[Info-vax] Poulson at hot-chips 2011

[Neil Rieck]

On Aug 26, 1:45 pm, John Wallace <johnwalla... at yahoo.co.uk> wrote:
> On Aug 26, 7:48 am, Michael Kraemer <M.Krae... at gsi.de> wrote:
>
>
>
>
>
>
>
>
>
> > glen herrmannsfeldt schrieb:
>
> > > Michael Kraemer <M.Krae... at gsi.de> wrote:
>
> > >>There's no evidence whatsoever that intel had stolen anything.
>
> > > In the case of patents, you can infringe without knowing about
> > > the patented design.   So, while "stolen" may sometimes be used,
> > > it is not a requirement.
>
> > Now this may well be possible in that intel vs DEC lawsuit.
> > There aren't too many different ways to build a functioning
> > microprocessor, I guess, the physics is the same for all.
> > But this is far from some people's notion that
> > intel are too stupid to build their own CPU and thus
> > have to "steal" the design from others.
>
> > > Many patents are bought to avoid the possibility of an infringement
> > > suit, even when it isn't known that the technology is actually
> > > infringing.  
>
> > > -- glen
>
> "the physics is the same for all. "
>
> Yes the physics is the same for all, but the details of things like
> branch prediction are not the same for all, and ignorance of the
> existence of a patent is no defence in the courts.
>
> "the ... notion that intel are too stupid to build their own CPU and
> thus have to "steal" the design from others. " (and re "no evidence
> that Intel had stolen anything").
>
> Presumably you're not aware, or have forgotten, that two people who
> had that notion, indeed perhaps started spreading it around, were
> Intel's Chief Executive, Andy Grove, and Chief Operating Officer,
> Craig Barrett, interviewed in the Wall Street Journal in 1996 (August
> 26th)?
>
> E.g.
>
> "the world's biggest chip maker copied and improved upon approaches
> already laid out by minicomputer, mainframe and supercomputer
> designers. But Intel has decided that won't cut it anymore.
>
> "Now we're at the head of the class, and there's nothing left to
> copy," said Craig Barrett, chief operating officer of the Santa Clara,
> Calif., company. Adds Chief Executive Andrew S. Grove: "We're a big
> banana now. . . . We can't rely on others to do our research and
> development for us."
>
> Can't find the original but it's referenced athttp://www.time.com/time/magazine/article/0,9171,986396,00.htmland
> elsewhere.
>
> More recently than that, can readers think of anything original and
> successful that Intel have come up with?
>
> IA64: original but if it had to stand on its own two feet (rather than
> relying on customers dependence on HP-UX and NSK and VMS) it wouldn't
> be classed as successful, it would be dead, like it is in the Linux
> and Windows markets.
>
> Xeon 64 (or whatever EM64T is called this week): introduced after
> years of Intel denying it was necessary or possible, basically a clone
> of AMD64.
>
> Quickpath/CSI: Intel's eventual response to AMD's Hypertransport (and
> where did Hypertransport come from? Well it's kind of EV7-inspired).

Thanks for this, you were reading my mind. It was my intention to do a
little Googling this morning to find some decade-old news articles
supporting the claim that Intel was caught ripping off DEC.

Your very last line, however, is a little wrong. CSI (DEC) and
HyperTransport (AMD lead consortium) were designed to compete with
each other while getting a leg up on Intel (in those days, Intel was
making so much money selling CPUs and chip sets based upon FSB, they
weren't conscious of any alternatives). After the official Alphacide
in 2001, some engineers went to Intel while others went to AMD. As I
understand it, Intel saw these incoming engineers as interlopers so
any new ideas were met with resistance. This is one reason why it took
so long for CSI to morph into QPI which first appeared in
November-2008 on Core-i7.

http://en.wikipedia.org/wiki/Common_System_Interface
quote: Intel developed QPI at its Massachusetts Microprocessor Design
Center (MMDC) by members of what had been DEC's Alpha Development
Group, which Intel acquired from Compaq and HP. Prior to the name's
announcement, Intel referred to it as Common System Interface (CSI).

Many ideas behind CSI influenced successive versions of
HyperTransport. I can only "assume" that this was due to the influence
of the DECies at AMD (but this assumption is just conjecture on my
part)

Anyway, a wise academic once said "the victor writes the history" so
surely Intel would want to present a more honorable narrative. We
would all do the same thing in the same position :-)

Neil Rieck
Kitchener / Waterloo / Cambridge,
Ontario, Canada.
http://www3.sympatico.ca/n.rieck/