[MLton-devel] icfp

[Suresh Jagannathan]

I was thinking a bit more about an icfp submission we might think
about putting together.  Here's a cut at an abstract:

=====================
The Effectiveness of Simply-Typed Intermediate Languages for SML

There has been significant interest in building strongly-typed
intermediate languages (ILs) for modern languages like SML, Haskell,
or Java.  These languages use types in their ILs to ensure the
correctness of the compiler, and to guide advanced optimizations.  In
these implementations, the ILs themselves are expressed using a rich
type system, often tailored to exploit specific characteristics of the
source language.  This approach suffers from two drawbacks.  First, it
complicates straightforward application of optimization techniques
developed for other more general intermediate representations.
Second, sophisticated approaches to represent and manipulate type
information in lower-level ILs are usually required.

In this paper, we present a systematic evaluation of the compilation
strategy used in \mlton, an optimizing compiler for Standard ML.
\mlton\ transforms SML programs early in the compilation process to a
first-order simply-typed IL.  We have developed an allocation and
time-based profiling tool for \mlton, and apply it to a range of
benchmarks, including the compiler itself, to help verify the
feasibility of this approach.  Profiling data indicates that \mlton's
compilation strategy is effective, yielding code efficient in both
runtime and allocation costs.
============================

The main question is how we justify these claims.   Do you have
ideas on the kind(s) of profile information we could generate to
bolster these arguments?  There are some obvious things such as
measuring the space overheads due to monomorphisation or
polyvariance in the defunctorizer, or defining a measure of type sizes at
different passes, but I'm wondering if there is other evidence 
made available by the profiler that would be otherwise difficult to
construct.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel