Why compare 1.9GHZ v.s. 3.0GHZ ?

[sean2002]

Why the hell would anyone post a comparsion between a 1.9GHZ Athlon and a Pentium 4 3.0GHZ knowing the P4 would win. All that does is tell some of the not so computer literate that the P4 is faster. Lets see a side by side of a P4 1.6GHZ v.s. an Athlon Xp-2000 (1600MHZ). Now lets post the results on the web so the people who do know as much about computers as some of us can see how the 2 compare clock for clock.

 

[Bovinicus]

It did bring up something interesting, wether the idea was dumb or not. It shows that if you plan on using a high speed Northwood CPU, you better get RDRAM, DDR333, or something better. As the clockspeed on the P4 increases, so does it's hunger for memory bandwidth.

 

[mchammer187]

1. clock for clock they are not the same nor are they meant to be
2. maximum overclock with XP vs max O'c with P4
3. AMD's fastest vs intel's fastest
thats all it is

i believe it is a fair test

 

[Sunner]

Sean, comparing a 1.6 GHz K7 against a 1.6 GHz P4 isn't fair.
The is designed to scale extremely high, and Intel willingly sacrificed IPC for clockspeed when they designed it.

What counts is absolute performance.

 

[dullard]

There are three major things people want to know:

1) How does a top-of-the-line Athlon compare to a top-of-the-line P4 when at their default speeds,
2) How do and Athlon and P4 at the same speed compare,
3) How will future top-of-the-line Athlons and top-of-the-line P4s compare?

#1 has been done a thousand times. A 2000+XP is identical in speed to a 2.2 GHz P4 on average.
#2 has been done a thousand times. A 1.6 GHz XP is faster than a 1.6 GHz P4 on most programs. We all know the P4 does fewer operations per cycle than the Athlon on most programs. Note: some programs they do the same number of operations per cycle and thus are nearly identical in speed when at the same frequency, but these are rarely shown in benchmarks.
#3 has never been done at a well known site.

That is the reason that Tom's compared the fastest overclocked Athlon to the fastest overclocked P4. It shows what we should expect if we want to overclock now, or what we should expect in 6 months for non-overclocked chips.



<< Why the hell would anyone post a comparsion between a 1.9GHZ Athlon and a Pentium 4 3.0GHZ knowing the P4 would win. >>


Like I said, since it hasn't been done, no one knew which would win! We didn't know how high the Athlon and P4 would go, we didn't know how they would perform at those speeds, we obviously didn't know how these unknown performances would compare against each other. The most interesting thing is that the 2.6 GHz P4 beat the 3.0 GHz P4 in many benchmarks! That shows that it would have been impossible to know exactly how the Athlon and P4 would compare.

 

[aldamon]

This article was very supportive of AMD. With Intel, WYSIWYG. Northwood IS the Pentium 4 for the forseeable future, all the way to 3 GHz. The overclocked AthlonXP, on the other hand, is a lame-duck processor that has yet to get its facelift. Like Tom said, 100 MHz more and the Athlon takes the prize.

The Thoroughbred core should clearly whipe the floor with Pentium 4.

Anyway, to answer the question at hand, a 3 GHz processor was compared to a 1.9 GHz processor because, like AMD is preaching, GHz doesn't matter. It's all about the performance baby!

 

[fkloster]

Great posts guys...keep it up

 

[Gonad the Barbarian]

I think a better comparison would be comparing chips based on current purchase price. Show how much CPU power say $250 will buy you from AMD and Intel respectively.

 

[sean2002]

I can kind of understand where you guys are comming from, I think the price to price comparsion would be cool too.

 

[Novgrod]

It seems the two comparisons that matter are performance/price and straight-out top performing chip from column A against top performing chip from column B. Mhz per mhz doesn't matter so much.

As the review mentioned, this comparison isn't necessarily how things will look in six months because well before then, the xp will be .13 micron and the rate at which it will scale up isn't well-known. Also, in six months the northwood will come at a default higher fsb, and that could have uncertain effects to its overclockability, I believe. In eight or ten months we'll have the clawhammer (or some time around then) which will be an entirely new entity

 

[x86]

<< Why the hell would anyone post a comparsion between a 1.9GHZ Athlon and a Pentium 4 3.0GHZ knowing the P4 would win. All that does is tell some of the not so computer literate that the P4 is faster. >>



I think the overall point of the comparison was missed. It was to see how high the current architecture can reach before becomming unstable. It also addressed the fact that when ascending to higher clockspeeds, DDR cannot deliver the bandwidth that the Pentium 4 requires, and that AMD has to release their 130nm technology in order to stay competitive.



<< Lets see a side by side of a P4 1.6GHZ v.s. an Athlon Xp-2000 (1600MHZ). Now lets post the results on the web so the people who do know as much about computers as some of us can see how the 2 compare clock for clock. >>



They are two totally different architectures, THAT would not be fair to compare.

-DocSmarts

 

[formulav8]

They are both x86 cpus. Why was it fair to compare a athlon to p3 mhz for mhz? But not a p4 and athlon? Because intel had to make the pipeline so long and the ipc low to rip off the unknowing consumer who thinks mhz = performance? I don't know but I guess the faster of each released would be what would make most people happy(Especially the p4 owners). So, I guess that is what I would go by. Or price/performance.



Jason

 

[xtreme2k]

The platform is fair. No matter what Mhz, it is INTEL's Best vs AMD's Best. What is unfair is the TESTs itself. Toms using some Intel favoured benchmarks...

 

[Diable]

What benchmark will be unfair to both processors? Bandwidth intensive benchmarks will favor the P4 and raw cpu benchmarks will favor the XP and as far as I can tell most if not all current benchmarks fall into one of those two categories.

 

[imgod2u]

The P4 was never meant to compete clock for clock and comparing it so is not feasible. A lot of people only take the MHz Myth one-way. By that, I mean they keep saying MHz doesn't matter when people mention the P4 has higher clock, but discard that when they mention IPC, saying the Athlon beats the P4 clock for clock. By people, I mean AMD zealots. I use to applaud AMD for their efforts to educate the average consumer that MHz truely isn't everything, but that was before this stupid PR system. I don't see how some people can claim it's a way of trying to "re-educate" people. Guys on TechTV are already starting to state that the AthlonXP 2000+ is "2000 MHz". This is not re-education, this is pulling the same (even worse because it's not actually 2000MHz with just lower IPC) trick as Intel did with the P4 (although to be fair, the P4's higher clock over-runs the lack of IPC making overall performance higher).

 

[Texmaster]

The problem with Tom's article guys is he didn't overclock the DDR Ram as you can on many i845 chipsets.

That is why the P42600 sometimes beat the 3000.

The 2600 benchmarks on the i850 are reliable, the 3000s are not.


Proof: Amused one's top 3dmark2001 score so far:

9498 P4@2400, 265/602 t500 i850

Texmaster's top 3dmark2001 score so far:

9522 P4@2700, 265/590 t500 i845

Clearly, Tom tried to pull a fast one of us with the 3000. He did not overclock the DDR ram.

 

[bigshooter]

I think that the only reason they even used DDR ram is that RDRAM doesn't overclock well. They needed to increase the fsb of the P4 to get up to 3Ghz, so they needed something that could handle the 133 fsb.

 

[christoph83]

Yes the RDRAM got stuck at 2.6ghz. Although the point was the more bandwith the better for the P4 , he didnt show the P4 with a P4B266 motherboard. It has the option to go 3:4 or 100/133 on the cpu and memory. Right now im at 160FSB on my memory and my DDR scores are much better. They are edging up to 2500/2500. Thats 25% more bandwith than what he showed for the DDR setup. It had 2000/2000 memory because the 3:4 option must had not been available on that board. My DDR combo has slightly higher bandwith than normal PC800 and can overclock a little better than RDRAM. It's kinda toss up, better overclocking and slightly lower bandwith or slighty worse overclocking and slightly higher bandwith.

About this comparison. If you want a price comparison the 2ghz northwood and 2000 XP are very close in price. The 2ghz northwood can hit 2.4ghz pretty easily. The athlon XP hits 2300 +. So these are two evenly price chips and the 2.4ghz beats the 2300 XP in a majority of the test. So which one would you buy?

Formulav8: Whats the deal? Where exactly does intel rip off people. Box makers continue to build and sell P4 systems and people continue to buy them. If they were getting ripped off wouldnt they stop buying them? If you think they are getting ripped off on performance, wouldnt they be able to tell? People havent so far as they cant tell a difference when getting beat by 10fps in game or a few seconds in decoding. And yet they have the stabilty right from the start with an intel cpu/chipset. Where are they getting ripped off again?

 

[Sid03]

<< Because intel had to make the pipeline so long and the ipc low to rip off the unknowing consumer >>

well you must really hate amd then!

i don't know if you've noticed, but they call their athlons by a higher number than the actual mhz (and hide the actual mhz). talk about trying to rip off the consumer! and don't give me "educate the consumer", because nowhere in best buy, at the local pc shops, or online vendors is there any clarification of their numbering scandal.

it was a scam when cyrix did it and it's a scam now.

look, i really like amd's cpu's... but let's not be so full of blind bias and talk about how intel is trying to rip off consumers.

 

[formulav8]

I am not bias adgainst Intel. But I do prefer Amd do to price/performance. And to me Amd's cpu's are more fun to mess with do to the multiplyer being unlocked. Amd's ratings are much more accurate then the P4's ratings. So, no Amd isn't ripping off the consumer. Intel is taking advantage of the people that look at mhz to justify the performance. Amd's rating doesn't say that their chips are 1900+ mhz or 2000+ mhz.



Jason

 

[josphII]

actually i think there are only two good tests.

top of the line vs top of the line
and
comparing two processors that cost the same

why somebody would compare one company's top of the line processor with that of another company's middle of the road processor is beyond me, well, unless the two chips cost the same and they wanna see how much performance they can get for their dollar

 

[Sid03]

you know full well that when amd calls their cpu a "2000+", they are inferring that it runs at 2000mhz. they are hoping that the guy walking into best buy sees both an intel running at 2000mhz and an amd 2000, and considers them equal.

the average guy walking into best buy will find no explanations about how they came up with this number. nor will he find any references to the actual clockspeed.

this is blatent misrepresentation to the consumer. amd makes great chips, and it's very sad to see them resort to such tactics. and it's even sadder to see the sheep who follow along, especially when they start referring to their overclocks in the pr rating.



<< I am not bias adgainst Intel. >>

i call BS on that one! i've read a good number of your posts, and you are one of the worst here. the fact that you think it's ok for amd to deceive consumers, and yet say that intel is "ripping them off" speaks volumes.

So, is mhz important or not?... make up your mind. imho, mhz isn't important. how the cpu performs, how much it costs, and how stable the platform is, are all that's important. i couldn't care less if the fastest cpu (for what i want it to do) was 4000mhz or 400mhz. but i do care that the corporation selling me that cpu is honest about what it is.

i own an athlon (not an xp) and i really like it. but i won't buy another as long as the company lies to consumers.

 

[josphII]

<< I am not bias adgainst Intel. But I do prefer Amd do to price/performance. And to me Amd's cpu's are more fun to mess with do to the multiplyer being unlocked. Amd's ratings are much more accurate then the P4's ratings. So, no Amd isn't ripping off the consumer. Intel is taking advantage of the people that look at mhz to justify the performance. Amd's rating doesn't say that their chips are 1900+ mhz or 2000+ mhz.



Jason >>



well given that Intel chips are performing damn well, your conclusion that "Intel is taking advantage of the people" is cr@p
the only company trying to scam the consumer is amd with their naming convention

 

[Wolfsraider]

hello everybody!

Can I ask some questions here?

How can you compare 2 processors so obviously different to each other,
yet expect 1 to be better than the other?

we all know the Amd processors use; more L1 cache, less L2 cache,and a smaller set of instructions per clock,etc...


And we all know intel use;less L1 cache,more L2 cache,more instructions per clock.and sacrificed ipc etc...

am i wrong therefore in assuming; any benchmarks comparing these two fine processors
would show differences in their strengths and weaknesses across the board as to their real world values?

And as such would it not be expected that as different as these two are,(almost opposites),
that results should be more varied?

That said isn't it funny that as vastly different as these two companies processors are, that real world tests show that they are remarkably similar?

So much so that differences are nominal EXCEPT to those choosing which one to buy?

And if they are so similar then whats all the fuss just go with the one you feel is best and enjoy.

personally though I look forward to what new toys they both bring to the world these next few years as the goodies list is getting.... longer lol

peace all

 

[FishTankX]

I but need to remind everyone here that the G4 slaughters the P4, *and* the Athlon clock for clock in almost *anything*.


Phtoshop test:Which beats the other? 800MHZ G4 or 1GHZ Athlon? OH MY GOSH! *DONGDONG!* The Athlon just got the beating of the millenium. Look at that G4 go! It gets even more embarrassing on a dualie platform. Ever compared Dualie G4 1GHZ to Dualie Athlon 1GHZ? It's just shameful. Can anyone say that the G4, in photoshop, beats the Athlon by nearly *50%*?!!

Clock for clock is so totally pointless. Even absolute performance is so totally pointless. It really depends on 3 things. now, this covers *any* CPU on *Any* platform for *anything*

1.What you do
2.What do you need?
3.What are you willing to spend?

If you're a 100K a year engineer who will get a 3% speed increase in his favorite program by using the Northwood 2.2GHZ over the AthlonXP 2000+ then that would be *totally* worthit. That's not more than 150$ per percentage point. When it's work your talking about... money don't matter. Thus, #2 is the dominant factor. G4 for photoshop, P4 for Caligari truespace, AthlonXP for 3dMax (That's a tossup, the Northwood P4 can get somewhere near the AthlonXP in 3dMax, SSE2 optomization might be responsible for that)

For budget conscious gamers, (UNLIKE ME! P4+Radeon8500 rock, even if the P4 cost me a pretty penny and gives me less performance, it's quieter. My friend is *SO* bothered by noise that he turns off my P4 whenever he gets the chance. He'd probably ban me from the office if I used an Athlon. Anyways, #3 is by far the most important. is again the most important. AthlonXP, when you're talking about dollar per percentage, can beat the P4 by nearly 50%.

For people who need settop boxes, or quiet boxes and DivX boxes, the C3 wins a unique victory. The ability to run without a fan. The ability to be totally silent (you can use a 100W fanless powersupply+PM133). For people like *US* who cherish silence and don't need excessive performance, the C3 wins a unique victory. #2 is the most important.

I hope this has englightened some of the less informed around here. Another day, another post.