Re: Machine descriptions for code generation?

"Evan Lavelle" <>
6 Nov 2002 12:01:07 -0500

          From comp.compilers

Related articles
Machine descriptions for code generation? (Evan Lavelle at dot) (2002-10-13)
Re: Machine descriptions for code generation? (Rainer Leupers) (2002-10-18)
Re: Machine descriptions for code generation? "" (Gert Goossens) (2002-10-25)
Re: Machine descriptions for code generation? (Evan Lavelle) (2002-11-06)
Re: Machine descriptions for code generation? (Gopi Bulusu) (2002-11-08)
| List of all articles for this month |

From: "Evan Lavelle" <>
Newsgroups: comp.compilers
Date: 6 Nov 2002 12:01:07 -0500
Organization: Riverside Machines Ltd.
References: 02-10-035
Keywords: architecture
Posted-Date: 06 Nov 2002 12:01:07 EST

On 13 Oct 2002 16:37:06 -0400, "Evan Lavelle at dot"
<> wrote:

Sorry guys, I've been away for a couple of weeks and missed your
replies. I'm familiar with LISA, and the work at Aachen, but I
couldn't find any real details on LISA and I concluded that it was
proprietary. Similarly, I know about nML, but I couldn't find any
details, and again assumed that it was proprietary. I've just
downloaded the nML presentation from but it doesn't give
any details on who owns, and who can use, the language.

I'm happy to pay real money for a compiler, but it has to support a
multi-issue architecture (I don't think Chess does that). And, of
course, it has to be better than GCC, or I'm wasting my money.

The next problem is that I also need to use an MDL internally, for my
own (non-compiler) tool development. So I want to write an MDL
description of my (new) architecture, in a non-proprietary language,
and then use that same description to target to a commercial or
non-commercial compiler.

In the short term, what I've actually done is to write my own simple
MDL. This is actually pretty easy, but it's not good enough for
compiler retargetting. I'm also going to have to do a manual GCC port
as well. Obviously, this is not a good solution, and I'd prefer to use
a public-domain MDL. Preferably, there should also be a public-domain
parser for the MDL, so that I don't have to write my own.

Any other ideas? From my limited research over the last 3 weeks
(corrections/additions very welcome):

MDL MDL public-domain? compiler? compiler ILP support?

lambda-RTL yes? Zephyr yes
ISDL no? none?
nML no? Chess/others? no?
LISA no? LPDP/others? no?
No N/A Pro/Open64 yes?
No N/A GCC no
No N/A SUIF yes
HMDES yes? IMPACT/Trimaran yes
Expression yes Express yes
RADL ? ?


Evan Lavelle
eml @ riverside-machines dot com

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.