[Wylug-discuss] Linux on Intel 64-Bit Xeon

[Mark P. Conmy]

On Thu, 23 Jun 2005, Dave Fisher wrote:
> 
> On Thu, Jun 23, 2005 at 12:33:11PM +0100, Phil Foxton wrote:
>> On Thu, 2005-06-23 at 12:15 +0100, Dave Fisher wrote:
>>>    1. Which Intel 64-bit architecture is a 'Dual 64-Bit Xeon'?
>>>
>> EM64T
>>>    As I understand it, Intel has two 64-bit architectures (one that is
>>>    and one that isn't 32-bit compatible).  Am I right about this?
>>>
>> Not that I am aware of, but it might be worth asking Dell if the box is
>> a EM64T arch. box. AFAIK (and I might be wrong so don't quote me) Intel
>> started down the route of 64 with a completely new design (Itanium) but
>> it was NOT backwardsly compatible with X86 so they ended up licensing
>> the X86_64 arch. from AMD.
>
> Thanks Phil, bit I am still a bit confused.
>
>  1. Is EM64T Itanium or X86_64?

X86_64 (or Xeon plus AMD 64 bit extensions. ;-))

>  2.  Isn't AMD's X86_64 capable of running 32-bit x86 code, and
>  therefore "32-bit compatible"? Whereas there never was a 32-bit
>  version of Itanium, so it its "32-bit _in_compatible"? What have I
>  failed to grasp?

Yes.  Yes, No*.  It's not related to Itanium.

* Itanium has a 32 bit emulator on board, but it's not very fast.

EMT64 chips are like the 32 bit version but with 64 bit extensions
which are amazingly similar to AMD's*.  This was a reaction to the
interest in the AMD offerings given Itanium's poor market share.

* Unkind observers have suggested Intel simply reverse engineered
   AMD's chips (which AMD can hardly complain about given *their*
   history), a claim given some weight (but for lawyers out there,
   I'll point out hasn't been verified or proven) by the existence
   in at least one EM64T chip of a bug^H^H^H^H feature that existed
   in the first generation AMD64 chips.

Mark