Advanced Micro Devices Inc (AMD)

[wbmw]

YB, Re: IPF is crearly replacing HP-RISC, Alpha and RS14000. The growth rate will be fast until it will reach the combined market of those chips. At that moment it will hit the wall and stay exactly at that point for years.

This is clearly your interpretation, but I don't think there is any evidence to say that Intel's approach with IPF needs to mirror HP's or SGI's approach with Precision or MIPS, respectively. 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. I think IPF has been gaining against Power and SPARC, too, which is why IBM and Sun have taken such defensive positions. 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. That's keeping in mind that the barrier to Xeon sales requires a very convincing TCO story of combined performance, uptime, and software, so don't take it lightly. IPF is still the "architecture for the next 20 years", but that doesn't happen overnight. Furthermore, I don't know if 64-bit x86 CPUs will take off, but if they do, it will only likely provide an interim solution until the IPF infrastructure can grow to effectively replace it. That doesn't require 20 years of infrastructure, either, since 99% of end-users use software that has been updated over the last few years. It's something to think about before assuming that IPF is nothing more than a high end niche.