Re: and emulators to try?

["Toru Nishimura" <locore64%alkyltechnology.com@localhost>]

Guys,

> >  I have installed NetBSD on i386 hundreds of times. And on amd64 maybe 10
> >  to 20 times. I don't have much familarity with installing NetBSD on other
> >  ports.
>
> under simh-vax:
>
> vaxnbsd# uname -a
> NetBSD vaxnbsd 4.0 NetBSD 4.0 (GENERIC) #0: Sun Dec 16 01:45:50 PST 2007  
> vaxnbsd# dmesg
> ...

It's always interesting to see how NetBSD boots on exotic computers.  dmesg
output is fun for NetBSD geeks who know what it means.  Now I would like
to dream a larger picture.  There are many "computer emulators" on screen,
however, their implementations and approachs are scattered at best.
Technically speaking it should be possible to try coordiation/merge/unification
to make a foundation for 'virtual computers'.  Just out of my head, well
articulated components should be like;

- CPU instruction emulator w/ various super-speedy-runtime technics
- framework to minics computer internals; whole interrupt scheme,
 bus orignizations and of course popular and/or exotic devices.

It might be more promising and challenging to aim beyond to build
"obsolute computer shop."  The market success of VMware and Xen
has shown people a clear perspective on how computer HW/SW can
be emulated and utilized on different breed systems.  NetBSD/xen
has a special signicance that the very same approach could be
extended to foster 'specially designed emulated computers built
upon Xen micro kernel .'  Existing "emulators" run as a userland
application.  Though the design is handy and useful to run, it has
subtile limitations by being so.  Then it' should be possible to build
"a guest OS to pretend other CPU and other computer" with
sophisticated instruction emulators/interpreters to do clockwork
ticking userlands.  It's norm today a modern computer to have at
least two CPU core inside.  Any taker?

Toru Nishimura / ALKYL Technology