joey j> "E core is finished (which should be a safe assumption -- any
objections?)"
rickag>No, but if you go to the link below and to page 4 it says; "-e500
core. Motorola's first G5 offering, optimized for performance and power,
specifically designed for embedded market"
... sounds like marketingspeak to me

. It does have "optimized for
performance and power", see?

. As far as "specifically designed for
embedded market goes", um, see below, i'm replying to you bottom-up for
some reason. The "specifically designed" features could be the
(apparently) DSP-like functionality in the 8540, which Apple most
likely won't be using (and is the most likely delay to the 8540, apart
from maybe the extra simd instructions apple won't require either).
>http://e-www.motorola.com/collateral/SNDFH1101.pdf
>Note that the e500 core is the one used in the MPC8450. Two things stand
out to me.
>1 "the e500 is the first G5 offering"
>2 "specifically designed for embedded market"
>This tells me that a different core will be used for chips for the
desktop market and they will come after the MPC8450.
( ... 8540.)
A different core? Hardly. I doubt Motorola's going to design an entire new
"core" (note the vagueness of the term), particularly if your conclusion
was reached on the basis of marketing speak (their press release was for
the benefit of their customers, i.e. largely the embedded market, they
want to hear that the 8540 is specifically "designed for the embedded
market" whether or not that particular book E implementation goes down
well in desktops or not.) Remember that the same second-generation G4 core
in the 7450 core serves both cisco and Apple, so I can't see why Moto
can't pull the same stunt with the G5 core.
>So for a chip for Apple = new core,
I disagree altogether. A "new core" means either a ground up redesign
(e.g. G5) or at least a major step forward from a previous architecture
(603e -> 7xx). It's not going to happen; the 8540 core is the "G5 core"
which Apple undoubtedly must have prototypes of (at least). Although
that's just my opinion 8)
> tying it to RapidI/O, OCeaN (On-Chip Network)., etc. In my very limited
knowledge this still would be very complicated.
Is apple going to use rapidio? Bear in mind that Apple is listed as part
of the hypertransport consortium. In any case I have read that rapidio is
"ready to go". I don't have the url but the quote was from someone from
the Microprocessor Report as i recall.