Related articles |
---|
Garbage collection Olin.Shivers@cs.cmu.edu (1992-11-24) |
Garbage collection numbers found Olin.Shivers@cs.cmu.edu (1992-11-29) |
Newsgroups: | comp.compilers |
From: | Olin.Shivers@cs.cmu.edu |
Organization: | Compilers Central |
Date: | Sun, 29 Nov 1992 08:34:55 GMT |
Keywords: | storage, bibliography |
References: | 92-11-138 |
Last week I posted looking for statistics on the debugging time saved by
providing automatic storage reclamation for the programmer. Kelvin Nilsen
sent me a pretty detailed reference, which I include below. I got a lot of
messages asking me to forward any answer that I got, so here it is.
-Olin
------------------------------------------------------------------------------
Reply-To: kelvin@cs.iastate.edu (Kelvin Don Nilsen)
Probably, the report you are thinking of is:
Paul Rovner. On adding garbage collection and run-time types to
a strongly-typed statically checked, concurrent language. TR CSL-84-7,
Xerox Palo Alto Research Center, Palo Alto, CA, July 1985.
in this report, Rovner estimates that developers using Mesa
spend about 40% of their development time implementing memory
management procedures and debugging errors related to explicit
storage reclamation.
Joel Bartlett (bartlett@src.dec.com) has been cited as reporting that
``a large fraction of software-caused total system failures are caused
by memory managment errors.''
Both of these observations are described in:
Benjamin Zorn. The Measured Cost of Conservative Garbage Collection.
CU-CS-573-92, April 1992. (zorn@boulder.colorado.edu)
I do not have the Rovner paper. I wish I did. If you figure out how
to order it, please let me know.
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.