Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!think.com!spool.mu.edu!snorkelwacker.mit.edu!bloom-beacon!eru!hagbard!sunic!mcsun!ukc!acorn!armltd!abaum
From: abaum (Allen Baum)
Newsgroups: comp.arch
Subject: Re: Compilers and efficiency
Message-ID: <168@armltd.uucp>
Date: 14 May 91 15:37:52 GMT
Article-I.D.: armltd.168
References: <1991May13.211555.28824@rice.edu>
Sender: abaum@armltd.uucp
Distribution: comp
Organization: A.R.M. Ltd, Swaffham Bulbeck, Cambs, UK
Lines: 19

In article <1991May13.211555.28824@rice.edu> preston@ariel.rice.edu (Preston Briggs) writes:

>It's easy to get old instructions (say int x int -> long int) in 1 cycle,
>just slow the cycle time down to old rates.

It's easy to do now, with the same cycle times. Cycle times are
not the issue in this case. The issue is one of returning two
results, and what to do with them. The are various solutions, almost
all are disliked by either compiler writers, architects, or chip
designers.

These solutions include specifying two destination registers in the
instruction, overwriting one or both of the source registers, writing
back to dest reg/dest reg+1 pair, keeping the upper half in a special
register..... you can probably think of more, and come up with
the reasons that each of the groups above dislike it. The reasons
for disliking them aren't terribly frivolous either- they just 
outweigh the advantages of returning both (in the eyes of those
particular groups, AT the particular time - times change!)3