Gavin Scott wrote:
>
> Courtland asks:
> > What was said of the status of IA64 for MPE?    I don't know of an
> > official status report since 1998.
>
> Work would appear to be well underway on that project though
> obviously their focus at the moment is the N class stuff which is
> also a prerequisite for IA-64 since IA-64 will arrive first in the
> N boxes most likely.

AFAIK, the N-class and kindred are the *only* machines in either the
9000 or 3000 platform which will be "IA-64 ready".  There are internal
bus, memory, and I/O issues unique to the IA-64 and related chipsets
that cannot be expanded to pre-N-class hardware without some form of
retrofit (which is unlikely).

But otherwise I agree completely, the N-class I/O system conversion
(which is actively underway in CSY) is certainly a pre-requisite.  If we
look back at classic-to-RISC as an example, we can 'presume' (or at
least hope) that much of the bootup process will be shared by both the
3000 and 9000 platforms (essentially the same today up through ISL)
although there will very likely be a requirement for an OS fork between
PA-RISC and IA-64 (on both platforms).  I would expect, like the
classics, that initially the bulk of MPE (and HP-UX) would be running in
PA-RISC mode (emulation?) or whatever equivalent on IA-64.

There is a proven basis for this especially with MPE, while this will be
somewhat of a first-pass effort for HP-UX (bearing mind that it was only
in recent years that HP-UX merged into a single code base, after dealing
with 680x0, Apollo, and PA-RISC issues and also workstation versus
server versions (9000/7xx and 9000/8xx).

But of course I'm only speculating :-)

Jeff Kell <[log in to unmask]>