[ Btw - whether you're Paul or not, let me just say it's
fun arguing with you. Arguing with Paul was always fun
and often educational - which is the point. So either
welcome back, or just welcome ;^]

someone (someone@somewhere.com) on 2/6/09 wrote:
>
>Here's the clue you are so desperately missing - Tukwila
>taped out more than a year before Beckton.

So that part I don't think really matters - what matters
is who ends up paying for it, and we always knew that x86
would have to do the whole proper point-to-point protocol
eventually anyway - AMD kind of forced their hand on it.

So whether they taped out IPF first or not is just an
implementation detail. They test out new processes by
taping out SRAM cells - that doesn't mean SRAM cells are
the driving force or the expected money-maker.

However:

>Here's gratuitously piling on:

Not gratuitous - much appreciated, and a much stronger
argument:

>1) IPF QPI and Xeon QPI don't use the same coherency
>protocol (Xeon uses a simplified snoopy scheme suitable
>for 2s/4s, IPF uses a more complex scalable distributed
>directory system). Where do you think most of the "high
>level" complexity is.

If true (and I have no real reason to doubt it), that's
just crazy, but I guess it meshes well with the whole
insane notion of Intel just traditionally not letting x86
grow up. I had really thought they got over that.

And if so, then yes, it does sound like IPF actually got
some real design resources put on it, rather than the
usual third-rate hacks (and I say "third-rate hacks" not
because I know any of them, but just from looking at the
sorry history of the dang thing).

>2) Who do you think the folks are who developed the
>"high level" part of QPI and what part of Intel do they
>work in?

Ex-alpha guys? Those people actually used to have a clue.

Linus