Burnt to a Crisp :D :D

[formulav8]

<< Mwahaha.. ok.. i'm not a staunch intel defender, but intel has been working on getting it's secret weapon out that will blow the AthlonXP2000+ away in terms of performance. It makes a 2.0A beat a 2100+ Athlon in terms of performance. Infact, it could make a 1.8A beat an AthlonXP 2100+ in terms of performance. Ready for the bomb?

SSE2.

If SSE2 optomization is used properly, the P4 leaves all of AMD's offerings far and away behind, atleast in current society. And you better pray to God, all you AMD supporters, that SSE2 takes off. Pray very hard. For there is one secret weapon that AMD is gonna use to even the battlefield.

Stealing SSE2 to use in the hammer.

Intel aint publishing FUD, they're just showing off they're P4's awesome capabilties once properly optomized.

So all of you nay sayers that say that there's no way the P4 could perform like that...

Welcome to
"Streaming Single Instruction Multiple Data Extensions V2.0" >>




How is it considered stealing SSE2 when Intel is licensing AMD to use it? Care to tell me? Also Intel must be a COPY_CAT since AMD had 3DNow! out before Intel has its SSE out? Care to exaplain? Pretty sad Intel needs SSE2 to make the P4 look half way decent IMO.


Jason


Jason

 

[AGodspeed]

 

[FishTankX]

Okay, FormulaV8, I admit, steal isn't the right word. I guess borrow would be more apropriate. And that's like saying that the Athlon needs it's 2 extra floating point units to beat the P4 in floating point apps. Is that right? no... it was designed with floating point power in mind. The P4 was designed with SSE2 in mind. And remember, if Intel is a copy cat of AMD, then AMD is a copycat of IBM and HP that had SIMD long before anyone else..

<<Pretty sad intel needs SSE2 to make the P4 look great IMHO>>

I think it's pretty sad that the Athlon needs 3 FPU pipes to pass the P4 when the P4 only has 1. I think it's pretty pathetic that the Athlon needed 3 FPU pipes to attack the PentiumIII when the PentiumIII only had 1 FPU pipe. That's downright sad.
(note, that was sarcasm. I wouldn't be caught dead using a P4 in 3dmax. What i'm saying is that your statment is rediclous. It only takes a recompile to integrate SSE2 and nearly double the P4's floating point performance. Don't knock a processor for using something that it was designed with in the first place. It's like saying that the P4 sucks because using SDRAM with the P4 cripplies it's performance. The P4 really wasn't meant for X87FPU apps.. it was meant for SSE2 optomized apps. Ever seen the P4 rule in lightwave?)

 

[CurtOien]

Something like that almost makes me want to never buy Intel.

 

[sandorski]

FishtankX: That's fine and dandy, but consider that the Athlon has fpu power out of the box. P4's SSE2 must be implemented by a third party, software makers. About a year ago, a number of major game publishers/makers were asked about 3dNow/SSE/SSE2 and they all agreed, they wouldn't optomize for any of them. So far Intel has made up some ground with the Northwood, but they did it through cpu improvements and not SSE2.

 

[FishTankX]

Yeah, the athlon does have more FPU power out of the box. But soon, with *all* highpowered CPU's running SSE or SSE2, the optomizations will have to begin somewhere. And intel is making a compiler that will automatically add SSE2 optomization in there, so it's only a matter of time before the P4 stretches it's leg. Do you think a CPU with only 1/3 the MFLOP rating of another CPU keeping up is impressive? I think so...

P4 2.2GHZ~12000MFLOP
AthlonXP 2000+=4000MFLOP if I remmeber right.. atleast 3000..

 

[formulav8]

<< Yeah, the athlon does have more FPU power out of the box. But soon, with *all* highpowered CPU's running SSE or SSE2, the optomizations will have to begin somewhere. And intel is making a compiler that will automatically add SSE2 optomization in there, so it's only a matter of time before the P4 stretches it's leg. Do you think a CPU with only 1/3 the MFLOP rating of another CPU keeping up is impressive? I think so...

P4 2.2GHZ~12000MFLOP
AthlonXP 2000+=4000MFLOP if I remmeber right.. atleast 3000.. >>




You've got to remember that the P4 is at a freq. of 2.2ghz while the Athlon is 1.64? or something like that. So, if you compare flops in that manner compare them at a similar freq. Also, you don't think those apps will benifit the SSE2 in the Hammer just as much as Intel? I think you will be quite surprised. Intel won't have the long leg that you are expecting.


Jason

 

[ssanches]

IMHO Intel is definitely spreading FUD in that document. AMD's Performance Rating is definitely a success by far. I personally dislike Intel's advertisements in the media stating "you can stay ahead of more demanding applications by using a P4." I feel Intel is a sore loser. Things will get worse for Intel once the Hammer comes out. AMD has definitely underplayed it's performance rating in the XP chips and it's based them on a more realistic method, by using real-world benches.

 

[ssanches]

<< Something like that almost makes me want to never buy Intel. >>



Me too :disgust:

 

[joohang]

LOL.. wtf?

If Intel engineers had the brains of their marketing department, I'd never buy an Intel CPU ever again.

But looking at Wingznut and pm, they clearly are intelligent folks.

 

[Diable]

Dose any else see the hypocrisy of this thread besides me? AMD picks some benchmarks that show its slower processors perform equal to or better then faster Intel processors. AMD didn't test every single app or game currently available before decrying their 1.6GHz XP is equal to Intels 2.0GHz P4, they just ran their hand picked suite of benchmarks and declared it fact and all of you believed it. Intel then picks a suite of benchmarks, runs them and inturn declares that AMD's processors are indeed slower then theirs and all of you start screaming FUD and that's bullsh@t.

Intel did exactly what AMD did. They both picked benchmarks that favored their own processors and neither should be believed IMO as fact. Is Intel spreading FUD? Unless someone reruns the benchmarks used in that PDF and proves they lied the answer is no.

 

[AGodspeed]

 

[Yossarian]

It's just marketing. The people who wrote that document have no interest in truth, they just want to give you reasons to buy Intel. I have no doubt AMD would do/has done the same.

 

[Diable]

<< Intel PDF uses obscure benchmarks and benchmarks that aren't indicative of real world performance (as well as controversial benchmarks). >>



AGodspeed, now that line is pure grade A FUD. What so obsure about these benchmarks:

SysMark - AnandTech, Tech Report, THG, Sharky Extreme and others uses this
SPECviewperf - AnandTech, Tech Report, THG and others uses this one to
Photoshop - GamePC, ZDNet and every site going a Mac vs. PC bake off uses this
Quake 3 - This ones is truely obscure
Divx4 - AnandTech, THG, GamePC and others do Divx4 tests. Intel's use of MAGIC video deLux instead of Flask, Xmpeg or Vub has no effect on benchmarks because the codec is the same and it does the work.

The WebMark and MAGIC MP3 Maker test are indeed obscure but neither you nor I know if they aren't indicative of real world performance, just because you havn't heard of them dosen't make them bad. If you don't trust Intel or their numbers just say it, don't act as though AMD is somehow more truthful because you don't who ran the benchmarks at either company or what they were thinking.

 

[SteelCityFan]

<< It's also interesting that Intel failed to mention the shift in processor types (P3 to P4). When AMD used Mhz to claim superiority, they based it on the P3 which had similar IPC and pipeline length when compared to the Athlon. The Pentium 4 had lower IPC, but could run at a higher frequency due to the 20-stage pipe. By now, everyone knows what this does to overall performance, even Intel. AMD didn't shift to the PR ratings until it became clear that people thought raw Mhz equaled total performance...which it doesn't. >>




I thought the Athlon has 15 stages and the P3 had 10? I think I remember reading the 15 number somewhere.

 

[AGodspeed]

 

[ST4RCUTTER]

I thought the Athlon has 15 stages and the P3 had 10? I think I remember reading the 15 number somewhere.


You're landing your balls close to the flag...

It's 10 stages for the integer pipeline and 15 stages for the FPU unit. Here's a pretty good diagram of the pipeline's structure.

 

[ssanches]

<<
Intel did exactly what AMD did. They both picked benchmarks that favored their own processors and neither should be believed IMO as fact. Is Intel spreading FUD? Unless someone reruns the benchmarks used in that PDF and proves they lied the answer is no. >>



AMD NEVER did what Intel is doing. Heck if AMD had a marketing budget like Intel's, everybody would have an Athlon XP. AMD uses benchmarks to highlight it's own performance, not to ridicule someone else's benches. AMD uses many benchmarks that are optimised for Pentium systems and others that are made by Intel's "bed-partners" like BAPCo. If I were given a choice between a AXP 2000+ and a P4 2.0A, I'd prefer the former anyday. This is just another lame attempt by Intel to defeat competition. Intel has NEVER done anything good for the consumer. It's only tried things that either make tougher for the competition or impose unjustified costs on the consumer(like RDRAM, modifying sockets, etc). This squarely falls into the definition of Monopolistic Trade Practices. IMHO, Intel is just a monopoly awakening to a world where it's monopoly would be no more, so it's running around scared and throwing FUD on the competition. Just wait till the CH comes out, then things shall become clearer.....

 

[JellyBaby]

<< Intel has NEVER done anything good for the consumer. >>

Utter BS. Even now they're playing consumer rights advocate in fighting Hollywood's draconian SSSCA proposals. Of course it's in Intel's best interests to battle against SSSCA, too.

Big corporations greatly fear competition. They'll do anything short of (and sometimes even) illegal to retain market share, hefty profits, mindshare, etc.

 

[apoppin]

After reading and rereading this "Intel Confidential" report, I am left with the feeling that it is not genuine . . . it just doesn't have the usual Intel marketing slickness . . .

I enjoyed it anyway.

 

[ssanches]

<<

<< Intel has NEVER done anything good for the consumer. >>

Utter BS. Even now they're playing consumer rights advocate in fighting Hollywood's draconian SSSCA proposals. Of course it's in Intel's best interests to battle against SSSCA, too.
>>



That's BS too! They're aren't fighting for the consumer! They are fighting for themselves. They too have much interest in the ruling as consumers like us have. Yes, I still say they they haven't done much good for us consumers. They force us RDRAM when the industry slated DDR-II for mainstream PCs by 2002 way back in their roadmaps of end 1999. They change sockets more often than they change their underwear. If it wasn't for AMD, Intel would still be selling sub-gigabertz CPUs. They are one slow moving Goliath They are definitely a monopoly that hates competition...

 

[NaughtyusMaximus]

<< That's BS too! They're aren't fighting for the consumer! They are fighting for themselves. They too have much interest in the ruling as consumers like us have. Yes, I still say they they haven't done much good for us consumers. They force us RDRAM when the industry slated DDR-II for mainstream PCs by 2002 way back in their roadmaps of end 1999. They change sockets more often than they change their underwear. If it wasn't for AMD, Intel would still be selling sub-gigabertz CPUs. They are one slow moving Goliath They are definitely a monopoly that hates competition... >>



Don't feed me crap and tell me its ice-cream. If Intel doesn't fight for themselves, we (consumers) never benefit. Look at the big picture.

 

[Midnight Rambler]

SHEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEESSSSSSSSSSSSSSSHHHHHHHHHHHHHHH !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!


Will you people never quit it with your AMD vs. Intel (or Intel vs. AMD) trolling ?

 

[Sunner]

I don't see what the big deal is?

It's marketing, if it's even genuine.

Back in the day of K6 vs P2/3, I remember AMD had a bunch of benchmarks on their site showing the K62 being supperior to the P2 in performance, and we all know, aside from a few select cases, thats not true.

Take it for what it is, marketing BS, marketing is all abut BS, no matter who's marketing it is.

 

[Pabster]

As usual, welcome to Flame Theater

What is so surprising about a manufacturer selecting benchmarks and testing environments which favor their product(s) in a better light? Might it be misleading? Sure. But it happens all the time.

AMD carefully chooses benchmarks and configurations in the same way.

Guys, right now the 2.2GHz P4 Northwood walks away with the speed crown, even at default clocks. With a little boost, nothing comes close. Who knows what Thoroughbred and Hammer will bring? That's the beauty of the market, and all the competition is great for us consumers. Even priced at $485, that is half (or less) of what Intel would charge if AMD hadn't burst on to the scene with the Athlon.

In closing, I wouldn't exactly call the PDF "FUD". If you are to do that, I could easily find a comparable set of figures on AMD's site (where they often measure performance in percentages rather than actual figures) which could easily be construed as "FUD" one way or another.