Re: Intel neuters Montvale, Itanic screams in alarm

From: Bill Todd (billtodd_at_metrocast.net)
Date: 06/10/05 

Date: Thu, 09 Jun 2005 22:49:15 -0400

Main, Kerry wrote:
>>-----Original Message-----
>>From: JF Mezei [mailto:jfmezei.spamnot@teksavvy.com]
>>Sent: June 9, 2005 6:47 PM
>>To: Info-VAX@Mvb.Saic.Com
>>Subject: Re: Intel neuters Montvale, Itanic screams in alarm
>>
>>"Main, Kerry" wrote:
>>
>>>Some share his negative view, others do not. Time will tell.
>>
>>Actually, I am quite positive. Intel continues to send out the signals
>>that by 2007, the 8086 will be able to replace IA64
>>completely, and that
>>a port of vMS to the 8086 is inevitable.
>>
>
>
> Huh? Where in the world did you pull both of those little gems from?
>
> Please tell the rest of us where you saw:
>
> 1. Where Intel has said 8086 will be able to replace IA64

Intel is developing a common system infrastructure which will support
both Itanic and IA32(extended) in the same socket (finally getting the
on-chip routing and memory-controller facilities that POWER4/4+/5, EV7,
AMD64, and even SPARC have enjoyed for years already) and using the same
chipsets. It's looking more and more like Itanic may get to that party
kind of late, but IA32 should be there in 2007.

This means that whatever advantage Itanic may have enjoyed due to the
lack of credible high-end x86 systems will tend to vanish by that point,
though in fact with IBM's currently-shipping Hurricane Xeon chipset
(which is hot on the heels of the top-of-the-line Itanic TPC-C score at
4 processors and has just edged ahead of it at 8 processors - it
supports up to 32 sockets, but they are apparently working their way up
the benchmark slowly) and the imminent appearance of the Horus Opteron
chipset (also up to 32 sockets), plus whatever larger Opteron facilities
Sun has in its back pocket that Andy B. has been working on for years,
one might suggest that any such advantage is being seriously eroded
starting right now.

  and,
> 2. Where Intel or HP or anyone (besides yourself) in any type of
> official position that has stated a port of OpenVMS to the 8086 is
> inevitable.

That's certainly a more justifiable challenge. It's probably as likely
that HP will just let VMS sink with Itanic as it is that they'll port it
again, and of course there's also the possibility that Itanic will
wallow along for years yet.

...

> If AMD had this attitude about 4 years ago (when Intel was kicking their
> *** in the Mhz race), they would have simply given up and said "whats
> the use?"

I think you may be a bit confused about where AMD was about 4 years ago:
  they were sitting pretty with performance that Pentium could only
dream about (and had been pretty much since Athlon debuted in 1999 and
started pulling away from PIII).

Pentium 4 had just appeared (at 1.5 GHz, vs. Athlon's 1.2 GHz, IIRC) and
while it offered a few more MHz than Athlon the performance was just not
there (and remained not there throughout P4's 180 nm. life, where it
topped out at just 2 GHz).

...

>>Their IA64 strategy became quite clear in early 2004. You are right,
>>they have no need to change that strategy because it is to phase out
>>IA64 by 2007 when the 8086 will be just as capable as IA64.
>>
>
>
> Everything official that comes out of Intel states otherwise

Perhaps it would be instructive for you to contemplate the similarities
with everything official that came out of Compaq about Alpha prior to
June 25, 2001 - and the foreshadowing that all those 'official'
pronouncements were clearly aimed at getting people to ignore.

- bill