eugene@ames.UUCP (Eugene Miya) (04/03/86)
Please send mail directly back to me, I do not normally read this
news.group.
I was wondering what it takes to port PC/MS/DOS [Time, software, license,
etc.]. Suppose I gave you a radically different architecture. Let's
say, as the joke when in the past, I gave you a Cray. A real Cray- X-MP.
I want be serious about this. Ignore the hardware cost [well I guess a
Cray is not that radically different: 64-bit word oriented rather than
byte oriented architecture, vector instruction set]. I assume some
but not all utilities such as those written in higher-level languages
might be portable. What would it take? I gave this thought because
I just gave a talk on supercomputers at CSU Fresno and this thought
about evolution occurred to me. If we can keep increasing micro Si performance,
is there anything to really stop MS/DOS from moving to a Cray performance
hardware architecture? I note that MS/DOS is looking more and more like Unix
as time progresses, does that stop?
Anyone familiar with the internals of VMS is also welcome to respond.
What do you think it would take to port VMS to a Cray.
I'm not implying that I plan to do this. It a combination of looking
back on some discussions we had three years ago and reflecting,
looking ahead a bit, and the fact that I can on occasion get a Cray
class machine to myself for my research, stand-alone (working on a
future-power micro computer, maybe?). I've been moving radically
languages to the 2, and this is not-unrealistic.
Again, I normally don't read this group, please send me mail directly,
and if I get more than 8 responses, I will repost a summary.
--eugene miya
NASA Ames Research Center
{hplabs,ihnp4,dual,hao,decwrl,allegra}!ames!aurora!eugene
eugene@ames-nas.ARPA