AMD’s Executives Admitted Inability to Compete – Intel.

[mav451]

Yeah I think I misused validation when I said that, heh. I'm well aware of the advantage of offering a complete platform. AMD fans (I used to be) suffered through the Via 4in1 nonsense, and b/c of that celebrated the nF2 when it was released. Of course I think AMD realized that, sooner or later, relying on a graphics company for chipsets just wasn't going to be a good end-game strategy.

 

[Gunbuster]

How do you compete when the competitors R&D, testing, and valadation budget could buy your company...

 

[alyarb]

VIA wasn't that bad. All my VIA boards would still start up today if I had any, and towards the end of Socket 462's life they really put up some decent chipsets culminating with KT266A->KT400 whereas most of the nforce4-based K8 machines I built are now dead. The 4n1 became a compact unified driver that installed just as easily as nforce. During the K7 generation you had consistently lower prices for the midrange parts from AMD. And, even for netburst standards, were able to see some pretty high clocks from AMD's process on 180 and 130 nm. AXIA Thunderbirds consistantly did 1.5 GHz and up, which was the fastest single socket machine at any price in early 2000. Thoroughbreds and Bartons at 2.5+ GHz were nothing to sneeze at either particularly compared to their 2.4-3 GHz contemporaries. The K7 really came a long way from where it started; from the very beginning it maintained comparatively larger cache and FPU resources and it sat on the faster bus. Having the big L1 was particularly badass when L2 was off-die and half speed, and i'm glad they've stayed with it. I think we can all agree that netburst relied entirely on SSE/SSE2 and in the beginning there was about a 16 month period where the implementation of SSE2 was quite slow. Just one of the mistakes coming out of the early days of netburst, causing it to really suck in the few apps where CPU performance mattered. A wilamette chip on a proper SSE2 encoder made today would probably run better than some beta of FlaskMPEG from 2002.

Moving away from 130nm, the Pentium 4 only got worse and K8 only got better, and prices were dramatically adjusted to reflect this, but back in the days of K7 and VIA, AMD had a superior product made with a fraction of the resources spent at Intel, and it seemed hardly anyone knew about it (no TV advertisements like intel and very little OEM penetration). For the typical multimedia machine, the pentium and athlon are practically indistinguishable, but content creation is a completely different story. AMD can still contend with intel today while their product is significantly more affordable for consumers, despite the fact that it is now intel with the better architecture.

 

[v8envy]

The problem with value pricing is: nowadays chips are cheap as... well... chips. The absolute dollar difference between a $99 value chip and a $150 cream of the crop chip is huge percentage wise, but really not that big of a deal in terms of complete system cost.

Sure, when the AMD 3800x2 was $360 and the P4 D 805 was $250 you could argue bang for buck. But when the difference of cheapest quad to a much better quad is $50 out of your whole system budget of $500-1000 it's harder to justify the low buck product.

 

[v8envy]

The problem with value pricing is: nowadays chips are cheap as... well... chips. The absolute dollar difference between a $99 value chip and a $150 cream of the crop chip is huge percentage wise, but really not that big of a deal in terms of complete system cost.

Sure, when the AMD 3800x2 was $360 and the P4 D 805 was $250 you could argue bang for buck. But when the difference of cheapest quad to a much better quad is $50 out of your whole system budget of $500-1000 it's harder to justify the low buck product.

 

[her209]

They weren't dubbed 'Preshott" for nothin'.

 

[Zebo]

With idiots like this working for them no wonder they are second best. Anyone who bought Intel 04-05 time frame when this statement was made was a total fan boy. AMD dominated performance, price and chip-set with nV's offerings. Today you can still make a strong argument for AMD based on price to performance ratios. 99% of people don't need the chips we have and would be well suited on cheap AMD's offerings. Certainly 'no reason to buy them'

 

[Edgy]

With idiots like this working for them no wonder they are second best. Anyone who bought Intel 04-05 time frame when this statement was made was a total fan boy. AMD dominated performance, price and chip-set with nV's offerings. Today you can still make a strong argument for AMD based on price to performance ratios. 99% of people don't need the chips we have and would be well suited on cheap AMD's offerings. Certainly 'no reason to buy them'

I'm not entirely sure I would agree with you there...

I've met Henri couple of times and he's far from an "idiot"... I didn't like him much (something about Sales/Marketing personnel, even if they are executives...), but he was definitely not the one of the worst types.. and he was knowledgeable.

Actually the article states that these comments were taken from AMD's INTERNAL communication during 2001-2007 I think...

If I was AMD, I would think that top management ACCURATELY criticizing the shortcomings of the company internally is good - and if that management works toward rectifying them, then even better.

AMD did have reputation as pushing inferior cheap products back then - true.
AMD did try to compete with Intel with CPU products only while Intel had advantage of complete platform product breadth - true.

As to the comment regarding his personal purchase would not be AMD products, I might be stretching things bit but that may have very well been a comment to get his point across to fellow executives to convince them to join him in doing something about it and a comment taken out of context.

Has AMD's reputation improved since as providing more quality products? Yes.
Has AMD found a way to increase breadth of products it controls to establish competitive platform offerings? Yes.

Was anything Henri Richard did at AMD directly or indirectly influence these results? I think we do not have enough evidence to conclude that but surely his comments which were made public during Intel's response to FTC is at least an evidence that at minimum, 1 executive at AMD during those times was aware of these shortcomings.

And yes - perhaps it was not the smartest thing to put such comments in writing even if it was internal, but back then subpoenae for internal electronic communication were not at the top of every corporations' concern list...

 

[Arkaign]

The problem with value pricing is: nowadays chips are cheap as... well... chips. The absolute dollar difference between a $99 value chip and a $150 cream of the crop chip is huge percentage wise, but really not that big of a deal in terms of complete system cost.

Sure, when the AMD 3800x2 was $360 and the P4 D 805 was $250 you could argue bang for buck. But when the difference of cheapest quad to a much better quad is $50 out of your whole system budget of $500-1000 it's harder to justify the low buck product.

PD-805 was never $250.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2736

The difference was HUGE. AMD had the better product with the X2, and felt perfectly comfortable charging more than double.

~$135 vs. $300+. Ick.

$165 saved buys a lot more GPU.

Of course a stock PD805 vs stock 3800x2, the x2 was just a plain faster chip. Overclocked to an easy 3,5ghz+, the 805 was suddenly fast enough for any game at the time, and that extra $ could go towards a nice x1900 or maybe even 8800gts

 

[Decembermouse]

And at what cost? Performance without price is a worthless comparison. You've consistently ignored price since the beginning of this discussion.

Tom's Hardware does wayyy too much of this in their reviews. "Let's see which is better if they're all at 2.8GHz, the i7 or the Phenom II! Aaaand the i7 sweeps the floor with the PII! Intel WINS!"

 

[jones377]

Tom's Hardware does wayyy too much of this in their reviews. "Let's see which is better if they're all at 2.8GHz, the i7 or the Phenom II! Aaaand the i7 sweeps the floor with the PII! Intel WINS!"

Personally I'm glad there is a site doing those kinds of articles. It's the only reason I read that site maybe once a quarter or so.

 

[alyarb]

i like their charts as long as they include the right test subjects. just don't read their text and there's certainly no reason to read the "conclusions" page.