[gclist] ref-counting performance cost

Kragen Sitaker kragen@pobox.com
Tue, 29 Aug 2000 16:05:52 -0400 (EDT)

A friend of mine is somewhat conservative on newfangled technologies
like garbage collection, which, after all, has only been reasonably
efficient for the last 25 years.

I asserted that reference-counting is almost always much slower than
well-implemented mark-and-sweep or copying collectors; he requested
quantitative evidence.  Real quantitative evidence requires
measurements of real programs, and doing those is a lot of work.  Has
someone done them?

I can't seem to find papers on such measurements, even in excellent
bibliographies like
http://www.xanalys.com/software_tools/mm/bib/authors.html ; even Zorn's
excellent Ph.D.  thesis has no useful information on the slowness of

<kragen@pobox.com>       Kragen Sitaker     <http://www.pobox.com/~kragen/>
Perilous to all of us are the devices of an art deeper than we ourselves
                -- Gandalf the Grey [J.R.R. Tolkien, "Lord of the Rings"]