Path: utzoo!utgpu!water!watmath!clyde!rutgers!sri-spam!ames!amdcad!amdahl!nsc!woolsey
From: woolsey@nsc.nsc.com (Jeff Woolsey)
Newsgroups: sci.electronics
Subject: Re: hidden features
Message-ID: <4910@nsc.nsc.com>
Date: 10 Jan 88 09:45:36 GMT
References: <268@stratix.UUCP> <4913@sol.ARPA> <9100@utzoo.UUCP> <2037@ttrdc.UUCP> <1602@brahma.cs.hw.ac.uk> <480@auvax.UUCP> <564@bearcat.rutgers.edu>
Reply-To: woolsey@nsc.UUCP (Jeff Woolsey)
Organization: National Semiconductor, Sunnyvale
Lines: 34

In article <564@bearcat.rutgers.edu> lou@bearcat.rutgers.edu (Lou Steinberg) writes:
>Then there was the CDC 6400 (I think) where the float multiply was
>standard but integer multiply was an extra cost option.  (No, I'm not
>kidding.)  When we purchased an upgrade, a service engineer came out,
>opened the back of the machine, turned key in a lock that was back
>there, and voila!  Instant upgrade.

It just so happens that I have my copy of the blue book ("Assembly
Language Programming for the Control Data 6000 Series and the Cyber 70
Series" by Ralph Grishman, Algorithmics Press, 1974 (yes, that's really
the title)) open to page 115, where it is written:

"3.15  INTEGER MULTIPLICATION AND DIVISION

"	As we mentioned earlier, the original 6000
and 7000 series machines had no single instructions
for integer multiplication and division.  These
operations were performed by converting the operands
to floating point, executing a floating multiply or
divide, respectively, and then converting the result
back to an integer.  Control Data subsequently
realized that by making a fairly simple change in
the floating multiply instruction it would be
possible to perform integer multiplies without the
conversions.  This modified instruction is included
in all recent 6000 and Cyber 70 series, and has
been installed as a "field change" in most earlier
machines."

-- 
-- 
I wish people would stop running in to my car!

Jeff Woolsey  National Semiconductor  woolsey@nsc.UUCP  woolsey@umn-cs.EDU