Path: utzoo!attcan!uunet!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!haven!adm!smoke!gwyn
From: gwyn@smoke.brl.mil (Doug Gwyn)
Newsgroups: comp.sys.apple2
Subject: Re: MacMusings
Message-ID: <14261@smoke.brl.mil>
Date: 27 Oct 90 21:28:58 GMT
References: <agate!darkstar!ucscb.UCSC.EDU!unknown@UCBVAX.BERKELEY.EDU> <9010271753.AA23864@apple.com>
Organization: U.S. Army Ballistic Research Laboratory, APG, MD.
Lines: 19

In article <9010271753.AA23864@apple.com> MQUINN%UTCVM@PUCC.PRINCETON.EDU writes:
>On Sat, 27 Oct 90 07:24:26 GMT The Unknown User said:
>>	But having more registers takes up more bits of an instruction
>>if you want instructions to be general and use any register.. (I should
>I don't think I follow.  I don't deny this, I just don't understand it and it
>sounds interesting.

The idea is quite simple.  A machine instruction must specify what is to
be done, to what the action will be applied, what input is used for the
operation, and a manner of doing the action.  All this is encoded in the
instruction "opcode".  When it is possible to infer some of this
information, then no bits are needed in the opcode for that part of the
specification.  Otherwise, some bits must be provided to distinguish
among the possibilities (for example, to select specific source and
destination registers).

One should note, however, that when fewer alternatives exist in the
instructions, in general it will take more instructions to accomplish
a given algorithm.  The trade-offs are not simple to evaluate.