[comp.arch] MIPS CORE ISA

lindsay@k.gp.cs.cmu.edu (Donald Lindsay) (01/08/89)

In article <296@quick.COM> srg@quick.COM (Spencer Garrett) writes:
>Anybody got a reference for the DARPA Microprocessor without
>Interlocked Pipe Stages (or MIPS) Core Instruction Set
>Architecture specification?

My copy says
	Core Set of Assembly Language Instructions 
	for MIPS-based Microprocessors
	Version 3.2
	30 October 1987

	Maintained by
	Robert Firth
	Software Engineering Institute
	Carnegie Mellon University
	Pittsburgh
	Pennsylvania 15213-3890
	(412) 268-6305

As I understand it, MIPS Corp. uses something slightly different, which
is close enough that a mechanical translation is practical. Also,
machines which adhere to this ISA may have differing native machine
languages.  (The standardization is at the assembler level, not at the
binary level. For example, all those famous NOPs are generated at
assembly time.)
-- 
Don		lindsay@k.gp.cs.cmu.edu    CMU Computer Science
-- 

scarter@gryphon.COM (Scott Carter) (01/12/89)

In article <3980@pt.cs.cmu.edu> lindsay@k.gp.cs.cmu.edu (Donald Lindsay) writes:
>In article <296@quick.COM> srg@quick.COM (Spencer Garrett) writes:
>>Anybody got a reference for the DARPA Microprocessor without
>>Interlocked Pipe Stages (or MIPS) Core Instruction Set
>>Architecture specification?
>
>My copy says
>	Core Set of Assembly Language Instructions 
>	for MIPS-based Microprocessors
>	Version 3.2
>	30 October 1987
>
>	Maintained by
>	Robert Firth
>	Software Engineering Institute
>	Carnegie Mellon University
>	Pittsburgh
>	Pennsylvania 15213-3890
>	(412) 268-6305
>
The current chairman of the SAE Aerospace commitee in charge of the ISA is

Terry Rasset
McDonnell Douglas Astronautics Company
5301 Bolsa Avenue
Huntington Beach, CA 92647

(or you can contact me, same address, 714-896-3097)

>As I understand it, MIPS Corp. uses something slightly different, which
>is close enough that a mechanical translation is practical. Also,
>machines which adhere to this ISA may have differing native machine
>languages.  (The standardization is at the assembler level, not at the
>binary level. For example, all those famous NOPs are generated at
>assembly time.)
>-- 
>Don		lindsay@k.gp.cs.cmu.edu    CMU Computer Science

Actually, the ISA (Instruction Set Architecture) is (or has become) a virtual 
machine definition where the virtual machine rather resembles the MIPS Inc ISA.
The virtual machine instructions are [mostly] specified using a syntax
similar to the MIPS assembler [note: _not_ MIPS machine instructions].  A
processor [or processor simulation :)] vendor provides a translator and
code reorganizer for the Core-to-target translation.  The virtual machine is
intended primarily as a compiler [i.e. Ada] target, although people do write
assembly directly in Core.

The basic Core ISA is pretty well established.  The major work in progress is 
1) including aliasing information directives to make code reorganization more
efficient; and 2) specifying the extended processor environment (exceptions,
interrupts, I/O instructions, MMU/cache control, clock/timers, multiprocessor
operations) in a way that is both machine-independent over the target set and
low-level enough to be useful.

The hope is that the Core ISA will become a military standard, essentially the
32-bit equivalent of the 1750A (some of us hope that Core will be useful
outside the aerospace community).

Further info on request.

Scott

These opinions are the opinion of the writer, and in no way reflect the
position of McDonnell Douglas Inc.