Dear TOM! You've made a mistake!

[razvan]

Dear Tom,



Your chart with FlasK MPEG is hard to follow. You shouldn't include the P4 results for SSE2 iDCT MedQ and LowQ since you provide no such benchmark for Athlon. You will see that many of your readers will just look at the best performance of P4 and Athlon and conclude that there is a difference of about

22.85-16.97 = 5.88fps

between P4 and Athlon and that's false. The P4 at 1.5Ghz, SSE2 iDCT HighQ has the score of 18.96 and the Athlon 3DNow! iDCT HighQ has a score of 16.97. P4 wins by 1.99fps! I wouldn't call that MUCH! The HUGE bandwidth of P4 should do much more.
I don't think that you are biased BUT you should think that most people are "result-orientated" and they don't care so much about details. Most people have a visual memory and that means in 2 weeks they will only remember that P4 was way ahead of ATHLON in this bench.
Your article shows Athlon in a very bad light in this benchmark and IMHO it's not the case. If you compare a

P4 1.5 SSE2 iDCT LowQ

and

Athlon 1.466 3DNow iDCT HighQ

what conclusion can you make??! Since you've changed the parameters of your bench the difference has no meaning at all. Please run the Athlon in the same conditions - I mean LowQ and MedQ. Also, if you would run the P4 at higher clocks speeds but in the same condition well, that would be something worth looking into.
About the Bapco tests. You said that in your knowledge is NOT Intel orientated. I'll take your word for that. At least NetBurst is not a marketing trick.
If I missed/don't understand something then I'm apologising. If not I must say I'm a little bit disappointed. You hurry too much lately like in the second P4 article.
And you dear reader please tell me your opinion. If I'm wrong I'd like to be corrected.




Razvan

 

[Kodiak]

agreed... see my post written at the same time
I haven't learned -anything- new from the 4th review... if anything, it left me confused as to the whole point of the exercise...

 

[JoeHead]

jsut a note: I beleive the Athlon optimizations don't have the same options fo lowQ and medQ. I could be wrong. And if people are not paying attention it is unfortunate that bad rumors are to be spread but Tom is not to cater to these people.

I have respect for the PIV today. Prior to the release I had serious doubts due to rumors. I have to commend Intel. especially for not being stuck in the past and looking beyound the 87 optimization towards SSE2. AMD is on it's way to SSE2 also now.

<b> Fragg at will!!! </b>

 

[Guest]

What will be interesting will be to see how the SSE2 enabled, next generation Athlon will stack up to the PIV. Unfortunately, it seems like AMD might be 1 step behind again in the processor wars, just like the days of the K6-2.

 

[razvan]

JoeHead,

"I beleive the Athlon optimizations don't have the same options fo lowQ and medQ."

I disagree. The lowQ and/or medQ use the same alg like highQ but the alg is using a lower resolution. I don't have Flask MPEG so I can't be sure but if you are right then those who wrote Flask are let's say just NOT VERY INTELLIGENT.
Will anyone throw some light here?





Razvan

 

[Guest]

I haven't used it much other than a brief trial that I found was giving me tons of illegal ops. But from what I've seen, it seems that the low vs. high quality is determined by whether it's using integers (lowQ) or floating point (highQ). Please correct me if I'm wrong.

"The answer is not in your hair."
"I'd rather jump in the lava than be fragged by you."

 

[AmdMELTDOWN]

why is it that amd puppies for some wierd reason can't understand that the P4 is a strong contender and it is also a misunderstood processor?

 

[Nikko]

Misunderstood? You've got to be kidding me! Look really, really hard, I'm sure there's something to like about that P4.

ForeverBuckeye - I'm not sure I understand where you're coming from. From my perspective the Athlon, in it's present state, is a better processor than the Pentium 4. It may not clock as high but big deal. In case you Intel droids haven't figured this out, the biggest thing going for the Pentium 4, the thing that allows it to look decent in any benchmarks, is it's quad pumped bus and bandwidth. Put the P4 on the same bus as the Athlon with the same memory and it get's completely, totally stomped in everything! Take a memory channel away from it or give it PC600 and see how it does. It sucks!

The Hammer family will be the final nail in Intel's coffin.

 

[Guest]

Duh... Ah'm sorry, yer raght! Clearly I'm in the wrong... DOOD! P4 ROX!

 

[Guest]

.... I share your concerns about the way Tom presented the mPEG graph, but maybe intel was threatening to pull advertising support or something. ..

.... There really only ONE thing to notice on that last mPEG graph:

.. When both are on High Quality x87 FP only with NO SIMD (SSE2 or 3DNow!) used, the Athlon @ 1.4Ghz outperforms the WhatIV 1.4Ghz !!

.... Not many other 'even' comparisons are possible since 3DNow! has no version comparable to the new SSE2, since AMD is simply switching to SSE & SSE2 now to beat intel at its own game again.

.... The Athlon 1.2Ghz OC'd to 1.4Ghz will set you back $289 now!! The WhatIV 1.4Ghz is $644 ! Then add a precious ii850 MoBo and high priced PC800 RDRAM.

.... The Athlon will run most every app. exceedingly well. Check www.inqst.com for tests of Athlon vs PIV & Dual PIII. The WhatIV excels in Quake3, comes 2nd in mPEG unless you have an intel SSE2 engineer locked in your basement, and falls farther and farther behind in other apps. until it looks like a PIII 600.

.................................................. -Buzz+

 

[RavenPrime]

BuzzLOL makes a good point about the prices but Tom made the same point, with less specifics, in his article.

Looking at the numbers and doing the calculations. For a 90 min movie at 15 fps (frames per second), that is 81000 frames. Times to encode that using Tom's numbers:

P4 1.5 HQ 71 min
A1.2DDr HQ 89 min
A1.466DDR HQ 80 min

A difference of nine minutes each. If you are doing 30 fps, that would be 18 minutes. How significant this is depends upon how long the recording is and its quality.

Bottom line, if you do a lot of encoding or other high memory bandwith work, the P4 may be worth the extra cost. How much is your time worth?

But I said it before and I will say it again. The average Joe doesn't know what a benchmark is, he just wants the coolest looking box with the highest GHz rating. And that's what Intel is marketing. Intel can get away with calling the P4 the fastest processor ever because it operates at the fastest GHz speed ever.

I actually did a survey of people in my office, only two (counting myself) out of 21 knew what a benchmark was!

James

 

[Guest]

I am asking myself... whould not it be nice to see some benchmarks of overclocked P4 CPU running Flask P4 optimised encoding... like at 1.7 GHz?

Bet you the numbers would be really interesting for P4 owners. I am sure you would be talking about amazing performance. Way beyound AMD 1.466 overclocked numbers!

Let see what each of these CPUs can deliver when running at top clock speed!