[MLton-devel] Optimizing garbage collection efficiency paper

Alain Deutsch deutsch@polyspace.com
Fri, 03 Jan 2003 18:46:47 +0100

I read an interesting article in the last issue of TOPLAS on the impact 
of liveness analysis on garbage collection efficiency.

The main result of the paper is that garbage-collection oriented 
liveness schemes focused on intraprocedural liveness analysis of local 
variables are weaker, in the sense that a bunch of useless data is 
retained, than interprocedural liveness for local + globals. Their 
benchmarks are performed on C/C++/Eiffel code though.

I presume that MLTON, as most ML implementations, uses for determining 
GC roots an intraprocedural analysis ?

I was wondering about your thoughts on the subject.

This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
MLton-devel mailing list