Re: The Inquirer: HP re-thinking its IA-64 strategy?
From: Chuck Chopp (ChuckChopp_at_rtfmcsi.com)
Date: 03/05/04
- Next message: JF Mezei: "Re: The Inquirer: HP re-thinking its IA-64 strategy?"
- Previous message: John Brandon: "ADMIN SHOW CONNECTIONS ???"
- In reply to: Andrew Harrison SUNUK Consultancy: "Re: The Inquirer: HP re-thinking its IA-64 strategy?"
- Next in thread: Bob Ceculski: "Re: The Inquirer: HP re-thinking its IA-64 strategy?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Fri, 05 Mar 2004 13:26:19 -0500
Andrew Harrison SUNUK Consultancy wrote:
> This would leave Intel trying to develop what would to all intents and
> purposes be a new CPU platform (no Windows, no commercial UNIX etc)
> and one that only as a revenue stream of ~250 million a year with
> the one available OS.
Windows was running on Alpha up through the beta cycles of Windows NT v5.0
[Win2K]. The 64-bit development work for the Windows NT platform family was
all initially done on the Alpha, too, and continued after the final release
of Win2K [which only ran on x86]. It was only after IA64 chips were leaking
out in small quantities to hardware developers that Microsoft ceased using
Alpha for 64-bit Windows development work, and even then there was
side-by-side co-development on Alpha and IA64 until the whole development
process could be sustained on native IA64 systems. I expect that if
Microsoft saw any advantage what so ever in supporting 64-bit Windows on
Alpha again then they'd embrace the architecture again.
Red Hat Linux was also running on Alpha around that same time period. The
Linux kernel runs on so many different architectures that it would be, I
think, the easiest one to port back onto Alpha again. The compiler
technologies that DEC developed still exist for Alpha and could easily be
brought back to life again, perhaps even brought into the open source world.
Again, if there were some competitive benefit perceived about supporting
Linux on Alpha, you'd see commercial distributions of enterprise Linux
products for Alpha.
Or, to look at it another way, Microsoft an Intel have been strange
bedfellows for some time. At times they seem to operate in collusion for
the Win-Tel monopoly, and at other times they seem to be at odds with each
other. When Microsoft dropped for support of PPC and MIPS and Alpha, they
somewhat pulled the plug on the broader acceptance of those architectures.
Having Microsoft ditch Alpha in favor of IA64 could have been more of an
exercise in back-scratching between Intel and Microsoft with the end goal
being to change peoples' perceptions so that Alpha was perceived as being
inferior to IA64.
>
> You then get to the issue of if Intel could do a new generation
> of Alpha processors and that would be tricky. The compiler teams
> for Alpha all work for HP and if HP dump Itanium then the Intel
> HP relationship is going to be strained.
This comes down to the same situation as with Microsoft. You've got
simultaneous cases of cooperation, collusion and adversarial relationships
occurring all the time in this industry. With the by-outs, take overs and
other mergers, there's even fewer players in the market so you've got these
complex relationships among a smaller # of companies. Take Intel and AMD on
the microchip side and take Microsoft, HP/Compaq/DEC/Tandem, Sun, Novel/SuSE
and Red Hat on the O.S. vendor side and you've got a fine mess to try and
figure out.
>
> In addition many of the Alpha engineers jumped ship rather than
> joining Intel, some have gone to startups, some have gone
> to AMD and some have gone to Sun. Why do you think Sun has a
> microprocessor development facility in the Boston area.
In the past, unless I'm mistaken, Samsung tooled up for Alpha production in
the late 1990's at a facility they had in east asia [Korea or Thailand,
IIRC]. Supposing that the intellectual rights to Alpha were licensed
properly, you could have Samsung or AMD producing Alphas.
>
> However I cannot se HP dumping Itanium its too integrated with
> their DNA for them to do so at this point.
That is the crux of the matter, isn't it? Even though Alpha was the better
architecture, Intel dumped it for what appeared to have been no better
reason than a bad case of "not invented here" syndrome. Raiding the
intellectual property of the Alpha hardware developers to try to make IA64
into something viable never resulted in IA64 being equal to or superior to
Alpha.
What always puzzled me was that Alpha was a real production chip while IA64
was still on the drawing boards, and even at that time Alpha was superior.
Reluctance to adopt IA64, slow response from software vendors to produce
O.S. and application software for IA64 and the relatively high cost of IA64
all seem to have contributed to making the first few years of IA64 pretty
dismal. Alpha had OpenVMS, Linux and Windows NT running on it, had
excellent compiler technology and would have been poised to be more broadly
supported by other O.S. & application vendors, but none of that had any
impact and thus we have no new Alpha development going on today.
-- Chuck Chopp ChuckChopp (at) rtfmcsi (dot) com http://www.rtfmcsi.com ICQ # 22321532 RTFM Consulting Services Inc. 864 801 2795 voice & voicemail 103 Autumn Hill Road 864 801 2774 fax Greer, SC 29651 800 774 0718 pager 8007740718 (at) skytel (dot) com Do not send me unsolicited commercial email.
- Next message: JF Mezei: "Re: The Inquirer: HP re-thinking its IA-64 strategy?"
- Previous message: John Brandon: "ADMIN SHOW CONNECTIONS ???"
- In reply to: Andrew Harrison SUNUK Consultancy: "Re: The Inquirer: HP re-thinking its IA-64 strategy?"
- Next in thread: Bob Ceculski: "Re: The Inquirer: HP re-thinking its IA-64 strategy?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|