Re: Curly soon to be out of a job

From: John Smith (a_at_nonymous.com)
Date: 02/08/05 

Date: Tue, 8 Feb 2005 08:54:25 -0500

JF Mezei wrote:
> Bill Todd wrote:
>> Certainly not in large commercial systems, which POWER will have
>> pretty much to itself for the foreseeable future, though Montecito
>> may manage to reach something like parity with it in small ones.
>
> There is a big difference between Power and IA64: Applications. Like
> Alpha, Power is mature and there has been plenty of time for an
> aplications forest to grow around it. IA64 is still very young and
> just doesn't yet have much of a portfolio.
>
> Yes, HP can brag about X ISVs porting to IA64. But IA64 will remain a
> subset of what is available on PaRisc and Alpha for quite some time,
> epecially for apps that ran on Tru64 and not HPUx, and apps that ran
> on VAX/Alpha but vendor no longer interested in VMS, hence no port to
> IA64.
>
> Last years string of announcements about IA64 have sent a strong
> message
> to ISVs:
> 1- customers are not interested in migrating to IA64. (hence really
> disapointing results)
> 2- a chip which is already retrenching from markets doesn't have a
> bright future
> 3- ISVs aren't interested in architectures that don't have a bright
> future.
>
> So far, all the big "wins" that Rob Young has show here are similar to
> what Cray or CDC would announce back in the 1970s or 1980s. Nice non
> commercial research labs that don't care much about what they run, as
> long as they are given sufficient CPU horsepower. Some run on 8086s,
> siome on Power from IBM, some on Power from Apple, and yes, some on
> that IA64 thing. In each case, the get one platform because it ends
> up being
> the cheapest to them due to teh subsidies the vendors are giving in
> exchange for bragging rights.
>
>
> And one can also interpret those recent VMS sales increases this way:
>
> After the alpha murder, customers delayed sales until there was some
> visibility on IA64's future. Now that there is some visibility,
> customers have decided to buy alphas instead of waiting some more for
> IA64. And should HP really decide to make 2006 a hard "last Alpha
> sale" date, they'd be shooting themselves in the foot since customers
> would
> most likely simply go to Island and buy Used alphas, thus generating
> no income for HP.
>
>> Actually, I was just being open-minded (or perhaps charitable might
>> be a better characterization) given that with the advent of 64-bit
>> Xeons the coffin is already sealed shut whether Opteron itself
>> achieves conspicuous success or not.
>
> Yes and no. As long as the 8086 lacks some of the enterprise
> scalability features that IA64 has, then Intel can probably still
> justify the IA64
> on that basis.
>
> But if AMD makes its 8086 scalable to wildfire/superdome class
> systems, Intel will have no choice but to follow suit in order to
> preserve its 8086 market share and not allow AMD to become the
> predominent 8086 supplier.
>
> And once the 8086 scales to galaxy class machines, there is no reason
> to keep IA64 and its expensive low volume compilers.
>
>
>> Not to me. I always viewed Opteron's likely success in terms of its
>> capabilities, and see no reason to change that assessment now.
>
> What is interesting is that the media now regards AMD as a strong
> competitor, which, in many ways, has a technologically superior
> product
> to Intel, and this causes greater curiosity by the media on the
> respective vendor's market share.
>
> And there are constant rumours that Dell is warming up to AMD. What
> AMD will do is erode Intel's profit margins. And in doing so, Intel
> will
> have more incentives to cut projects that don't yield much profits and
> are a drain on the 8086's thinning margins.
>
> If AMD attacks Intel with a single product line while Intel must
> develop 2 lines, one of which being a real dog (IA64), there will be
> pressure on Intel to kill off the unprofitable IA64 in order to allow
> Intel to focus
> on the 8086 market against the lean and mean AMD.

That's what Intel's directors would do in their role as fiduciaries for
their shareholders.