Replies to post #1503 on Intel Corp (INTC)

[Yousef]

Charles,

Re: " ... As to the reason why they (AMD) have been floundering around the
last 18 months or so, I don't know ... It can't be something as simple as "their
.13 is broken" because the delays were widespread and started way before the .13 process."

The reason that AMD is "floundering" is most likely due to their .13um
process and risks that AMD took with SOI/Copper/LowK ... Sorry, but this
is "most likely" the TRUTH. No one can seem to get "the truth" from AMD. <ggg>
The foundry industry development that should worry the 'Roids the most is the
announced slips by foundries (TSMC,UMC,CHRT) at 90nm. AMD will definitely fall
further behind INTC ... INTC is NOT waiting for the foundries at 90nm.

Make It So,
Yousef

[wbmw]

Constantine, Re: Anyway, all of AMD's products that I mentioned are not "powerpoint" products but real products that have been demonstrated at numerous venues.

...which means nothing, of course. Not to trivialize the importance of showing a "proof of concept" demonstration, but functionality need not be complete to get to this phase. Intel demonstrated Madison silicon running on modified hardware back in the Spring, yet that product won't be launched until Q2 next year. Intel even had it running on popular database software (I believe DB2 from IBM), and it ran without a hitch. In the Fall IDF, Intel demonstrated hot-swappable nodes, allowing someone to upgrade the processors without ever turning the machine off. This is a very high end RAS feature, yet it worked great in a 32-way SMP system. That demonstrates high confidence in silicon, yet it was still at least six months away from a product launch. The point is that demonstrations are nice as a "proof of concept", but it doesn't mean that the silicon is healthy enough for launch. As far as I know, Intel doesn't have any .13u manufacturing problems, so I can only assume that Madison is still exhibiting functional problems with rare corner cases that only pop up under extreme conditions. AMD may be at that point with Hammer, but that's usually the longest and most difficult part of the validation process. So I would not rule out functional problems in Hammer, even though it has been demonstrated multiple times.

wbmw

[Nitt]

Re:"Anyway, all of AMD's products that I mentioned are not "powerpoint" products but real products that have been demonstrated at numerous venues. The only valid question that you have is whether AMD can manufacture these things. My guess is yes."

I'm not sure how a "demonstration" means a lot. This chip will compete with Pentium on the desktop first if AMD expects to get enough volume to matter. It's not clear the chip will matter on the desktop from a performance standpoint because it is not clear there will be any 64bit code that will make people care. The server space will take time to get a foothold in. Network owners do not take risks... it will take time to get SW in place and then it will take more time to get users to evaluate it completely and then deploy (24 months would be fast). You have to ask yourself, how compelling will Hammer be?

I also believe that despite the press releases regarding support, most companies "supporting" Hammer are really watching their spending and waiting to see if AMD can get it out. The first versions of any SW will not be best in class out of the gates… and some will be beta. There will be less support for Hammer than there is for Athlon... less motherboards, less chipsets, less OEMs offering systems. This economy keeps everyone risk averse... that's why Intel's big support machine adds so much value in the worst of times and AMD has no answer for that.

Nitt

[Elmer Phud]

Constantine -

Anyway, all of AMD's products that I mentioned are not "powerpoint" products but real products that have been demonstrated at numerous venues.

Many of us distinguish between prototypes, engineering samples etc and real products. It must be manufacturable to be a product.

Every single product introduction in the last 18 months has been delayed with no explanation given. It seems to me that must have some internal bs going on that we don't know about. It can't be something as simple as "their .13 is broken" because the delays were widespread and started way before the .13 process.

You are correct about the continuous delays but it's not true to say no explanation has been given. The have constantly explained by saying yields are fine, yields are the best in the world, Hammer is on schedule, Barton is on schedule, .13u is on schedule etc etc. The problem is that the explanations don't match up with the facts. Yields were fine but the fab couldn't produce half of what it should have been producing. The .13u process is fine but Athlon won't scale. Hammer is on schedule but only sampling at 800MHz. You name it, they have an explanation, it's just doesn't wash.

EP