Even at 2GHz....

[Kosh Naranek]

Almost forgot ......

Anandtech has also posted up a review that is even more favorable to the Athlon thanks to them using a SiS 735 based board. Take that *ntel liars!

What ever do they mean ...... ?

Oh I get it ... so now it's the platform which is to blame for the speed and not the CPU.

I suppose SIS 645 don't mean diddly ! LOL

About the Review on Anandtech .... I do not get those MPEG4 scores ..... and I find the part about no performance difference between MMX iDCT and *intel iDCT very weird.

[superfly]

The thing to remember as well is that the existing version of the p4 differs quite significantly from the original specifications of the chip,which was originally intended to be built at 0.13 micron.

But since intel was forced to release the chip early and build it at 0.18 micron but had to remove quite a few features in the process,to keep the chip from being even bigger than it already is...

I remember seeing an article on the original specs and it mentioned that the chip was to feature(among other things) a 16kb L1,128 kb L2 and a 512 kbL3 cache...

The existing one cuts the L1 in half and doubles the L2 to 256...


I'll wait and see if the northwood version of the p4 ends up integrating the features that were part of the original desing....

[Dr Mordrid]

KOSH,

Go to the top of the d**m page and read Joels post that started the thread. My post about Intels updated compiler was totally in context with Joels statement that;

======
AMD is still able to hold it's own with Intel.

Found this at www.amdzone.com.....

(article about P4 vs. Athlon performance)
======

The reason is that current software is not compiled with P4 optimizations, not any real advantages of the Athlons. The same thing happened when the PIII came out. As compilers were updated things changed.

Sheesh.

Dr. Mordrid

[Bubba Feelgood]

I'll take Anand's review over Tom's any day. Tom is so busy patting himself on the back he's liable to get repetitive motion injury.

[Topha]

maybe the reaseon the palomino is so late, is that amd wants to sell most of the tbirds first, that explains the drastic price drops over the last few weeks.

and since they dont have to decide wheter they make a palomino, mobile or whatever cpu until the last production step, maybe they are just stocking up enought of the different cpus (mobile, desktop, mp (i think i read somewhere (toms???) that the desktop palomino will not be a mp cpu, i hope thats wrong)) before they start selling them.

i just wish they would have started using 13mikron technology earlier...

[Dr Mordrid]

============================================
I'll take Anand's review over Tom's any day. Tom is so busy patting himself on the back he's liable to get repetitive motion injury.
============================================

OUCH!!.....but oh so true

Problem is you could target some arrows at Anand too. Some of his test procedures at times leave something to be desired.

Dr. Mordrid

[mdhome]

For some of you who are a little confused about the subject then there is a good article (with a rich message board content ) over at Aceshardware. It gives the rundown on both the Athlon and P4.


regards MD

[Bubba Feelgood]

Ain't competition great? If we only had 1 CPU manufacturer we'd all be running 300MHz systems costing 3 times as much as we pay now.

[Kosh Naranek]

http://213.219.40.69/29080104.htm

This really sucks .... AMD must be in real deep shit to pull something like this !.


( I can almost predict the responses to this )

[Greebe]

@Kosh

*sigh*

[Indiana]

Kosh, this kind of sh*t is produced by Intel, releasing CPUs with about 30GHz that perform like, say 500MHz ones (pun intented). You know that the *current* P4 is a joke and even Intel has said that they'll leave the MHz over performance path in the future. And why equip a CPU with additional CPU/FPU cores that can't actually be used at the same time due to design flaws (or is this a "feature"?).
And I don't really believe that the "AMD can't get more than 1.4 MHz" story is true 'cause the AYJHA-Y steppings seemt to constantly reach 1.6 -1.8GHz without the need for excessive voltages or cooling.
When Intel releases the *final* P4 the picture might change but until then AMD is in the lead by now for anyone that's only halfway realistic.

And for a stable Athlon chipset, how about the VIA KT266? I know it doesn't perform much better than the KT133A but most people are VERY happy with the KT133A's performance, it's just the (missing) stability that annoys them. I've not yet heard (or experienced) any major problems with the KT266.

[Kosh Naranek]

INDIANA

When Intel releases the *final* P4 the picture might change

My point exactly ... but the current P4 is ALWAYS considered to be the one and only version of the P4 by 99.999 pct of all AMD fans.

and even Intel has said that they'll leave the MHz over performance path in the future

Yes but that's all got to do with the Itanium ! and that may also be the reason why they didn't announce the 1.1 Ghz Celeron although you can buy it in some of the nordic countries.

This put aside I still think this performance rating thing will only confuse matters more like in the good old CYRIX days , their ratings didn't even come close to actual performance.

So who will be the judge of this performance rating system ?
... the average user ? I think not.

Of course those who know will know, but it still leaves a lot of the decision up to AMD whether they will be ahead or behind current P4's in regard to ' performance ' and it would also mean that AMD would have to guess about the performance of not yet released P4's ( if today they decided to rate a 1.6 Ghz cpu as a 2.2 Ghz Northwood P4 )

[Greebe]

This put aside I still think this performance rating thing will only confuse matters more like in the good old CYRIX days , their ratings didn't even come close to actual performance.

Actually not true... the Cyrix chips did hold their own when it came to the typical business app using the PR rating system. Additionally Intel forced the Performance Rating system on Cyrix through their licensing agreement.

AMD rewrote the book on how the front end was addressing the code and thus didn't have to abide by the PR system Intel had in place... which holds true to this day. But like you, I can't think why they would return to this standard other than there are far more ignorant buyers out there than those of us that understand what the real deal is.

The future has already made it clear that we're speed limited and thus the claims of however many hertz they'll do will take a backseat to how the processor executes code.

Like Wombat has stated many times, take out the slow front end on the Athlon and it's a real scream demon!


And another thing... we have people like ALBPM (and Oboy to thank for the speed increases the upcoming P4 will have. I certainly welcome it!

[Nuno]

I have mixed feeling about the dreaded PR rating return. Mainly because it remembers me of one of the most deceiving feature on hardware, the Cyrix and AMD PR. Sure, they had similar integer performance, but when something requiring FPU kicked, a Cyrix PR 200 performed like a Pentium 100. How lame.

Now what we have is a different story. Looking at Tom´s news they are thinking of "Models" and Palomino 1400 would be "Model 1600" because it would perform similar to a PIV 1600 (I think AMD is being too modest here, if a TB1400 is not always beaten by a PIV 2000, a Pally 1400 should easily keep with a Model 2000).

Truth is that you´re actually getting the same performance between a much lower clocked AMD CPU and a Pentium IV. Period. Price (CPU and platform) is still a huge AMD advantage.

So I know and you know a PIV 2ghz is not much faster than a TB1400, I know and you know Palomino will be significantly faster than a Thunderbird at the same clock speed, I know and you know that a Palomino 1.6 Ghz on a fast DDR motherboard would kick PIV 2000 every time... but... You can bet 90% of computer users (buyers) don´t even know that. Ooooo, 2000 Mhz wow and it´s Intel, wow, I want one.

So I can understand this move from AMD. Believe me if they had a Palomino 2 Ghz ready tomorrow, there still would be many people that would take the Intel chip instead. PIV was a great marketing plot from Intel, because they always knew the average computer users only wants to know 3 things: If it has a Intel "pentium" on it, how many Mhz and how much memory.

So, as far as I´m concerned, they can call it Model 1600/1800/2000, PR 25 Ghz, the bios can display AMD Athon "fast enough", I don´t care. As long as I know what I´m really buying and it costs 4 times less than the similar performing Intel CPU, I´ll buy it.

[Technoid]

Kosh Naranek, What Cpu have you got in your Computer?