Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!asuvax!mcdphx!udc!neutrino.urbana.mcd.mot.com!dfields
From: dfields@neutrino.urbana.mcd.mot.com (David Fields)
Newsgroups: comp.sys.m68k
Subject: Re: Why no fcmp/fbcc on 68040?
Message-ID: <1279@urbana.mcd.mot.com>
Date: 14 Jun 90 14:53:42 GMT
References: <16340002@hpfinote.HP.COM>
Sender: netnews@urbana.mcd.mot.com
Reply-To: dfields@urbana.mcd.mot.com
Lines: 16

In article <16340002@hpfinote.HP.COM>, den@hpfinote.HP.COM (Don Novy) writes:
|>
|>  I have read some articles about the 68040 and am confused about one point.
|>Why have the fcmp and fbcc instructions been eliminated from the
floating point
|>unit? I have found these instructions to be VERY helpful in speeding up some
|>double precision floating point intensive software. One program was
sped up by
|>23% by using fcmp and fbcc instead of 68020 instructions (cmp.l).

I'm not sure which execution unit the fcmp and fbcc instructions are in
but according to the '040 manual I've got on my desk, the instructions
are still implemented in hardware.
            
Dave Fields // Motorola MCD // uiucuxc!udc!dfields //
dfields@urbana.mcd.mot.com