Klaus, Re: Thanks. 2006 will be the time Opteron most probably will be dual core already (90nm). If it works out like this, possible improvements from a clean IA-64-design would come too late to be impressive in my book, even if it is a LV-quad-core CPU
Wasn't Opteron's motto, "Better late than never?" Trust me, if IPF shows a clear price/performance advantage with Tukwila, it will be successful no matter when it's introduced. Late is only relative to an individual's expectations, anyway.
Right now, there are a few markets where IPF is showing a price/performance advantage, but there seems to be a lot of hidden costs and fine print. The infrastructure still lacks in some areas, which will add cost any way you look at it. I think that most of the IPF early adopters buy it because they see IPF as inevitable, so it pays to get the IT staff familiar with the architecture. That's what generally makes sense for new architectures. I have a feeling that some are buying Opteron for the same reason.
Even so, I wouldn't assume that the market is only big enough for one major architecture. I think in ten years we'll still see x86 alongside IPF, each specializing in different aspects of the market. I could be wrong, but I think that Intel would benefit the most from keeping IPF higher margin to use it as an incremental ASP lift, rather than completely commoditizing it top to bottom.
Market Data 
Markets 
AI
