On Sun, 3 Feb 2019 12:41:09 +0000 (UTC), Martin Gregorie
 declaimed the following:

>
>Just compare what we were able to do with mainframes, minis and 8-bit
>microcomputers back in the day when interactive devices were 24x80
>textmode terminals or teletypes:
>
>- my university's scientific computer, an Elliott 503, had 8Kwords of 39
>  bit memory (call it 48KB equivalent because each word held two
>  instructions). It used paper tape for input and a lineprinter for
>  output and everything we ran on it was written in Algol 60.
>

 Xerox Sigma-6... Relatively fat machine for the mid-70s
(four-bank/four-port, interleaved) 512kB (though I could swear ours was
somehow doubled to 1Mb*). On a good day we'd often have between 40 and 60
terminals active, along with the batch queue (we actually had a Honeywell
Level-6 being used as a terminal server, replacing racks of Gandalf
equipment https://en.wikipedia.org/wiki/Gandalf_Technologies )


>- after university, the computer bureau where I worked had four
>  teletype terminals plus an RJE device (remote cardreader and
>  printer)and ran additional background jobs under a multiuser,
>  multitasking OS on a 32 Kword (96KB equivalent) 1902S with 2
>  60MB disks, 6 tape drives plus card and paper tape readers and a
>  fast (1250 lpm) printer.
>

 Pretty sure we had 6 100MB drives, and it was a big thing when we
obtained a 300MB drive to be used for swap space.


>- Back in the MS-DOS days Bill Gates was on record as saying that nobody
>  could ever need more than 640 KB of RAM (the limit for an IBM PC-AT).
>
 The chip could access 1MB, but probably used some of that for memory
mapped I/O (rather than using the separate I/O port/control that Intel
chips supported).



* The memory cabinets were "refrigerator" sized, originally holding
magnetic core memory... However, by the late 70s, the core had been
replaced with 18x18" circuit cards with RAM chips, making for a very empty
looking cabinet.

 The multi-port interleave allowed the CPU and three I/O Processors to
chase each other through memory -- while the CPU accesses, say, word 4, the
IOPs can each access words 2, 3, and 5 (again as an example) -- without
having to put wait states for any other. Closest I've seen to that is the
Amiga -- with the first 512kB being "chip" memory (dual ported, accessed by
CPU and by special chips; extended memory was only CPU accessed)


--
 Wulfraed                 Dennis Lee Bieber         AF6VN
 wlfraed@ix.netcom.com    HTTP://wlfraed.home.netcom.com/

--- SoupGate-Win32 v1.05