RE: IBM SAGE 1955

["Frazee, Walt" <Walt@VistaControls.com>]

I don't know about IBM in 1955, but I am pretty sure the term was in use at
Raytheon Computer when I worked there in the 60's.  Being an Analog type, I
designed Data Acquisition Systems, DACs, and such, for the most part, so
wasn't directly involved with the computer and memory design (except I did
design a core memory sense amp and a current mode line driver).  But I
recall the digital types talking about DMA as short for Direct Memory
Access.

I may have just read too much into the name, but my impression (and after
all this time, that is all I have left to go on) was that "direct" really
meant direct, and going by way of a bus shared with other peripherals
doesn't strike me as very direct.  I think it was more like a dual port
memory, with one port being DMA, dedicated for high speed data I/O.  In
those days, it would have been biax or other magnetic core memory, so "high
speed" didn't mean what it does today.  If that is correct, the DMA
controller would probably have been built into the memory system.

Walt Frazee
Vista Controls,
A Lau Defence Systems Company



-----Original Message-----
From: Kevin Normoyle [mailto:Kevin.Normoyle@Eng.Sun.COM]
Sent: Saturday, December 16, 2000 10:01 PM
To: pci-sig@znyx.com
Subject: IBM SAGE 1955


In just looking around, it appears there
are some people who grant the IBM SAGE (1955)
the claim of first to use "DMA"
(I/O directly accessing memory.. "stealing memory cycles")