[Info-vax] OT: Intel Demos 48-Core Prototype Chip

[Neil Rieck]

On Dec 4, 8:14 am, JF Mezei <jfmezei.spam... at vaxination.ca> wrote:
> Neil Rieck wrote:
> > Not sure if this technology will ever make it into Itanium, but lack
> > of Itanium updates makes me think Intel is treating Itanium like a
> > "red headed step child".
>
> It isn't just IF  technology makes it to that IA64 thing. It is also WHEN.
>
> Quadcore 64 bit 8086s with Quickpath (CSI stuff) have been on consumer
> desktops and business servers for over a year. Equivalent IA64s are
> still months away. And they'll be one (or is it 2?) FAB process behind
> current 8086s.
>
> So, by the time the next IA64 comes out, people are likely to yawn at
> the announcement of a chip whose  technology that is already 2 years
> behind the 8086 toy controller.
>
> This message was typed on of of them fancy 8086 toy controllers.
>
> Resistance is futile. You will be assimilated into the Unix collective
> running on 8086s :-) :-) :-) :-)
>
> With regards to skunk works, It isn't so clear if HP still has software
> skunk works.  There are no longer any secret offices in the basement of
> ZKO (the fact that their existance was always forcefully denied is a
> sign that they really did exist :-)
>
> On the other hand, the folks in india, being detached from the
> community, would have an easier time keeping covert projects a secret.

To your points:

1) We all know that Microsoft winning the browser wars (which resulted
in the demise of Netscape) stagnated web technology for ten years.
After wining the browser wars, Microsoft shifted its focus to Sun/
Java. Meanwhile, browser technology continued develop under the guise
of Mozilla/Firefox. Web technology was further developed when Google
introduced AJAX (which Microsoft invented but kept for its own
purposes) into what many call "web 2.0".

Getting back to Itanium, a similar case could be made for Intel. They
were so focused on competition from enterprise architectures (SPARC,
Ultra-SPARC, Alpha, PA-RISC, POWER, etc.) that they were willing to
throw billions at Itanium while underestimating their true competitor,
AMD. Intel thought AMD would only make incremental changes to their
streaming instructions (from "The Pentium Chronicles"). Meanwhile, AMD
surprised everyone while introducing 64-bit extensions to x86 and
HyperTransport. And let's face it, it was much cheaper for AMD to add
64-bit extensions to x86 than it was for Intel to produce a new 64-bit
chip. I now wonder about Itanium's ROI.

2) Let's be realistic. There are only two operating systems in the
word: *nix based (including Linux), and Windows. All the rest are just
two-bit (no pun) players. The only thing these operating systems have
in common is that they already run on x86-64. So-called PC technology
does lock-up on occasion but this has more to do with cheap hardware
(memory with no parity checking just to name one pet peeve). I have
seen Windows-Server-2003 run for 3-years on a hardened platform so you
can't always blame the OS. One reason why Mac OS-X is so popular is
that Apple has more control over what hardware makes it into their box
(kind of like Alpha Servers now that I'm thinking about it). If HP
management has a passion for computing that is similar to IBM
management, then they will already be porting OpenVMS to x86-64.
QUESTION: Why would they do this? ANSWER: There is still value in
selling OpenVMS-based software. By making sure OpenVMS software runs
on the most popular CPU architecture in the world, HP will ensure
future profitability. (this last statement hinges on the difference
between short-term and long-term objectives)

3) HP was always crowing about the lower costs of workers in India. If
this is true, turn loose those programmers on doing a port.

NSR