Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uflorida!haven!purdue!mentor.cc.purdue.edu!l.cc.purdue.edu!cik
From: cik@l.cc.purdue.edu (Herman Rubin)
Newsgroups: comp.arch
Subject: Re: Endian reversing MOVEs
Summary: Either it is not too important, or it should be in hardware
Message-ID: <1125@l.cc.purdue.edu>
Date: 10 Feb 89 11:35:16 GMT
References: <759@atanasoff.cs.iastate.edu>
Organization: Purdue University Statistics Department
Lines: 17


We have seen a substantial discussion on exactly how to reverse endianness
in software.  Various methods have been proposed, different ones better for
different machines.  Now several possibilities occur to me.

1.  It is not too important to do quickly.  In this case, any reasonable
method is OK.  Whether it pays to do an efficient job becomes debatable, but
since it is so short, assembly coding is no problem.

2.  It is important to do quickly.  That is to say, a significant amount of
running time will be tied up by the lack of this.  In that case, it should
be added to the hardware.  A small amount of "wiring" makes this as fast
an instruction as a move.
-- 
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907
Phone: (317)494-6054
hrubin@l.cc.purdue.edu (Internet, bitnet, UUCP)