On compilation Re: [plt-scheme] 301.4

From: Alexander Friedman (alex at inga.mit.edu)
Date: Wed Feb 1 06:16:56 EST 2006

On Feb  1, Noel Welsh wrote:
> --- Matthew Flatt <mflatt at cs.utah.edu> wrote:
> > MzScheme (non-3m) now includes a just-in-time native-code
> > compiler for i386 and PowerPC. 
> ...
> > For now, typical speedups from JIT compilation are in the
> > range 1x (i.e., no speedup) to 1.3x
> This is nice to see, though the performance isn't very
> impressive.  I assume the intended replacement is Larceny.
> [...]
> I'm not sure there really is a point to this email, other
> than to suggest the compiler developers look at LLVM.  Had
> I the time, I investigate myself.  I've got as far as
> compiling LLVM -- so far that's the one downside of the
> project - it takes a whole day to compile on my machine.

Ack - I'm sorry you wasted your time writing this email.

The compiler I am writing does in fact target LLVM (for all the
reasons you mentioned). In fact, it also uses the scheme->scheme
portion of the Larceny compiler for optimization :)

It currently (mostly) works, but needs a few more features, lots more
polishing, tuning, etc.

One thing I would actually like to have more of is 'real-word'
benchmarks to run it on.

FWIW, llvm is quite slow to compile, but it takes less than 10 minutes
on my speedy new machine. I suppose binary distributions will become
slightly more popular.



Posted on the users mailing list.