Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!decvax!cca!ima!inmet!bhyde
From: bhyde@inmet.UUCP
Newsgroups: net.arch
Subject: Re: Univac 1100 registers - (nf)
Message-ID: <610@inmet.UUCP>
Date: Mon, 12-Dec-83 23:42:16 EST
Article-I.D.: inmet.610
Posted: Mon Dec 12 23:42:16 1983
Date-Received: Wed, 14-Dec-83 02:17:55 EST
Lines: 14

#R:umcp-cs:-429500:inmet:2500009:000:680
inmet!bhyde    Dec 10 16:20:00 1983

In the 1108 that I learned to program on the fact that registers were
memory was only sort of implemented.  The I/O calls to the operating 
system if pointed at the "registers" would read/write the memory
that was really there, but the instructions would see the registers.
This was a lot fun for passing secret messages between processes.  A
note left in the memory under the registers could often be picked up
a week later, love that core.
  Of course the PDP-10 treated registers as low core, no self respecting
compiler would fail to implement a tight loop by coping the instructions
into registers and branching into them, it would make the operator's
eyes glow.				ben hyde