Is it fair to benchmark P4+RDRAM vs AthlonXP+DDR?

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

fkloster

Diamond Member
Dec 16, 1999
4,171
0
0
Originally posted by: Macro2
Does Intel even support 1066 RDRAM?

Mac

Who cares....Chevy did not intend customers to put 'blowers' on their 67' camaros but when customers do put blowers on ....they go faster....should we go and remove the blowers because chevy didn't put blower modifications into their camaro white papers? (whew...off the deep end :) )

 

SexyK

Golden Member
Jul 30, 2001
1,343
4
76
I am so sick and tired of people claiming the P4's arcitechture gives it "unfair advantages" over its competitors. The P4 was designed to scale up in clock speed, you all can think it was designed that was purely for marketing purposes, but Intel isn't stupid enough to do that. Intel designed the chip to work on a faster bus arcitechture, how does that give it an unfair advantage? Its just a fact. It was designed to funtion at its peak with high bandwidth memory, so how is it unfair to AMD to show it using RDRAM? I just dont understand youre arguement, espicially things like "well that benchmark uses SSE2 and AMD doesnt support that!" Well, news flash, there are plenty of programs that USE SSE2 that professionals and hobbiests use every day. That means, in those apps, the P4 will be way faster, it's not unfair, its not misleading, its just a faster chip at some tasks. Those are the facts. Grow up and live with it.

And as for all those innocent people that have been "tricked into" buying P4 and SDRAM, well, I don't think those people are going onto Anandtech or Aceshardware before their big purchase either way, so I dont see how showing PC1066 RDRAM is unfair in those cases. These sites cater to the enthusiasts that know what they are doing, and for a LOT of people who read such sites, PC1066 is a reality already paired with their P4 chips.

Kramer
 

dexvx

Diamond Member
Feb 2, 2000
3,899
0
0
You know, Anandtech, HardOCP, tomshardware, Aceshardware, etc are called "enthusiast" sites. Theres a reason why they go more in depth than the average joe. Moreover, you are NOT running the CPU out of spec. Intel makes processors. If you run their processors in spec, than they are in spec regardless of memory speed. So IMO, DDR333, DDR400, PC1066 are perfectly legitimate so long as you run the CPU in spec. Intel has NO control of what the mobo manufacturers do just like AMD has no control over what VIA / SiS does. So long as you run the CPU in spec, you are giving a fair CPU comparison. Whatever motherboard features individual motherboards have have nothing to do with the CPU, even if it is an intel based board.
 

ToBeMe

Diamond Member
Jun 21, 2000
5,711
0
0
Originally posted by: ElFenix
Originally posted by: Macro2
Does Intel even support 1066 RDRAM?

Mac

no
That's right Intel does not *officially* aknowledge PC1066 until Oct. 7th according to their new roadmaps and apparently Asus/Abit/and others have boards ready to go.........



  • On October 7th, Intel will introduce its 845PE, 845GE and 845GV chipsets, which support DDR 333 memory types. On the same day it will position its 850E with PC 1066 Rambus official validations as the highest performing desktop chipset.

As for Maj's claims to a "paper launch" of the 2.8 as well as others..........odd, especially since the 2.8's have been available at Ingram and other places for a while now and several places online have had them in stock. Still can't find that 2600+ anywhere though...........;)
 

Macro2

Diamond Member
May 20, 2000
4,874
0
0
RE:"I am so sick and tired of people claiming the P4's arcitechture gives it "unfair advantages" over its competitors"

No one said that. P4s architecture doesn't give it an unfair adavantage. It's just when you model benchmarks to showcase those couple of things it does well...that things get pretty fuzzy.P4s encode MP3's and play games all day the P4 does pretty well. For the rest of the apps we all use it doesn't do so well.
Let's face the facts...Intel dictates a lot of benchmarks. Bapco shares the same address as Intel for crying out loud.

Like I said before. I can understand the average Joe falling for "MHz sells" but no people on this BB. In light of that, if the P4 is for you then go for it.




Not sure that Intel planned it that way but the scaling of the P4 as a marketing tool was a stroke of genius. OTOH, now they have some lawsuits about misleading benchmarks so who knows.
 

Ilmater

Diamond Member
Jun 13, 2002
7,516
1
0
First off, I am pro-AMD because I am pro-capitalism (please don't flame me, but it's the best system in an imperfect world). If Intel were the only one in the business, we'd all be paying whatever Intel wanted us to pay. So, I will always buy an AMD unless they get a much better market share (say 40% or so). It's not just because I like the underdog, it's because we NEED the underdog.

That being said, I thinkt that the benchmarks are fine using RDRAM PC1066 compared to a DDR333 Athlon XP. It's all about what the absolute fastest combination is. I don't even fault the manufaturers of benchmarks (synthetic or otherwise) for using them. Think about it; if you're Blizzard or Id or any other software designer, you want to make the best game possible that will run on as many systems as possible. If I can make my game run faster on 80% of all of the systems out there by utilizing SSE2 instructions, why in the world wouldn't I? That would be ignorant of me. Not only that, but if it were a popular game, then I believe that it should most definitely be included in most if not all benchmarks done by the various web sites. It's completely valid because the REAL WORLD PERFORMANCE of that game and many other games and other applications truly would be faster to the end-user. It would not be fair, however, to say that those shouldn't be used just because they show the P4 having a performance advantage. That would be like saying if Excel had a benchmark, and it ran better on an AMD than a P4 because of better branch prediction, it couldn't be included. That's just crazy talk. If I ran only Excel and I wanted a machine to do it faster, then if the P4 does it faster then that would be the way I would look.

I do have one thing that I do agree with, however. I do think that it should be noted how much of a price differnece there is between the different parts used in a system. If I can get 10 more fps in Q3 on a P4 system that costs $500 - $600 more than a given Athlon system (that's $50 - $60 per extra frame), then you had better be an idiot to be buying the Intel system IMO. Hell, you could buy the latest graphics card (which will last longer anyway) instead, which would get you just as much of a performance boost. Sure, P4s will O/C better and their chipsets are more stable (PLEASE MAKE YOUR OWN CHIPSETS AMD!!!), but are they $500 - $600 more overclockable and more stable? Dream on. Sorry Intel lovers, but for the majority of users, that just isn't the case. And yet, tomorrow I will wake up, and Intel will have around an 80% market share and AMD will still be playing catch-up. If these prices were listed, I think that it would really put the whole thing in perspective.
 

dexvx

Diamond Member
Feb 2, 2000
3,899
0
0
Retail Intel Pentium4 2.53Ghz - $250 shipped Available Now
Retail Intel Pentium4 2.8Ghz - $625 shipped Available Now
Retail AMD Athlon 2600+ - $2?? shipped Available in Mid-Late Sept

I dont see where you're getting a $500-600 differential. The 2.8 is way expensive simply because its the newest tech available and no consumer proc can match it.
 

Rand

Lifer
Oct 11, 1999
11,071
1
81
Whether it's fair depends upon what the purpose of your comparison is IMHO.

Are you comparing the fastest availavle plaotforms officially supported by the processor/chipset manufacturer?
If so then you'll be using a P4 2.8GHz + I850/PC800 DRDRAM vs. AXP 2600+ KT333/PC2700 DDR.

Are you comparing the fastest platform possible without running anything out of spec by manufacturers recommendations?
If so, it's P4 2.8GHz + I850/PC1033 DRDRAM.... though you'll have to specifically ensure that the motherboard manufacturer does guarantee PC1033.
On the AMD side it would again be be AXP 2600+ KT333/PC2700 DDR.

Are you comparing Price/Performance?
If so then you'll likely be comparing a P4 + I845E/PC2100 vs. AXP KT266A/PC2100. In addition, you'll probably end up running a P4 clocked lower then the comparable AXP's Model rating.


What's "fair' depends entirely upon what the purpose of your comparison is, and whom your target audience is.
 

Sevenhunt

Member
Aug 23, 2002
104
0
0
Sure its fair.
rdram800 on a p4 loses to PC2700 on a P4 845 g chip board.

So DDR vs DDR would`n`t be fair.
 

sandorski

No Lifer
Oct 10, 1999
70,655
6,222
126
Originally posted by: Rand
Whether it's fair depends upon what the purpose of your comparison is IMHO.

Are you comparing the fastest availavle plaotforms officially supported by the processor/chipset manufacturer?
If so then you'll be using a P4 2.8GHz + I850/PC800 DRDRAM vs. AXP 2600+ KT333/PC2700 DDR.

Are you comparing the fastest platform possible without running anything out of spec by manufacturers recommendations?
If so, it's P4 2.8GHz + I850/PC1033 DRDRAM.... though you'll have to specifically ensure that the motherboard manufacturer does guarantee PC1033.
On the AMD side it would again be be AXP 2600+ KT333/PC2700 DDR.

Are you comparing Price/Performance?
If so then you'll likely be comparing a P4 + I845E/PC2100 vs. AXP KT266A/PC2100. In addition, you'll probably end up running a P4 clocked lower then the comparable AXP's Model rating.


What's "fair' depends entirely upon what the purpose of your comparison is, and whom your target audience is.


There you go, the best answer so far.
 

jbond04

Senior member
Oct 18, 2000
505
0
71
Originally posted by: SexyK
I am so sick and tired of people claiming the P4's arcitechture gives it "unfair advantages" over its competitors. The P4 was designed to scale up in clock speed, you all can think it was designed that was purely for marketing purposes, but Intel isn't stupid enough to do that. Intel designed the chip to work on a faster bus arcitechture, how does that give it an unfair advantage? Its just a fact. It was designed to funtion at its peak with high bandwidth memory, so how is it unfair to AMD to show it using RDRAM? I just dont understand youre arguement, espicially things like "well that benchmark uses SSE2 and AMD doesnt support that!" Well, news flash, there are plenty of programs that USE SSE2 that professionals and hobbiests use every day. That means, in those apps, the P4 will be way faster, it's not unfair, its not misleading, its just a faster chip at some tasks. Those are the facts. Grow up and live with it.

This is the most intelligent post I have seen yet. SexyK is right on target with his argument.


This debate is regoddamndiculous if you ask me. The P4's architecture inherently performs well on a high bandwidth platform (like PC1066 RDRAM). If that causes the P4 to perform better than the Athlon XP, that's because Intel designed it that way.

Even though Intel has not validated the use of the i850E chipsets with PC1066 RDRAM, motherboard manufacturers like Asus have. Therefore, using PC1066 RDRAM is not "out of spec", because Asus, Gigabyte, and Iwill motherboards all officially support it. I don't care whether or not Intel has given the blessing for PC1066 RDRAM to be run on the i850E, because the motherboard manufacturers have done their own testing and validation. If an i850E motherboard with PC1066 RDRAM is unstable, then that is the fault of the motherboard manufacturers, not Intel. If one KT333 motherboard was unstable with PC2700 DDR RAM, is it fair to call all KT333 motherboards unstable, and then point a finger at Via? The fault would obviously lie with the motherboard manufacturer.

Also, why are the use of SSE2 enabled benchmarks unfair? Majewski9, are you trying to tell me that no one out there uses Lightwave, 3ds Max, or other SSE2 enabled programs? Do you think that Newtek (the makers of Lightwave) or Discreet (the makers of 3ds Max) put SSE2 code into their programs just for the sake of making it look better on the P4 architecture? Give me a break...they use SSE2 code to help consumers get the most out of their programs.

The point of all of this is that it is only fair to benchmark CPU's on the fastest possible platform if you are trying to declare a performance winner. If the P4 was benchmarked with the i845E chipset using PC2100 DDR and lost, are you going to tell me that the P4 is slower than the Athlon XP, when there are obviously other faster, officially validated platforms that would cause the P4 to outperform the Athlon? Use some sense here, people.
 

Diable

Senior member
Sep 28, 2001
753
0
0
Ace's used a i845G mortherboard in there last bake off and even with the reduced bandwidth of DDRAM the 2.4GHz and above P4's were faster(in most test)than any currently available XP.
 

Macro2

Diamond Member
May 20, 2000
4,874
0
0
Right.
I think it's fair you go out and buy Rambus 1066 and a new P4 2800 because that's the fastest gamer/bandwidth benchmarker. Go for it.
$600 for the chip. A couple hundred for some Rambus...what the heck, it's only money. Helps the Intel cause. Keeps great companies like Rambus in business too!

I dunno what's worse...the P4 bandwidth/Bapco scam or the Rambus scambus.
 

NOX

Diamond Member
Oct 11, 1999
4,077
0
0
Originally posted by: Macro2
Right.
I think it's fair you go out and buy Rambus 1066 and a new P4 2800 because that's the fastest gamer/bandwidth benchmarker. Go for it.
$600 for the chip. A couple hundred for some Rambus...what the heck, it's only money. Helps the Intel cause. Keeps great companies like Rambus in business too!

I dunno what's worse...the P4 bandwidth/Bapco scam or the Rambus scambus.
That remark really exhibits the actual intent of this thread.

It?s really sad that people can?t get over the fact that Intel has better means to take its chip to a new climax.

Like I told Hardware in another thread, don?t get mad at Intel, get mad at AMD for relying on sub-par third party chipsets/components for their chips!

Until AMD takes full control of the situation, Intel will continue to offer faster chips, and more bandwidth.
 

Ilmater

Diamond Member
Jun 13, 2002
7,516
1
0
Rand and jbond04, you're both right, but I already said the same things. I don't think that it's unfair at all to compare the Athlon XP2600+ with DDR333 to the P4 2.8 with PC1066 RDRAM. I also laid out all of the reasons why. I guess that when you read that I was pro-AMD you just decided not to read the rest of the post. Nothing you said was new, no offense.

In fact, I didn't even say that all benchmarkers have to post the differences in prices between the two systems. I only said that they should because that is where AMDs really shine. Intel processors were made the way they were for two reasons: high clocks and marketing. They will scale forever. They already know that with enough cooling, the P4 will scale to 4GHz. It saves them some money, agreeably, because they don't have to redesign anything to stay ahead of AMD forever, whereas AMD just had to increase the layers of process in order to get more clocks from their proc. o that is nice. But they also designed it for the dipsh!ts of the world that have no idea what any of it means. They see 3000MHz and immediately say, "That is faster than 2800." Proof you ask? Their pipeline is 10 stages long. Therefore, not only does it take longer for an initial command to get through, but it takes more of a hit if there is an error - say at stage 8 - because then it has to lose all of the caluculations of the first 8 stages and start all over again. But, they knew that they could scale a processor higher - MUCH HIGHER - with the extra stages. It will run slower clock-for-clock than the competition (even back then they knew AMD was their only real competition), but it will show higher clock speeds. C'mon, an Athlon 2600+ runs 677 MHz slower than its Intel counterpart at the same speed. One wanted real speed and one wanted higher clocks. It's that simple.

Again, reviewers don't have to put price into the equation, but let's look at the first review I found with PC1066, which is from the Pabster, here. The P4 setup would cost $927 for the 2.8 P4 setup and $630 for the 2.533 P4 setup, both including shipping. Mind you, that's using Samsung instead of the Kingston that they used, which would have been $46 more with no difference in performance. I am fair. On the other hand, the Athlon 2600+ (which is preordering for right now for $300) setup would cost $494, while the 2400+ setup (again it's on preorder) would be $394. That is using Samsung RAM as well, since I couldn't find ANY winbond in 512MB sticks. You were asking where the $500 - $600 differences were? Well there you have it. Just in case you want to see like-performing systems, the same systems with a P4 2.2Ghz and an AXP 2200+ would cost $587 and $339 respectively. Nearly a $250 difference! I don't care how much money you have, you're crazy to spend an extra $250 just to have the P4 name. There's not enough overclockability or stability to justify that in my mind. Plus, most users don't overclock anyway. But if I'm joe schmo and I want the fastest computer I can buy, I stumble upon the Pabster's website and say, "I guess that would be best for me," without knowing what the differences in price are.

Now, with that said, please, don't respond to this without thinking about what you're going to say. Don't take this personally as I honestly don't mean it that way, but this is a ridiculously weak arguement from Sevenhunt:
Sure its fair.
rdram800 on a p4 loses to PC2700 on a P4 845 g chip board.

So DDR vs DDR would`n`t be fair.
There's absolutely no logic to back that up. I'm not even entirely sure if that's a pro-Intel comment or a sarcastic pro-AMD comment. So come on, I know you're all dying to beat that arguement down, so let's see what you got.
 

jbond04

Senior member
Oct 18, 2000
505
0
71
Originally posted by: Ilmater
Rand and jbond04, you're both right, but I already said the same things. I don't think that it's unfair at all to compare the Athlon XP2600+ with DDR333 to the P4 2.8 with PC1066 RDRAM. I also laid out all of the reasons why. I guess that when you read that I was pro-AMD you just decided not to read the rest of the post. Nothing you said was new, no offense.

None taken. I read your post (it was very good as well :)), but I needed to vent a little anger. :disgust:

Originally posted by: Ilmater
I don't care how much money you have, you're crazy to spend an extra $250 just to have the P4 name.

I would have to disagree with your point about spending $250 extra on a P4 setup being excessive. The reason why I went with my 2.53GHz/i850E/PC1066 RDRAM setup (see my sig) is because I wanted the overclockability, stability, and the SSE2 instructions that Intel could offer. I was tired of dealing with Via chipsets; and the Athlon process didn't appear have a whole lot of headroom left in it when I ordered (early June). Even now, I still feel that this purchase was a good one. It really didn't matter to me how much money I had to spend to buy the computer that I wanted, so it's not exactly fair to tell everyone out there that the cost/performance ratio is all that matters, because that's not true. Everyone has their own beat that they march to, and mine was "best performance, period". Others on this forum are like that, too. It's not right to go around saying that those who buy a 2.8GHz P4 setup are wasting their money. After all, it is theirs to waste. :)
 

Ilmater

Diamond Member
Jun 13, 2002
7,516
1
0
None taken. I read your post (it was very good as well :)), but I needed to vent a little anger. :disgust:

Completely understandable, as that is what I was doing.

I also have to complement you on your well-written response. You make an excellent point: everyone is looking for something different. Actually, back in June, things were to the point where I would have even considered a P4 system if I was buying then. Luckily for me (IMO anyway) I'm not going to be doing any major upgrades until December. I think it's best to say that my stance is, if you're going to be a truly great reviewer, then go to the trouble to find some details about the products you're testing (ie. pricing, stability, etc.) and present them all to the reader so that they can make their own decision. I just pieced a brand new system together (3 days ago to be exact), and at that time, I found the Athlon XP 2000+ to be the best performer for the money. There was no way I was going to go the P4 route for the guy I was building it for, because he doesn't want to shell out that much cash. For me - and him - it isn't worth it to pay the extra money. I'm always here to fix any problems he has with the KT333 board I got him, and he saves in the neighborhood of $300.

As you said, this is not the case for everybody. If I were an overclocker, I'd consider getting the P4. If I was filthy rich, I'd almost definitely do it since it does have the current performance crown, and I wouldn't have to mess with any stability issues. But even at that, I want to be shown every aspect of the system. That includes some subjectivity, which makes the entire process fun (IMO) for the reviewer. You can't objectively say that one is more stable than the other - really - since there are some people who have no stability issues with VIA chipsets at all, even though a majority of the time that is not the case.

So, the only thing I really wanted to say by posting here in the first place is something people on any side of the debate can agree with: the best review is one that gives as much objective information as possible, while still leaving room for some subjectivity, as long as it is labeled as such. Tell me that one is faster, tell me that one is cheaper, tell me that one is more hassle-free, and then tell me what you personally think. Then I'm happy no matter what components you use.
 

SexyK

Golden Member
Jul 30, 2001
1,343
4
76
Intel processors were made the way they were for two reasons: high clocks and marketing. They will scale forever. They already know that with enough cooling, the P4 will scale to 4GHz. It saves them some money, agreeably, because they don't have to redesign anything to stay ahead of AMD forever, whereas AMD just had to increase the layers of process in order to get more clocks from their proc. o that is nice. But they also designed it for the dipsh!ts of the world that have no idea what any of it means. They see 3000MHz and immediately say, "That is faster than 2800." Proof you ask? Their pipeline is 10 stages long

Well, first of all, the P4 has a 20 stage pipeline and the Athlon has a 12 stage pipeline if i recall correctly. Anyway, if you you Intel has somehow saved money by designing the P4 so "cheaply" then i think you're mistaken. The P6 core lasted relatively unchanged from the P-Pro's introduction until just recently when the Tualitins were finally taken out of production. The P4 architechture is not even 2 years old yet, so i'd be worried if it looked like it wasn't going to scale quite a bit higher, because it should have a coupe yeaers left in its lifetime.

C'mon, an Athlon 2600+ runs 677 MHz slower than its Intel counterpart at the same speed. One wanted real speed and one wanted higher clocks. It's that simple.

This is the biggest load of bull-*$^ arguement i've ever read. How you can possibly say Intel designed for clockspeed and AMD designed for "real speed" baffels me considering the nearly identically priced 2.53GHz P4 and XP2600+ have nearly identical performance. If GHz don't matter, as you so zealously defend, then why should it matter that the P4 requires a higher clock speed to perform at the same level as the XP? All that matters is how the chip performs, not it's speed in GHz, so to say that AMD chips have more "real speed" than Intel chips is absurd, considering that Intel makes the chips with the most performance out there, and their middle price range chips compete very strongly with AMD's top end. I would say i agree with you that a lower IPC chip isn't desierable if they were both running at 1.8GHz, but the facts are you can go out and by a 2.8GHz P4 right now, and thats way more real speed than anything AMD is offering.

Kramer
 

Ilmater

Diamond Member
Jun 13, 2002
7,516
1
0
Well, first of all, the P4 has a 20 stage pipeline and the Athlon has a 12 stage pipeline if i recall correctly.
I will begin by appologizing for the misinformation. It seems we both need to check our information, though you made it much more clear that you weren't sure of the actual number of stages in each respective pipeline. I meant to say that I wasn't sure, but that I knew it was more. The actual numbers are 20 and 15 for the P4 and the AXP, respectively. What, exactly, do these extra stages do, you ask? Basically nothing. They just move data around. There are no computations involved. But, since I didn't back up what I said before, I'll go to the trouble this time. First of all, from Geek.com, it proves that the P4 has 2 extra stages for no other reason than to move data around on the chip. From the Pabster's site here, it reaffirms what I was saying. It reads as follows:
The reason for the longer pipeline is Intel's wish of Pentium 4 to deliver highest clock rates. The smaller or shorter each pipeline stage, the fewer transistors or 'gates' it needs and the faster it is able to run. However, there is also one big disadvantage to long pipelines. As soon as it turns out at the end of the pipeline that the software will branch to an address that was not predicted, the whole pipeline needs to be flushed and refilled. The longer the pipeline the more 'in-flight' instructions will be lost and the longer it takes until the pipeline is filled again.
As I said before, Intel knew that adding useless stages to the pipeline would make for less IPCs, but went for it anyway. Why? Because Intel isn't worried about making the most innovations, they're worried about how many chips they can sell, again, IMHO. They knew that even if the processors didn't get the most IPCs, they'd be able to push it as having the higher number to the morons of the world (ie. the 80% of people that do no real research before buying a computer). In fact, they're in litigation with a class action lawsuit filed against them that claims (correctly) that Intel marketed their P4 as being faster than the P3, while in actuality, clock for clock, they were designed specifically do be slower, but have more clocks. I know, some won't believe me, so here a link from the INQwell. Thank god for AMD's PR ratings as they're the only way that AMD can compete for the "dumba$$" market segment.

Speaking of which, AMD, on the other hand, has commited itself, not only with its XPs, but with the Hammers as well, to actually ramping up performance rather than blindly increasing clock speeds. I'm not saying that Intel hasn't worked towards more performance - clearly Intel is at the top of the performance rankings right now - but they don't put much real thought into their processors. How else do you explain the fact that a company with an R & D department the size of a small country has a processor that has a piece-o-sh!t branch predictor? Of course I'm exagerating, but even though Intel could clearly spend some time to opitmizing it, taking less of a hit in performance because of less "flushes and refills," they don't.

I know that I already sound like I'm terribly biased towards AMD, and maybe I am. But as I said before, you'll all be sorry if AMD goes down the tubes. Intel's already lazy researchers will just get laid off, while it lays back and watches the cash roll in. I know that sounds like a little too much fire and brimstone, but if you don't think that it's anything but the truth, you're sadly mistaken. What's even worse is, the only one with enough power and enough need for faster processors is Microsoft. Linux doesn't need faster processors, it just needs more software support. However, Microsoft products could potentially last forever, as long as they stop trying to slow down computers by adding "fluf" to systems, such as the Longhorn's new entirely 3D GUI. Then they market it as a "must have" and Intel will work just hard enough to supply it with power. It will work with Microsoft to optimize one for the other so that they can: 1) convince people that they need the new software and hardware, and 2) get a firmer grasp on the entire market by making people that use Microsoft "need" (in marketing speak) Intel processors to run it. A lot of conjecture, but just try and come up with a spot where my logical guesses falls off. I dare you (that should really bring the flamers out!).

But, in spite of all of that, I am the LAST person to say that you should buy AMD just for AMD's sake, because that's not the point at all. The fact is, for the user that just wants a cheap computer - a HUGE part of the market - AMD is exactly what they want. I will mention one caveat to that: that is on an AMD mobo and NOT a VIA mobo. These home users don't need the extra performance a VIA gives them at the cost of the reliability of an AMD chipset. But you mention one true point:
How you can possibly say Intel designed for clockspeed and AMD designed for "real speed" baffels me considering the nearly identically priced 2.53GHz P4 and XP2600+ have nearly identical performance.
You're right, and actually, the 2600+, 2400+ (both preorder), and 2.53GHz P4 are priced at $300, $200, and $240, respectively, according to Pricewatch.com. So the top of the line AMD is slightly over-priced (though it wouldn't be too bad if you add in the cost of RDRAM or the slight performance hit when using the same DDR as the AXP would use. Now look down the line. The 2400+ is only $4 more than the 2.4GHz P4 (before shipping), while the real benefits start at the 2200+ ($141) vs. the 2.2GHz ($191), and only get better from there. To me and many other well-informed and price-conscious buyers, AMD is simply the best choice. The problem is, AMD doesn't have the marketing money that Intel does, so many aren't convinced of its abilities. That's why I post like I do in an effort to persuade some who might not know as much that stumble upon a review realize that, while currently P4s hold the performance crown, the premium people have to pay for that isn't worth it in many people's eyes. So I want Intel and AMD to continue to struggle against each other, and I want reviewers to be as informative as possible. There, I've said it agian.

You bring up good points, and truly the P4 is the only REAL processor (not some white paper p.o.s.) with speeds above 2.2GHz.
 

MadRat

Lifer
Oct 14, 1999
11,965
278
126
I want to see reviews with the P4 using SDR, DDR, and RDRAM. The only site that does massive benchmarking suites and uses a good variety of platforms is (unfortunately) Tomshardware.
 

Askheart

Junior Member
Sep 4, 2002
15
0
0
Say hi to everyone. I'm new here.

I was lurking around, then I read this comment which interested me.

"First off I dont really like the set of benches that Ace's used. Mostly gaming benhcmarks where we know favor the wide bus of the P4. So I dismiss those as being trivial since they really dont test what I use my machine for. Anyway the difference in many game benches are only a couple frames. Who can honestly tell the difference between 103 frames and 107? No one! Then they have some 3d studio max apps which the AthlonXP wins hand over fist, but then we get to the SSE2 benches which are absurd seing how AthlonXP doesnt use SSE2 as of yet. "

Aces 2600+ review

Aces 2.8 GHz review

The Warrior Kings gaming benche is interesting since it looks to be cpu-intensive (low frame rates) and includes SSE2 optimizations, anyhow the Athlon 2.13 GHz did outperform the P4 2.53 GHz by 40%. Another interesting thing is that a P4 rig was chosen by Matthew as the "Anandtech gaming rig of the month" and it hardly runs such a game.

I also have a complaint towards the "Intel recommended" Lightwave scenes used by the majority of reviews (there are situations in Lightwave where the Athlon XP wipes the floor with the P4).

Furthermore, there is no problem IMO in benchmarking a PC-1066 system vs. DDR333, or in using SSE2-enabled benchmarks (save the Sysmark fiasco) but I think that hardware reviews might put more emphasis on giving old x87 FPU hard time, definitely lives out there random software which counts on x87 FPU instead of 3dNow! Professional or SSE2 optimizations.

If someone wants the Athlon XP to do win the benchmarks, ask the software supporters to optimize for its architecture. I've found that the Athlon architecture is superior to P4, but you'll never see its real performance, if very few optimize specifically for the AMD processors.
 

BFG10K

Lifer
Aug 14, 2000
22,709
3,000
126
Well lets face it RDRAM especially PC1066 has more memory bandwidth than DDR.
And? That's the Athlon's fault for not supporting it and/or not taking advantage of it.

All the benches I see for p4's anymore use PC1066 which is not supported offically by the chipset.
True but it is officially supported by multiple motherboard vendors.

Most P4 systems sold use DDR and not RDRAM.
So what? I don't buy what most systems have. I buy what I want. If PC1066 paried with a 2.8 GHz P4 is the fastest platform available then I most certainly would like to see the benchmarks for it. I'm not going to pair it with SDRAM (for example) because Grandma down the road has a rig like that.

Also it seems to me like this is another classic case of "it's not fair because my favourite platform is losing". It's funny that during the Willammette days when AMD was winning nobody thought using RDRAM was unfair in those benchmarks. So why the sudden interest now?
 

Askheart

Junior Member
Sep 4, 2002
15
0
0
"Also it seems to me like this is another classic case of "it's not fair because my favourite platform is losing". It's funny that during the Willammette days when AMD was winning nobody thought using RDRAM was unfair in those benchmarks. So why the sudden interest now? "

Good catch :)
 

Tbirdkid

Diamond Member
Apr 16, 2002
3,758
4
81
I am sure this topic has been talked about but for my 2 cent and all its worth i dont think it is fair. That is why i compared my p4 1.6a with ddr to my xp 1600 with ddr and my xp blew it out of the water... P4s love memory bandwidth so obviously rdr will win the battle because of its quad speeds. DDR cant compare to it yet... however it will very soon and that is why everyone is going to ddr chipsets.