Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 7/1/84; site rlgvax.UUCP
Path: utzoo!watmath!clyde!burl!mgnetp!ihnp4!houxm!houxz!vax135!cornell!uw-beaver!tektronix!hplabs!hao!seismo!rlgvax!guy
From: guy@rlgvax.UUCP (Guy Harris)
Newsgroups: net.micro
Subject: Re: "Smalltalk Coming to Micros!" ?... HA!
Message-ID: <7@rlgvax.UUCP>
Date: Sat, 21-Jul-84 17:31:20 EDT
Article-I.D.: rlgvax.7
Posted: Sat Jul 21 17:31:20 1984
Date-Received: Mon, 23-Jul-84 02:10:36 EDT
References: <191@sri-arpa.UUCP>
Organization: CCI Office Systems Group, Reston, VA
Lines: 25

> I suspect that a "native mode" implementation of a smalltalk like system
> could run pretty well on a 10Mhz 68K, but would be more work than anyone
> wants to do.

In the proceedings from the 11th Annual International Symposium on
Computer Architecture, published as SIGARCH Newsletter Vol. 12, No. 3,
June 1984, there's a paper called "Architecture of SOAR: Smalltalk on
a RISC" which describes Berkeley's architectural design for a RISC machine
oriented towards running Smalltalk in "native mode".  Instead of compiling
Smalltalk into bytecodes interpreted by an interpreter program or microcode,
it compiles them into the SOAR instruction set which is a general-register-based
RISC.

> Of course, if you have a processer with user writeable microcode, you might
> be able to do some interesting things... (thats the dorado advantage,
> right?).

That, plus a 70ns cycle time (about 1/3 that of an 11/780).

> Did the vax version use user microcode?

Not from what I can tell from the article in "Bits of History...".

	Guy Harris
	{seismo,ihnp4,allegra}!rlgvax!guy