At 3:43 PM +0000 4/4/03, olympus-digest wrote:
>Date: Fri, 4 Apr 2003 10:30:51 +0200
>From: "Roger D. Key" <rdk@xxxxxxxxxx>
>Subject: RE: [OM] OT Ancient computers - was so digital ...
>[snip]
>I remember thinking that the IBM360/40 was a backwards step from my
>previous systems - I had been programming since 1962 for LEO Computers, a
>British company, with their LEO III and subsequently LEO 326, both
>multiprogramming systems. The [LEO III] was their first model with transistors
>and used real core memory; the core memory was made up small core rings
>with three (I think) wires running through each core, made up into a
>physical cube.
Three wires is correct: X, Y, and sense. This was standard for magnetic core
memory. The X-wire and Y-wire were each driven with a pulse half the strength
needed to "flip" the core (reverse the sense of magnetization in the donut).
At all places but the intersection, the pulse was too weak to flip a core. At
the intersection, the pulses summed to full strength, causing that one core to
flip, inducing a pulse on the sense wire. Read was destructive: one tried to
clear (zero) all cores representing the bits in the computer word. Where one
saw a pulse on the sense wire, the core had been set to a "one". Where there
was no pulse, the core must already have been set to a "zero". After reading,
all cores were set to zero. Later, the just-read data was written back to the
cores.
Joe Gwinn
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|