Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!uakari.primate.wisc.edu!brutus.cs.uiuc.edu!ux1.cso.uiuc.edu!ux1.cso.uiuc.edu!m.cs.uiuc.edu!p.cs.uiuc.edu!gillies
From: gillies@p.cs.uiuc.edu
Newsgroups: comp.arch
Subject: Re: Are RISC programs bigger than C
Message-ID: <76700118@p.cs.uiuc.edu>
Date: 18 Jan 90 02:55:00 GMT
References: <606@torch.co.uk>
Lines: 14
Nf-ID: #R:torch.co.uk:606:p.cs.uiuc.edu:76700118:000:582
Nf-From: p.cs.uiuc.edu!gillies    Jan 17 20:55:00 1990


Re: Compiled Code density RISC v.s. CISC.

Is your question theoretical or practical?

The 68K is a CISC, but it's not really a code-density king.  I'd
rather hear specs about the NS32532, or even VAX code.

Try comparing a SPARC to a Xerox Dandelion (with bytecode
instructions).  The Dandelion was "tuned" the instruction set in
insane ways, based on statistical analysis of thousands of lines of
MESA (glorified PASCAL).  For instance, they might add instructions
like load-immediate-0, -1, -2, -4, -7, since 0,1,2,4, 7 were the the
most common constants found in most programs.