someone (someone@somewhere.com) on 2/7/09 wrote:
---------------------------
>Big system users don't switch architectures on whim, they
>really do "marry" them for 10 or 20 years or more. Someone
>choosing a platform today will ask themselves "will Intel
>still be in the chip making business 10 and 20 years from
>now? What about IBM?"


If someone wants to "marry" an architecture for 10-20 years, why are they going to ask whether Intel and IBM are in the chip making business? If you marry IPF, whether Intel is still in the business 10-20 years down the road has nothing to do with whether they will still be developing and selling Itanium CPUs.

I'll bet they'll still be sold in 10 years, but it will be chips made and stockpiled to meet contractual support obligations, no new development will be taking place.

There are a few diehard IPF supporters who keep saying "just wait 'til next year", essentially. We heard it for Mckinley, then Montecito, then Montvale, until a year ago it was all about Tukwila, now Poulsen will be the one to prove IPF. When it goes late as IPF chips always so, and underperforms expectations, as they always do, whatever the name of the one that follows it will be the call.

One day it'll just be cancelled out of the blue just like Alpha was. The PA-RISC upgrade cycle is pretty much done now (anyone know when the last sale date is for PA-RISC servers?) That means there's no room for growth for IPF, its all a slow decline now, unless Sparc or POWER get canned first.