Article 8799 of alt.sys.pdp10:
Path: news3.best.com!news1.best.com!newsfeed.mathworks.com!howland.erols.net!outgoing.news.rcn.net.MISMATCH!feed1.news.rcn.net!rcn!not-for-mail
From: "Alan H. Martin" <AMartin@MA.UltraNet.Com>
Newsgroups: alt.sys.pdp10
Subject: Re: death of architectures...
Date: Mon, 22 May 2000 01:17:17 -0400
Lines: 31
Message-ID: <3928C2DD.B1FC7708@MA.UltraNet.Com>
References: <547leyms3t.fsf@flipper.cisco.com> <38F5F2E9.A051A9BB@jetnet.ab.ca> <38f7cd59$0$227@nntp1.ba.best.com> <8d9hi9$b6f$5@bob.news.rcn.net> <38fa0630$0$200@nntp1.ba.best.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: Zw2WKut0AzgwP2H58x82tIl9GZCRjep1ymi+qUf7OlE=
X-Complaints-To: abuse@rcn.com
NNTP-Posting-Date: 22 May 2000 05:17:56 GMT
X-Accept-Language:  en,en-US,en-GB,es
X-Mailer:  Mozilla 4.7 [en] (Win95; U)
Xref: news3.best.com alt.sys.pdp10:8799

Joe Smith wrote:
> 
> In article <8d9hi9$b6f$5@bob.news.rcn.net>,  <jmfbahciv@aol.com> wrote:
...
> For instance, the PDP-10 had 16 instructions that dealt with boolean
> operations on two variables.  Some of them were downright worthless;
> they existed to make the set of 16 orthogonal.  But compilers used
> only 5 or 6 of the 16 possible instructions.  The design of the SPARC
> CPU (used in Sun workstations) implemented only the 6.  No use wasting
> transistors to implement op-codes that weren't going to make the system
> as a whole run faster was the thinking.

FORTRAN-10/20 boiled down to these 10 with various addressing and
destination modes:

AND	IOR	EQV	XOR	ANDCM	ORCM
SETCA	SETCM	SETZ	SETO

BLISS-10 used at least these 13 (plus at least some of their addressing
and destination variations):

AND	OR	EQV	XOR	ANDCM	ORCM
ANDCA	ORCA	ANDCB	ORCB
SETCA	SETZ	SETO


(I'm more convinced than ever that all 16 operation families takes
little hardware if you provide all of them...)
				/AHM
-- 
Alan Howard Martin				AMartin@MA.UltraNet.Com