🚀 go-pugleaf

savel@hal.COM (Bharat Baliga-Savel) writes:

>hmmm! i was under the impression that it was specifically a separate 
>Instruction Cache and Data Cache. 
>-- 

I think a lot of people use it that way now, but the term predated the
use of caches.  For example, when I was back at Intel many of the
microcontrollers had a EPROM instruction store and a data RAM array...
neither memory could be used for another purpose. This was called a
Harvard architecture at the time (though I'm sure by that time caches
existed).
-- 
-Paul Ries DISCLAIMER: <generic disclaimer, I speak for me only, etc>
UUCP: 	pries@mti.sgi.com                 Phone:	415-390-4481 
USPS: 	Silicon Graphics Computer Systems, MS 10L-135 2011 N. Shoreline Blvd,
	Mountain View, CA 94039-7311

In article <1k1ljgINNf76@gorn.hal.com>, savel@hal.COM (Bharat Baliga-Savel) writes:
|> hmmm! i was under the impression that it was specifically a separate 
|> Instruction Cache and Data Cache. 

That is how the term Harvard architecture is used today, but as
Hennessy and Patterson discuss in Section 1.8 of their book CA:AQA,
the term originated as a description for the Mark series of machines
built at Harvard by Howard Aiken.  These machines had "separate
memories for instructions and data, ... [and they] were regarded
as reactionary by the advocates of stored-program computers." (p. 25)

For more info, you may want to look up some of H&P's references, or
look through the IEEE Annals of the History of Computers.

Mike Smith
smith@das.harvard.edu