Re: porting a compiler to a new architecture

Norman Ramsey <nr@labrador.cs.virginia.edu>
28 Mar 1999 16:57:03 -0500

          From comp.compilers

Related articles
porting a compiler to a new architecture mathai@ecf.toronto.edu (1999-03-09)
Re: porting a compiler to a new architecture albaugh@agames.com (1999-03-10)
Re: porting a compiler to a new architecture jsgray@acm.org (Jan Gray) (1999-03-22)
Re: porting a compiler to a new architecture hawa@celsiustech.se (Hans Walheim) (1999-03-22)
Re: porting a compiler to a new architecture nr@labrador.cs.virginia.edu (Norman Ramsey) (1999-03-28)
| List of all articles for this month |

From: Norman Ramsey <nr@labrador.cs.virginia.edu>
Newsgroups: comp.compilers
Date: 28 Mar 1999 16:57:03 -0500
Organization: University of Virginia Computer Science
References: 99-03-036 99-03-049
Keywords: architecture, optimize

Jan Gray <jsgray@acm.org> wrote:
>I only wish there was a retargetable assembler as easy to port as is lcc. I
>looked at gas but it seemed rather large. Any pointers? Instead I wrote a
>simple two-pass assembler from scratch. This took twice as long as the
>compiler port...


The New Jersey Machine-Code Toolkit can generate the machine-dependent
parts of an assembler from a short machine description. It lacks a
standard syntax for pseudo-ops, etc, into which the instruction
grammar can be plugged. It also lacks an on-disk object-file format.
I have been wanting to remedy these lacks for some time and would be
pleased to work with someone who wanted to put together a retargetable
assembler. (The Toolkit seems mostly to have been used to build
dynamic code generators that go direct to binary, with a smattering of
other applications as well.)


More info at http://www.cs.virginia.edu/nr/toolkit




Norman


Post a followup to this message

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