[NTLUG:Discuss] Killing Off Linux: It's All Academic

[Eric Schnoebelen]

Steve Baker writes:
- Eric Schnoebelen wrote:
- > Steve Baker writes:
- > - I'm a little suprised when you say that V6 didn't have virtual memory.
- > - That old PDP-11 only had 64K words of memory - 128Kbytes.  We often
- > - had several people logged into it at once - it's hard to believe that
- > - without virtual memory.
- > 
- >         It does (did?) have virtual memory, although not as we
- > think of it now..  What V6 did was swap out entire programs when
- > it needed to bring in another program and space wasn't
- > available.  Additionally, processes were limited real memory
- > (likely 32K words on your little PDP-11 above)
- 
- Oh - I see - so it paged physical memory - but not with virtual memory.

	Correct.

- Hmmm - No MMU I guess.

	There was an MMU, but its task was to keep different
processes from touching each other.  I believe it was bright
enough to be able to be able to do VM, but the memory space (32k
words text, 32k words data) was rather too small to be useful?
(remember, VM does depend upon a large, linear,  address space.)

	Given that the later machines suppored up to a several
megabytes of memory, there was an MMU, but it was still mapping
everything into user space as a single 32k word chunk.

- I think our machine really did have 64k words of memory - there
- was some hack on the 11/70 to have a 16 bit data address space
- as well as a 16 bit code space.  I could be wrong about that though.

	Yup, it was called Split I/D space, aka, 32K words of
instructions (text) and 32K words of data.  Some extremely
complex programs were written by using overlays in the 32k words
of text.  You'd have a small portion (8K?) of fixed text, and
the rest of the instruction space would be used for various
routines that would be swapped in and out (overlayed) as needed.
(lots of time spent with loader maps to figure out what
should/could go into each overlay.)

- Compared to that, UNIX was like magic.  The UNIX machine was
- chronically short of hard disk space - so we dumped our entire
- workspace to DEC-tape at the end of every session - and restored
- it again at the start of the next one.  DEC-tapes were really
- neat storage devices. Fast and *nearly* as portable as floppies.

	Yup.. DEC tapes were neat.. Under VMS, they could even
have a filesystem on them, and be randomly accessable. (I don't
remember if the UNIXen of the time could handle that or not..

- We had a PDP-11/20 too - we used it for learning to write OS's
- and device drivers.  It was a totally "bare" machine. No boot
- ROM, no OS, no hard drive.  You had to key in a paper tape loader
- from the front panel switches to get your code loaded.  Between
- those two PDP-11's I learned more practical programming than
- in all of the lectures in 3 years of my degree course. Played
- a lot of core-wars on the 11/20 too!

	My first exposure to UNIX was on a PDP-11/40 running V6 (I
think), acting as a development system for the PDP-11 assembly
programming class.  We had LSI-11/05's with ADM5 terminals at the
workstations, and they were connected to the PDP-11/40.  We assembled
our programs on the UNIX system, and downloaded them via the terminal
line to the LSI.  Four years later, that set up was replaced by an
array of M68k single boards, backed by an HP9000/200 system (running
HP/Unix, I believe). Still fun and cool, but different.. :-)

--
Eric Schnoebelen		eric at cirr.com		http://www.cirr.com
	    Remember when Windows were for washing; a mouse was
		for trapping and a UNIX guarded the harem?