Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!tut.cis.ohio-state.edu!ucbvax!SUN.COM!wmb
From: wmb@SUN.COM (Mitch Bradley)
Newsgroups: comp.lang.forth
Subject: Re: Forth Programs (was Forth learning curve)
Message-ID: <9001021451.AA18382@jade.berkeley.edu>
Date: 31 Dec 89 02:35:14 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Reply-To: Forth Interest Group International List <FIGI-L%SCFVM.bitnet@jade.berkeley.edu>
Organization: The Internet
Lines: 32

> I'm running Microport "Real Unix" System V/AT on my '286 box.  It eats
> about 30% of the CPU cycles.

This isn't so much the fault of Unix, and maybe not even the fault of the
port.  It's the fault of the AT hardware, which was optimized for DOS,
not Unix.

Unix is a multi-tasking operating system with processes which are
protected from each other.  Some hardware support is required in order
to do this efficiently.

One of the reasons that Sun was successful was because the original
Sun hardware was designed by people who understood about MMUs and how
to build them cost-effectively using standard microcomputer technology.
Consequently, Sun boxes ran Unix well.  It really doesn't take that much
hardware, it just takes the right hardware.

Nowadays, the industry in general has learned how to do this, and
the current generation of microcomputer hardware (e.g. the 386)
can support Unix just fine.  In particular, Intel has finally figured out
how to build a decent MMU and how not to cripple a processor with 64K
segments.

I note that Forth doesn't have much of an answer for the 64K segment problem
either.  At every Forth conference, there seems to be some sort of working
group of people trying in vain to cope with Intel segment brain-damage.

Actually, the AT could probably run V7 Unix just fine, in 16-bit mode
just like the PDP-11.  However, the ante has been upped, and nobody wants
V7 anymore.

Mitch