Blade Servers

[mike-DlQxw/23Tq2aMJb+Lgu22Q at public.gmane.org]

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Lennart Sorensen wrote:
> | On Tue, Feb 15, 2005 at 10:09:57AM -0500, mike-DlQxw/23Tq2aMJb+Lgu22Q at public.gmane.org wrote:
> |
> |>Not exactly.  There is a push to go forward with x86-64 hardware, which
> |>doens't neccesarily disqualify intel now that they've co-opted amd's
> |>instruction set.
> |
> |
> | Instruction set yes, performance no.  Throwing an instruction set for
> | 64bit onto a 32bit chip design doesn't make a 64bit chip.  So far the
> | amd64 debian port has found the amd speeds up in 64 bit mode on average,
> | while the intel em64t chips slow down.
>
> No kidding. It's a 32bit core.

What makes you think em64t is a 32bit core?

> Anyway, Intel totally missed the point.
> It's not only the 64bit instruction set, it's the integrated MMU and
> Hypertransport architecture.

Um, I hope you mean AMD's integration of the northbridge on die, which
happens to have an IOMMU.

FWIW:
the 80286 had an integrated segmenting mmu.
the 80386 had both an segmenting and paging mmu.

> Multiprocessor Opteron systems absolutely
> dominate multiprocessor Xeons.
>

Yes, Yes they do.

--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml