Re: : HP's The Machine kicks Microsoft to the curb in favor of Linux - TechRepublic (fwd)

On Mon, Jun 23, 2014 at 12:04:06AM -0400, D. Hugh Redelmeier wrote:
> You need to break down "they".  HP's goals and Intel's goals appeared
> to be different.  And goals evolved.
> 
> In the beginning, it was an HP project to replace their PA-RISC
> architecture with a big new idea: VLIW.  They decided that they could
> not do the whole job, including fabrication, and partnered with Intel.
> Intel was always #1 in process engineering (making silicon fast).
> 
> Intel appeared to have a different goals: to (eventually) replace x86
> with something in which they could keep out other suppliers (AMD,
> NatSemi, ...).  And to have a seriously performant entry in the
> workstation and server races.
> 
> Besides, x86 was crufty.  Extending an architecture too many times is
> difficult.  The rule of thumb is that you can do it once well (see
> Bell and Newell).  8008 -> 8080 -> 8086 -> 80286 -> i386 was already
> pushing it.

4004?  80186?  8085?

> Worse was the historical baggage from the IPM PC and PC/AT.  Remember,
> EFI came out of the Itanium project.  Look how long it has taken to
> land.
> 
> I have a lot of respect for some amazing folks who worked on Itanium
> in the early days (Muchnick (HP), Fisher (ex-Multiflow), Rau
> (ex-Cydrome), etc).  Compiler technology was crucial.
> 
> Many ambitious projects fail.  Some have the goalposts adjusted.
> Failing fast is often a blessing.  The Itanium has suffered from an
> extremely long failure -- 20+ years!

Well someone kept throwing money at it for some reason.

It did succeed at one goal.  It killed a number of better CPU designs.

-- 
Len Sorensen
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists