octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Octave Compiler And Octave Interpreter Source Doku


From: Paul Thomas
Subject: Re: Octave Compiler And Octave Interpreter Source Doku
Date: Thu, 27 May 2004 07:31:02 +0200
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225

Jens Ruecknagel wrote:

Hi,

I am done. My Student research project is done and I can publish it.

Well done! I took a quick spin through it this morning. I think that this is a valuable contribution.



I implemented a minimal Octave to C++ compiler. It is not really good - just a proof of concept. The compiler supports a subset of Octave. It keeps Dynamic Typing. It builds a stack machine to evaluate expressions. This way the execution speed of loops is doubled.


Second I wrote a documentation of the Octave Interpreter source code.
It is the first Chapter of my student research project paper.
Do you think I should separate the doku from the paper? Do you want to include it into the octave Documentation?

For what it is worth, I think that it should be available somewhere.



The paper and the compiler can be downloaded from http://www.stud.tu-ilmenau.de/~rueckn/


As my masters paper I will try to build a compiler with type inference. I am happy about any advice.
I hereby want to start the compiler discussion again.
I got the Majic-Paper, I am reading it right now. Are there any other papers I should read. Is there an octave/matlab Specification, I can use to build the type inference system?

George Almasi is communicative until you ask for a copy of the Majic source. This is true for all his collaborators and his thesis advisor too. Majic is said to be the basis for the just-in-time compiler of Matlab R13. The only documentation I could find on the JIT was a technical note on what it does and does not do. It is an active research topic at Mathworks - they have been hiring for this work.

I think that you will have to do this job from scratch, using Almasi's thesis as a basis (perhaps!). I think that you can be encouraged that he did the type inferenceing, the interpreter and the JIT for his doctoral thesis - thus, a good compiler, including type inferenceing, should be possible for a masters.

Good luck!


Thanks

Jens






reply via email to

[Prev in Thread] Current Thread [Next in Thread]