Replies to post #17390 on Advanced Micro Devices Inc (AMD)

[yourbankruptcy]

wbmw, I agree with some points of your post except this: I think the architecture will grow to the point where it can start replacing Xeon, which is a very sizable market, and that's the key of my point.

Today, I see nothing that indicates that IPF can be used where Xeon is used today.

The criteria is very simple: where there are plans or indications that in the future it will be possible to buy a $5,000 Itanium server with speed equal to $5,000 Xeon or Opteron server?

I just went to dell.com and configured dual-Madison PE3250 2U server. Ups, it was $20,815.

This server is $15,815 and 1U off the Xeon and Opteron market.

[sgolds]

wbmw, IPF comments -

In fact, it doesn't take an IPF advocate to tell you that Intel has larger plans for IPF then to just take control of the aforementioned products.

Plans are one thing, execution is another. Intel doesn't make servers so they can't do that alone. Outside HP and some supercomputer design wins, I just don't see where the IPF penetration is happening. There are other vendors offering IPF, but they look more like checklist offerings, not strategies.

I think IPF has been gaining against Power LOL!! and SPARC, too, which is why IBM and Sun have taken such defensive positions.

Sun is being creamed by IBM's Power servers. You can't show me any significant market movement from Sun to IPF, but the numbers consistantly show IBM gaining while Sun loses. How much IPF does IBM ship? No, if you are going to support this claim then you will have to show a pretty detailed marketing report that breaks down server sales by architecture, or show that HP is making big gains with their IPF line. So far I see neither.

In a couple years, I think the architecture will grow to the point where it can start replacing Xeon, which is a very sizable market.

You'd better hope that neither of these things happen: Opteron takes off and replaces Xeon, or Intel produces a 64-bit Xeon. Your statement sounds more like wishful thinking than market analysis.

Some true-blue Intel shops will bite the Itanium bullet, I am sure. There undoubtably will be more x86-dependent companies who will opt for a form of x86-64, and right now there is only one solution for them.