To: [EMAIL PROTECTED]
Subject: Re: Evaluating Haskell
David Wilczynski <[EMAIL PROTECTED]> wrote:
> 1) JAVA -- Are there any plans to compile Haskell into byte codes for
> execution on the Java Virtual Machine? The Java issue is very important.
Over the last year, I have developed not one, but two compilers from
Haskell to Java Virtual Machine code. Both are based on the Chalmers
HBC compiler.
My first compiler goes via the conventional G-machine. Using SUN's latest
just-in-time Java 1.1.3 implementation, it goes about half the speed of
the Nottingham Hugs interpreter for small programs, and about 1/8th of
the speed for some of the large programs in the "nofib" suite.
My second compiler goes via a version of the nu-G-machine. This is not
the obvious thing to do, but it works out more cleanly. At the moment,
its performance is approximately comperable to that of the first
compiler, with the exception of the "infer" program from the "nofib"
suite where it is much worse.
I shall give a talk describing the first compiler, and mentioning the
second at the
9th International Workshop on the Implementation of Functional Languages
at St. Andrews in early September.
Ciao,
David
-------------------------------------------------------------------------------
Dr David Wakeling, Department of Computer Science, University of Exeter,
Exeter EX4 4PT, United Kingdom. http://www.dcs.ex.ac.uk/~david
You have to be careful what you say around here, or the system administrators
