7T PRO2 RU (POS), 2x256 MB CRUCIAL PC2100 CL2.5 memory, Asus V6800 DDR Delux (GF 256) video card, 6.4GB+27GB WD HD, 40GB IBM HD (all 7200RPM). My computer is an acronym
May 13, 2003 3:45:52 PM

I actually agree with the conclusion. Maybe AMD claims that they are trying to be consistent with the model numbers and relating them to the old thunderbird. The problem is that the whole reason for introducing the model numbers in the first place was to better compare with the P4. If Intel would have found a way to improve upon the P3 instead of launching the P4, there would have been no need.

I just look at the numbers. I already thought the Barton 3000+ was rated too high. Now the 3200+ is faster by a whole 33MHz (that takes into account the faster FSB). I thought we were done with 33MHz jumps a long time ago.
May 13, 2003 4:14:07 PM

Hm... it´s not the first time that a conclusion from THG seems a little off...

One thing they unfortunately can´t conclude in this case, however, is that the 3200+ is not a disappointment...
Related resources
May 13, 2003 4:28:22 PM

Actually for a processor that runs at 2.2 GHz, the processor performs pretty well. I find the rating system to be much more of a dissapointment than the cpu. It is because of the rating system that you expect it to perform so much better than it does. And with AMD marketing it as the fastest desktop CPU, it just helps destroy any remaining faith one might have had in AMD's intelligence.

The CPU is overrated, overpriced, and is definately not the fastest. It is still fast though and more than adequate for any normal task.
May 13, 2003 4:52:19 PM

It's not totally that the rating system seems to lie, it's more the fact they base the price on the PR rating, <i>not</i> on the actual performance. If it was considerably cheaper, then people wouldn't bitch so much. it's the way the price/perf ratio seems to be getting worse as AMDs new cpu releases continue - that's what annoys people. it looks like AMD are trying to over-hype their products so they can screw more money out of customers.

---
:smile: :tongue: :smile:
May 13, 2003 6:19:26 PM

Overall, the only AMD problem is it's PRating...

AMD should have this CPU 3050+ or 3100+ or even 3000+/FSB400.

This would have pleased eveyone! PowerUser would have loved the idea of a 200 MHz FSB, and comparison with Intel would have been more "fair".

Yes, I like my idea... I would have the choice to buy a 3000+/FSB400 or a 3000+/FSB333 (for my old A7N8X Rev. 1.04).

--
Would you buy a GPS enabled soap bar?<P ID="edit"><FONT SIZE=-1><EM>Edited by TheRod on 05/13/03 02:20 PM.</EM></FONT></P>
May 13, 2003 8:26:24 PM

Maybe it's me, but I just don't get what you have against the conclusion. You didn't outline at all where THG fell short.

I totally agree with THG on this one and am glad they blasted AMD. Hope they even make a very public rant about it, and continue bugging AMD.

--
This post is brought to you by Eden, on a Via Eden, in the garden of Eden. :smile:
May 14, 2003 1:34:15 AM

It seems to get results. Look at the AMD Thermal Death video for an example. To tired tonight to look up the link. It's here at THG though.

If it ain't broke, take it apart & see why not!
May 14, 2003 2:12:49 AM

I also agree with the conclusion.

But if THG tells them to correct the PR system, how AMD will fix it(if they want to fix it)?

The PR system was very nice till T-bred "B". Barton with 400 MHz FSB is an very impressive CPU, but AMD destroyed it with dreadful PR system

----------------
<b><A HREF="http://geocities.com/spitfire_x86" target="_new"> My Website</A></b>

<b><A HREF="http://geocities.com/spitfire_x86/myrig.html" target="_new"> My Rig</A></b>
May 14, 2003 2:42:07 AM

Ok, here goes. AMD already fixed it, it's not the result of the POMPUS and BIG HEADED THG group either. Note the numbering system on the Opterons? I bet THG will say that was solely result of their rant, nevermind it was though of long before the article...

Now, here's why I think it's bad, but acceptable.

AMD is at least trying to get a relatively straightforward comparisons scheme. Not between AMD and Intel (as it was initially) but at this point, just between AMD CPUs.

Initially
P4 vs Model number initial - Model number underated the Athlon
Next P4 revission, they were comparable
Next, the P4 had a slight advantage
Now, P4 has a significant advantage


probably missed a revision in there, but I cound both core and bus changes as a revision.



Now, After the P4 surpassed the Athlon rating, AMD had a couple of options:

1) Chuck the rating system, go back to Mhz, and allow Intel to mislead people.

2) Revise the rating system, make the slower Athlons look faster and the faster athlons look slower, confusing customers by giving them the numbers related to the current core.

3) Come up with a new rating scheme for the athlons, not even related to the current scheme.

4) Stick with the current scheme.


Problems with each.
1- Face it, this really misleads the customers into buying a usually worse CPU performance wise, and spending a lot more money in many cases (both on the mobos and the processors, higher end would be just the mobos)

2- Afther the 3rd revision of the P4, the following athlon numbers would go down, say having (as an example, not necessarily correct), the 2000+ be rated as 1700+, even though it's faster than the previous 1800+... Yeah, that's brillian.

3- This wouldn't be bad, but it might confuse people, there was enough of a problem with the model numbers initially, what will happen when they don't compare at all to the P4? Not to mention, just switching from pseudomhz to not at all in one architechture could end up confusing people, mhz to pseudomhz doesn't seem as bad.

4- Well, it's misleading, unless you compare to Athlons only. Now, typically, this is how you should compare anyway, since architechture to architechture has a wide variety of differences, and while both P4 and Ath. are x86, they vary in strengths and weaknesses. Example: For the tasks I'm going to be doing, an Athlon 2500+ Barton would smash any non HT P4, and the extra price of the HTP4 and good chipset isn't worth it. Now for what my aunt does, the P4 (HT won't really affect it) is the optimal choice.



So, AMD took the last option for now, and decided to change it for the Opteron/x86-64.




The THG rant wasn't very good journalism in my oppinion, as it was more of an ANTI-AMD rant, than a constructive peice of work. What do they think AMD should do for the correct model numbering I wonder?

Athlon XP 1600+, MSI K7T PRO2 RU (POS), 2x256 MB CRUCIAL PC2100 CL2.5 memory, Asus V6800 DDR Delux (GF 256) video card, 6.4GB+27GB WD HD, 40GB IBM HD (all 7200RPM). My computer is an acronym
May 14, 2003 4:55:57 AM

it wasnt much of an anti amd rant though, they summed up the article by saying its a good chip with a screwy model rating as compared to Intel,

i like the idea of just leaving it as a 3000XP w/ 200mhz FSB though, that would have gone over much better
and why is it anti-AMD to find fault with their marketing scheme? i like AMD to but its deceptive advertising,

you and i arent as bothered by it because we research and understand the P4 vs XP debate, but general consumers look a the biggest number


"Going to war without France is like going hunting without an accordion."
- Gen. (Ret) Norman Schwartzkopf
May 14, 2003 5:15:37 AM

The simple fact is this: AMD has some rather large gonads charging what they charge for their high-end chips when you compare them to the performance of high-end P4's. You can buy a 3.06 HT Northwood for $367, and it beats the 3200+ in almost all disciplines, sometimes badly. How much is the 3200+ going to cost? $475?

And there's the problem. It's not about deceptive PR Ratings so much as it is about Price/Performance. It used to be that AMD was a better value at the high-end, but now, they are trying to make more money from an inferior product. It creates a situation where there's no rational reason to buy a high-end AMD processor.



<font color=green>The Netherlands is where you go when you're too good for heaven.</font color=green> :tongue:
May 14, 2003 10:25:53 AM

that point, I agree with at least, they shouldn't charge so much for it.

Athlon XP 1600+, MSI K7T PRO2 RU (POS), 2x256 MB CRUCIAL PC2100 CL2.5 memory, Asus V6800 DDR Delux (GF 256) video card, 6.4GB+27GB WD HD, 40GB IBM HD (all 7200RPM). My computer is an acronym
May 14, 2003 11:09:48 AM

After reading the review here on THG I was pretty dissapointed about this processor. But if you look the review at X-bit Labs (www.xbitlabs.com), the processor is not a dissapointment at all, especially when it comes to gaming & business apps. Of course, it doesn't beat P4 3.0GHz in digital content creation applications, streaming data processing tasks and 3D rendering but if AMD gave this CPU a 2800+ rating as a few in here suggested, then they would be unfair towards their product because in some cases it performs as a "3200+" or AT LEAST as a "3000+".

The way I see it is that you have to choose your processor according to your daily tasks. If you are mostly in gaming then I think this new CPU should deffinetely be your number one choice (at least from what I've seen at X-bit Labs review), otherwise go for the P4 which is anyway a more expensive platform overall (I think).
May 14, 2003 11:53:14 AM

You seem like a very pro-AMD person in order to view THG's conclusion this badly.

They are only giving the truth, and personally, the harder it is, the more the company gets blamed in OUR WORD as well as THG's, since we can't voice it in any other way. THG is doing the right thing blaming for once AMD, since some thought they were AMD-biased, and the fact they stuck also the regular testing and not the AMD recommende optimized modes, shows even more honesty.

You're wrong, IMO, you're only pro-AMD if this conclusion bothers you. THG delivers the HARD TRUTH, period.

--
This post is brought to you by Eden, on a Via Eden, in the garden of Eden. :smile:
May 14, 2003 3:11:49 PM

"your wrong"

Nice, obviously you know nothing about civilized and adault discussions. I can see why this article's conclusion pleases you.

Now, the whole thing was a rant. The model numbers were there to make a straight comparison between AMD processors, and second generation P4 processors. Now, if THG made some decent suggestion on what AMD should do (and not some silly '2800+' suggestion as they put it), it might not be so bad.

Maybe if they toned it down instead of making it sound like such a rant, it wouldn't be so bad.

But this article's conclusion has not place in a serious news site.

Oh, wait, this is THG, I'm sorry. They do great data gathering, but the news is sometimes questionable.


Now, as for me being pro-AMD, yes, for the past few years, AMD has provided me with the best alternative, and yes, I am pro-AMD. However, that does not affect my judgment on the article, as I would be similarly peeved seeing something against Intel on the same note as what was up there (last time I did, I didn't use these boards though).

As for the benchmarks, only one seems to show any relevance to any user I've seen. For the tasks I do, an Athlon of lower model number can pretty much beat any P4 without hyperthreading (and a Barton will easily compete with an HT P4), now maybe the problem is THG is too limited on their benchmarks and puts too much emphasis on game benchmarks, and should do some scientific stuff instead. I think the AMD-Oriented benchmarks are bad if places are doing them, but those benchmarks seem to relate more towards what I'm doing.


Realize this: THG benchmarks are no where near comprehensive, and therefore they should be used in conjunction with other sources.

Athlon XP 1600+, MSI K7T PRO2 RU (POS), 2x256 MB CRUCIAL PC2100 CL2.5 memory, Asus V6800 DDR Delux (GF 256) video card, 6.4GB+27GB WD HD, 40GB IBM HD (all 7200RPM). My computer is an acronym
May 14, 2003 3:55:17 PM

Quote:
"your wrong"

Nice, obviously you know nothing about civilized and adault discussions. I can see why this article's conclusion pleases you.

While not the best grammar in the world (it should be you're as in a contraction of you and are, not your as in signifying possession of) Eden has a pretty simple and valid observation. You <i>were</i> wrong. Like it or not, it's just that simple. If you take a simple observation and correction so personally then clearly we can see just how much you actually know of civilized adult discussions and we can see why the simple but painful truth delivered in THG's conclusion annoys you. Do you prefer fantasy worlds without confrontation where we just ignore problems and hope that they go away instead of pointing them out and dealing with them like adults?

Quote:
The model numbers were there to make a straight comparison between AMD processors, and second generation P4 processors.

The model numbers were theoretically there to make a straight comparison between AMD processors, period. Call me crazy, but so far all that I can see is that AMD has done a piss-poor job of sticking to their own theory of making processor comparisons with their own CPUs <i>easier</i> by following their rating system.

Quote:
Now, if THG made some decent suggestion on what AMD should do (and not some silly '2800+' suggestion as they put it), it might not be so bad.

Is there anything else that AMD could have done to rectify the errors in their rating system? I would love to hear what you consider to be a valid suggestion for just it should be corrected.

Quote:
Now, as for me being pro-AMD, yes, for the past few years, AMD has provided me with the best alternative, and yes, I am pro-AMD. However, that does not affect my judgment on the article, as I would be similarly peeved seeing something against Intel on the same note as what was up there (last time I did, I didn't use these boards though).

That's an awfully easy statement to make when you have absolutely no means of backing it up. Forgive me if I am in no way impressed by it nor give it any value.

Quote:
As for the benchmarks, only one seems to show any relevance to any user I've seen.

If only one benchmark shows any relevance to any user that you've seen then clearly you haven't seen many users at all.

Quote:
Realize this: THG benchmarks are no where near comprehensive, and therefore they should be used in conjunction with other sources.

No one, including THG themselves, have ever stated that THG benchmarks are all-inclusive. They aren't. No site comes even close. In fact, no site could. It is impossible.

THG's benchmarks however are aimed at the extreme majority of consumers who will be reading their reviews. Since there is an incredible weight towards internet use, office software use, compression/decompression use, and gaming in PC users THG aims their benchmarks at those users. THG makes their articles worth-while to the largest market possible without a wasteful expendature of time, energy, and money.

Just because you are in the incredibly small minority which heavily uses scientific apps does not mean that any one )and especially all) review sites should have to cater to <i>your</i> monumentally small minority. (And just so that you know, I'm an analytical x-ray diffraction software engineer, so I know <i>just</i> how small the scientific application miniority is. This is why I don't rely upon benchmark sites to know how well a CPU runs my software.)

<font color=purple><pre><b>There are 10 types of people in this world: those who can understand binary and those who can't.</b></pre><p></font color=purple>
May 14, 2003 6:25:34 PM

what i wanna know about the last cpu review was why they didnt list a 2600 xp with 166fsb?
May 14, 2003 6:40:53 PM

My view on the cpu's. So the PR rating didnt really match its performance compared to a p4. SOO What. I look at like this. You have a cpu (AMD) running at 2200 MHZ keeping up with a cpu(Intel) that is not only running 800 MHZ fsster but also on a faster fsb.
MHZ to MHZ AMD is kicking Intels a*&. Yes Intel has the "fastest" cpu, but is it the most powerfull? What happens when Intel cant increase thier speed anymore because of Ohms law? Amd will still have a lot of room to increase while Intel will have to rely on some new technology.
AMD has the horsepower to get the work done at lower clock speeds while Intel must run the higher Mhz just to keep up.
I like amd but intel will do just fine. If intel had better cpus I would use theirs.
I use amd right now and have no plans to switch until Intel has a better, more powerfull cpu than AMD.

All done. REady to be slammed.
May 14, 2003 7:21:40 PM

We're not talking about "when Intel can't increase Mhz anymore..." We're talking about <i>now.</i> I must add that Intel has consistently been able to offset this seeming vulnerability with its superior execution of smaller, more efficient core processes. And Intel's core design is innately more conducive to higher clock speeds anyway.

Incidentally, the P4 <i>still</i> has more overclocking headroom than Barton. Look, the Mhz-Mhz argument just doesn't hold any weight any more. The only thing that matters is the real-world, bottom-line performance of the part, and the P4 wins. It might be different if there was some material advantage to be had from AMD's lower clock-speeds, but as far as I can see, there's no advantage; no power savings, no thermal advantages, no increase in stability, no reduction in production costs, and certainly no benefit in terms of performance. So where is the advantage? Enlighten us.

Seriously, if AMD could ramp the Barton to 2.6-2.8 GHZ, the IPC argument would hold water, but since AMD can barely get the Barton to 2.2 Ghz, the argument is specious. You are using an argument that has no impact whatsoever on the proposition.

<font color=green>The Netherlands is where you go when you're too good for heaven.</font color=green> :tongue:


<P ID="edit"><FONT SIZE=-1><EM>Edited by Twitch on 05/14/03 03:23 PM.</EM></FONT></P>
May 15, 2003 10:27:17 AM

If I'm wrong, counter my argument, don't just say "you're wrong", that doesn't provide a constructive counter, as it doesn't say why.


How do the model numbers make comparisions more difficult? In general higher model numbers tend to work better, though the plain model number does give all the information, does Intel's plain Mhz? No, you can get a higher Mhz processor that in many cases will be slower (cite: 2.4Ghz HTP4 Vs. 2.5GHz non-HT P4).


If you want to see some valid suggestions, read my previous posts.


YOu are right, I've no means of backing up my statement of objectivity, but then again, given that statement, you've no means of countering it. Since Intel hasn't been in this position since I really started using these boards, you can't much say that I haven't shown signs against it either?


I've seen a lot of users, theoretical benchmarks tend to be a start, but often lack in real world accuracy, and the rest here are games, Q3 is the only one I know of that anyone I know plays. I know a lot of users, most do scientific stuff. Lots of AMD fans, many people who just want something, don't do much reasearch, and end up with Intel, and a few Intel fands. Except the ones that have the cash for Itanium IIs (gotta admint, those things are amazing)...


Maybe they shouldn't have to cater to scientific apps because they are a small minority, however they do provide a good alternative benchmark, and maybe one of the dozens of games could be dropped to put on in. Since you say you work with a scientific group, you should know quite well, then often times a broad sample of many variable types (with only a few in each type), may give a better general overview of the results than a overview of all but a few variable types, and, though slightly more in each type, not nearly enough to be comprehencive in each.

Athlon XP 1600+, MSI K7T PRO2 RU (POS), 2x256 MB CRUCIAL PC2100 CL2.5 memory, Asus V6800 DDR Delux (GF 256) video card, 6.4GB+27GB WD HD, 40GB IBM HD (all 7200RPM). My computer is an acronym
Anonymous
a b à CPUs
May 15, 2003 1:35:05 PM

Whats wrong with the authors conclusion?

It says:

The intel has a new FSB.

The AMD is having itself on with its claim of a 3200Mhz equivalent CPU Mhz speed when it should be 2800Mhz.

My conclusion:

Intel has misled with the use of a 3Ghz CPU clock, because they already had a 3.06Mhz CPU. Just like in the theme of overclocking they split and send the multipliers wandering and set the FSB higher, huh and what is their claim now, that the 3Ghz is faster than the 3.06Ghz?
May 15, 2003 2:27:45 PM

Actualy when it comes to THG reading the conclusions are the best thing since his results are so one sided towards INTEL and optimized for them they are useless at least the conclusions can give you something to laugh at.
May 15, 2003 3:03:23 PM

Oh boy... I´ve heard that a thousand times before...

OK, folks, so hypothetically speaking, if you had a 2Ghz processor that could do 3 operations every clock and a 3Ghz one that could do 2.5 operations per clock, you´d buy the 2Ghz one? Boy, that´s just stupid! It´s "the most powerful, clock by clock", obviously, but what about "the one with higher clock"? The only thing that matters is NET PERFORMANCE!!! You said "fastest" and then used "powerful". And then you said that
Quote:
I use amd right now and have no plans to switch until Intel has a better, more powerfull cpu than AMD.

Well, the 800Mhz FSB CPUs clocked at 3Ghz and 3.2Ghz BOTH deliver MORE PERFORMANCE than the 3000+ and 3200+. AND guess what? Mhz and Ghz DOESN´T MATTER!!! (AMD´s own motto) If Intel´s top of the line delivers more performance, then those processors are more powerful - at least in my dictionary - than AMD´s top of the line.
May 15, 2003 3:13:12 PM

Ahhh yes but there is a fundamental difference between Intel´s Mhz and AMD´s PR rating. Intel´s Ghz can´t be argued with because they are physical characteristics of the CPUs they´re selling - they can´t possibly lie, and they don´t. Of course they don´t tell the whole story all by themselves, but then again, that´s impossible.

The problem with the PR rating is that it is indeed disconnected from something solely physical the processor in the usual way, and is meant to include a small idea of the IPC that Athlon processors are capable of. The problem is that IPC is not a straightforward matter as clock speeds, and that makes AMD´s life harder. It´s hard to keep that PR rating going, and they should be allowed some headroom.

<i>However</i>, the latest processors from AMD are rated in such a manner that most people around here (from what I´ve read) think they went beyond that headroom that we should accept. The 3200+ doesn´t outperform the 3.0Ghz by some 5~10%, as you´d expect from its rating, but rather falls behind considerably.

So that´s it. 'nuff said, already wrote too much.
May 15, 2003 3:43:12 PM

Ok. Ill agree Intel has the fastes, most powerfull cpus out right now. But look what Intel has to run to beat AMD. 800 MHz faster cpu speed and faster fsb. Yes the PR rating is miss leading, especially the 3200+, and needs to be changed. But it pretty much keeps up with the faster Intel.
What I want to see discussed is the fact that the AMD cpus are pretty much keeping up with the Intel chips at a lot lower clock speeds and lower fsb. This tells me that Amd is planning for the future.
Thank God we have a choice of Intel or AMD or else we'd be paying $1000 for an Intel 2Ghz. Competition keeps prices down makes companys improve their products.
May 15, 2003 5:41:48 PM

Quote:
Ok. Ill agree Intel has the fastes, most powerfull cpus out right now. But look what Intel has to run to beat AMD. 800 MHz faster cpu speed and faster fsb. Yes the PR rating is miss leading, especially the 3200+, and needs to be changed. But it pretty much keeps up with the faster Intel.

I do understand what you mean, but I think many people seem to be overly impressed with Athlon´s high IPC, which is of course just one side of the story and is as meaningless as Mhz if you don´t consider net performance. It´s just that AXPs and P4s have different architectures. P4s have high clock and low IPCs, and AXPs have high IPC and low clock. It´s a design choice. And because I´m not quite aware of every engineering intricacy, I usually worry just about the performance I get with a purchase, and that´s all. I don´t know much about the headroom that these two companies have and I frankly don´t care, as long as I get a good product for a good price. Is AMD really looking more to the future than Intel with their design choices? I can´t see the answer for that one clearly. I don´t think it has an obvious answer...

Anyway, I completely agree with your last sentence:
Quote:
Thank God we have a choice of Intel or AMD or else we'd be paying $1000 for an Intel 2Ghz. Competition keeps prices down makes companys improve their products.

Good products for good prices! :cool:
May 15, 2003 8:22:23 PM

I or imgod2u could right now put you in a learning session to show you why the P4 needs so much, and why it is not even important.

I don't feel like typing a book now explaining on the P4's fetching FSB algorithms, pre-fetching demands, cache importance, whatever, or the whole pipeline deal and the P4's IPC.
All I can tell you is that you are wrong to say that the P4 isn't impressive if it needs all this to fight back.

--
This post is brought to you by Eden, on a Via Eden, in the garden of Eden. :smile:
!