On Tuesday, 7 November 2017 at 14:03:31 UTC, rikki cattermole
wrote:
The way I think of it is that Xeon's get all the newest and
greatest features, with them slowly trickling down to the
i-series. Invest in the Xeon production line one generation and
in next use it for i7's ext. Basically R&D cost go all on the
Xeon's and then eventually once its paid off it goes straight
to the consumers.
I see that some features, like instructions, are tested out on
some Xeons first, but others are really only on Xeons and not on
all Xeons either. So I think the Xeons primarily are used as a
tool for differentiating in the high end to maximize profits
(turning on/off different feature sets, possibly from the same
die). I think this largely is the case because AMD isn't
competitive in that segment.
The CPU-architecture generations follow a tic-toc pattern where
the tics mean you have a new architecture and the toc means you
have an improved manufacturing process. I don't think that has
something to do with Xeon.
Looks like they are changing tactic after the last 10 years or
so. I do wonder if you're on the right track and turning a Xeon
into an i9 is just a firmware upgrade...
I imagine that they would try to not use too much die space for
the i9s, and the Xeons seem to require stuff that aren't needed,
so perhaps not likely, but who knows?
