[Info-vax] Poulson at hot-chips 2011

[JF Mezei]

Neil Rieck wrote:
> Many of us were
> unaware of Alphacide until it happened in 2001.

The deal between DEC and Intel gave Intel immunity from having stolen
Alpha designs for its Pentium 3, and gave Intel permission to continue
to use such designs. I can't recall if DEC gave Intel permission to use
newer Alpha stuff.

In exchange, Intel agreed to relieve DEC of its Hudson FAB plant, and
DEC agreed to port DEC Unix (Tru64) to IA64.  VMS was not part of the deal.

DEC argued it wants its Unix to be multi platform. Remember that at the
time, IA64 was still thought to replace the 8086 as mainstream commodity
industry standard.

Why VMS was left out had more to do with "Digital is killing VMS" than
"Digital is killing Alpha".

But yeah, I guess that the writing was on the wall, but many of us
didn't see it because we were too focused on VMS.


> states: "The Alpha architecture was sold, along with most parts of
> DEC, to Compaq in 1998. Compaq, already an Intel customer, decided to
> phase out Alpha in favor of the forthcoming Hewlett-Packard/Intel
> Itanium architecture, and sold all Alpha intellectual property to
> Intel in 2001, effectively killing the product."


This is where I disagree somewhat. When Pfeiffer bought Digital, he had
plans for Alpha and made it quite clear. He even started a porting of
NSK to Alpha and made sure the next version fo Alpha would have
"lockstep" to support NSK.

He spoke of plans to make Alpha PCs etc. There was a lot of hope
initially. (remember the ads with the "VMS" petrol pump (not "openVMS",
it was written "VMS").

Then Pfeiffer got ousted, all those plans got ousted too. And when they
failed to find a replacement for Pfeiffer and named the accountant Curly
as CEO, his first task was likely to call in the bankers and find a
buyer. I know IB was approached because this is documented in Gerstner's
book.

Curly had been convinced not to kill VMS because it was Compaq's cash
cow. And he would have likely kept Alpha alive as well to support VMS.
But when he ht it off with LaCarly and they were making wedding plans,
it became clear Alpha had to go.

In hindsight, HP should have really taken Alpha and ditched IA64 since
it had been clear for quite some time already that IA64 wouldn't be a
breakthrough.  But the "not invented here" syndrome won.



> (what is the difference between "Alpha architecture was sold" and
> "sold Alpha all intellectual property"? I wonder if the author meant
> to write "Alpha architecture" was licensed rather than sold)

I believe there was another deal at the time of the Alphacide which gave
Intel access to all Alpha intellectual property.  This was more "open"
than the deal between Palmer. (It would have included the work to EV7
which was not yet released).

HP kept the ownership of Alpha.

> So did DEC sell Alpha in 1997?

No. only gave permission for Intel to copy parts of Alpha.

> Or did Compaq sell Alpha in 1998?

Nop. In 1998 Compaq had great plans to leverage the full potential of Alpha


> Or did HP demand that Compaq sell Alpha during pre-merger talks in
> 2001 making it an HPQ decision?

My 8 balls says "signs point to that".


> I'm not sure, but HP kept selling PA-RISC based products in the market
> place for at least two years longer than Alpha.

Because HP-UX was bigger market than Alpha and HP wanted to keep those
customers.