Interesting. The first version of the Foster chips look like a
loser from the outset: It only supports 2 way multi-processing
(no 4x or greater) and only has 256K on-die cache. I can't imagine
why they're even bothering with this, except to gain earlier
bragging rights that a P4 "Xeon" is available. Sort of like
the current "Xeons" with 133Mhz bus speed and 256K cache. These
are repackaged PIIIs and are relatively unpopular owing to the
small cache. But Intel can claim they have very fast "Xeons" available.
The "Foster MP", which they claim will follow the initial Xeon
Foster "shortly" (but given past delays in Xeon introductions I
wouldn't hold my breath) is the one that actually brings P4
to the server market in a meaningful way.
Kwad